Riverbed® Cascade® Product Family

Deployment Guide

December 2013
© 2013 Riverbed Technology. All rights reserved.
Riverbed®, Cloud Steelhead®, Granite™, Interceptor®, RiOS®, Steelhead®, Think Fast®, Virtual Steelhead®, Whitewater®, Mazu®,
Cascade®, Shark®, AirPcap®, BlockStream™, SkipWare®, TurboCap®, WinPcap®, Wireshark®, TrafficScript®, Flyscript®, WWOS™,
and Stingray™ are trademarks or registered trademarks of Riverbed Technology, Inc. in the United States and other countries. Riverbed and
any Riverbed product or service name or logo used herein are trademarks of Riverbed Technology. All other trademarks used herein belong to
their respective owners. The trademarks and logos displayed herein cannot be used without the prior written consent of Riverbed Technology
or their respective owners.
Akamai® and the Akamai wave logo are registered trademarks of Akamai Technologies, Inc. SureRoute is a service mark of Akamai. Apple
and Mac are registered trademarks of Apple, Incorporated in the United States and in other countries. Cisco is a registered trademark of Cisco
Systems, Inc. and its affiliates in the United States and in other countries. EMC, Symmetrix, and SRDF are registered trademarks of EMC
Corporation and its affiliates in the United States and in other countries. IBM, iSeries, and AS/400 are registered trademarks of IBM
Corporation and its affiliates in the United States and in other countries. Linux is a trademark of Linus Torvalds in the United States and in
other countries. Microsoft, Windows, Vista, Outlook, and Internet Explorer are trademarks or registered trademarks of Microsoft Corporation
in the United States and in other countries. Oracle and JInitiator are trademarks or registered trademarks of Oracle Corporation in the United
States and in other countries. UNIX is a registered trademark in the United States and in other countries, exclusively licensed through X/Open
Company, Ltd. VMware, ESX, ESXi are trademarks or registered trademarks of VMware, Incorporated in the United States and in other
countries.
Portions of Cascade® Profiler Version 9 contain copyrighted information of third parties. Title thereto is retained, and all rights therein are
reserved, by the respective copyright owner. PostgreSQL is (1) Copyright © 1996-2009 The PostgreSQL Development Group, and (2)
Copyright © 1994-1996 the Regents of the University of California; PHP is Copyright © 1999-2009 The PHP Group; gnuplot is Copyright ©
1986-1993, 1998, 2004 Thomas Williams, Colin Kelley; ChartDirector is Copyright © 2007 Advanced Software Engineering; Net-SNMP is
(1) Copyright © 1989, 1991, 1992 Carnegie Mellon University, Derivative Work 1996, 1998-2000 Copyright © 1996, 1998-2000 The Regents
of The University of California, (2) Copyright © 2001-2003 Network Associates Technology, Inc., (3) Copyright © 2001-2003 Cambridge
Broadband Ltd., (4) Copyright © 2003 Sun Microsystems, Inc., (5) Copyright © 2003-2008 Sparta, Inc. and (6) Copyright © 2004 Cisco, Inc.
and Information Network Center of Beijing University of Posts and Telecommunications, (7) Copyright © Fabasoft R&D Software; Apache
is Copyright © 1999-2005 by The Apache Software Foundation; ILOG JView Component Suite is Copyright © 1987-2009 ILOG, Inc.; Click
is (1) Copyright © 1999-2007 Massachusetts Institute of Technology, (2) Copyright © 2000-2007 Riverbed Technology, Inc., (3) Copyright
© 2001-2007 International Computer Science Institute, and (4) Copyright © 2004-2007 Regents of the University of California; OpenSSL is
(1) Copyright © 1998-2005 The OpenSSL Project and (2) Copyright © 1995-1998 Eric Young ([email protected]); Netdisco is (1) Copyright
© 2003, 2004 Max Baker and (2) Copyright © 2002, 2003 The Regents of The University of California; SNMP::Info is (1) Copyright © 2003-
2008 Max Baker and (2) Copyright © 2002, 2003 The Regents of The University of California; mm is (1) Copyright © 1999-2006 Ralf S.
Engelschall and (2) Copyright © 1999-2006 The OSSP Project; ares is Copyright © 1998 Massachusetts Institute of Technology; libpq++ is
(1) Copyright © 1996-2004 The PostgreSQL Global Development Group, and (2) Copyright © 1994 the Regents of the University of
California; Yahoo is Copyright © 2006 Yahoo! Inc.; pd4ml is Copyright © 2004-2008 zefer.org; Rapid7 is Copyright © 2001-2008 Rapid7
LLC; CmdTool2 is Copyright © 2008 Intel Corporation; QLogic is Copyright © 2003-2006 QLogic Corporation; Tarari is Copyright © 2008
LSI Corporation; Crypt_CHAP is Copyright © 2002-2003, Michael Bretterklieber; Auth_SASL is Copyright © 2002-2003 Richard Heyes;
Net_SMTP is Copyright © 1997-2003 The PHP Group; XML_RPC is (1) Copyright © 1999-2001 Edd Dumbill, (2) Copyright © 2001-2006
The PHP Group; Crypt_HMAC is Copyright © 1997-2005 The PHP Group; Net_Socket is Copyright © 1997-2003 The PHP Group;
PEAR::Mail is Copyright © 1997-2003 The PHP Group; libradius is Copyright © 1998 Juniper Networks.
For detailed copyright and license agreements or modified source code (where required), see the Riverbed Support site at https://
support.riverbed.com. Certain libraries were used in the development of this software, licensed under GNU Lesser General Public License,
Version 2.1, February 1999. For a list of libraries, see the Riverbed Support at https://support.riverbed.com. You must log in to the support site
to request modified source code.
This documentation is furnished “AS IS” and is subject to change without notice and should not be construed as a commitment by Riverbed
Technology. This documentation may not be copied, modified or distributed without the express authorization of Riverbed Technology and
may be used only in connection with Riverbed products and services. Use, duplication, reproduction, release, modification, disclosure or
transfer of this documentation is restricted in accordance with the Federal Acquisition Regulations as applied to civilian agencies and the
Defense Federal Acquisition Regulation Supplement as applied to military agencies. This documentation qualifies as “commercial computer
software documentation” and any use by the government shall be governed solely by these terms. All other use is prohibited. Riverbed
Technology assumes no responsibility or liability for any errors or inaccuracies that may appear in this documentation.

Riverbed Technology
199 Fremont Street
San Francisco, CA 94105

Phone: 415.247.8800
Fax: 415.247.8801 Part Number
Web: http://www.riverbed.com 712-00113-05
 Contents

Contents

Preface.........................................................................................................................................................1
About This Guide .............................................................................................................................................1
Audience....................................................................................................................................................2
Document Conventions .............................................................................................................................2
Additional Resources........................................................................................................................................3
Release Notes ............................................................................................................................................3
Riverbed Documentation and Support Knowledge Base ..........................................................................3
Online Documentation...............................................................................................................................3
Contacting Riverbed .........................................................................................................................................3
Internet.......................................................................................................................................................3
Technical Support ......................................................................................................................................4
Professional Services.................................................................................................................................4
Documentation ..........................................................................................................................................4
What Is New .....................................................................................................................................................4

Chapter 1 - Cascade Product Suite Overview.......................................................................................... 5

What Is the Cascade Product Suite? .................................................................................................................5
The Cascade Appliances...................................................................................................................................6
Profiler and Profiler-VE Overview ...........................................................................................................6
Gateway and Gateway-VE Overview .......................................................................................................7
Sensor Overview .......................................................................................................................................7
Shark and Shark-VE Overview .................................................................................................................7
Pilot Overview...........................................................................................................................................7

Chapter 2 - Cascade Appliance Deployment Scenarios ......................................................................... 9

Choosing the Right Equipment.........................................................................................................................9
Choosing the Profiler...............................................................................................................................12
Choosing the Gateway.............................................................................................................................14
Choosing the Sensor ................................................................................................................................14
Choosing the Shark..................................................................................................................................15
Choosing Shark-VE.................................................................................................................................15
Choosing Pilot .........................................................................................................................................16
Deployment Scenarios ....................................................................................................................................16
Deploying the Shark and Pilot.................................................................................................................16
Deploying the Express.............................................................................................................................18
Deploying the Shark and Sensor .............................................................................................................20
Deploying the Express 360, Shark, and Pilot ..........................................................................................20
Deploying the Standard Profiler and Gateway ........................................................................................23

Riverbed Cascade Product Family Deployment Guide iii

Contents

Deploying the Enterprise Profiler and Gateway......................................................................................24

Deploying the Profiler, Gateway, Shark, and Pilot..................................................................................25
Deploying the Profiler, Gateway, Shark, Shark-VE, and Pilot................................................................26
Port and Protocol Dependencies.....................................................................................................................28
Shark and Pilot Port Dependencies .........................................................................................................28
Profiler and Gateway Port Dependencies................................................................................................29
Cascade Appliance Full-Solution Port Dependencies.............................................................................30
Cascade Appliance Enterprise Solution Port Dependencies ...................................................................31
Profiler and Express Flow Storage .................................................................................................................31
Types of Flow Storage.............................................................................................................................31
Flow Rate Estimation ..............................................................................................................................32
Flow Storage Size Estimation .................................................................................................................33
Flow Redundancy with Cascade Appliances..................................................................................................33

Chapter 3 - Flow Collection for Cascade Appliances ........................................................................... 37

Base Requirements .........................................................................................................................................37
Flow Data Fields Consumed by Cascade Appliances ....................................................................................39
Flow Type Considerations ..............................................................................................................................40
Flow Collection Considerations .....................................................................................................................40
Flow Collection in Virtual Environments.......................................................................................................41
Validating Flow Collection .............................................................................................................................41
Sample Third-Party Configurations................................................................................................................43
Configuring Cisco 6500 Series Switches Running Native Cisco IOS ....................................................43
Configuring Cisco 6500 Series Switches in Hybrid Mode .....................................................................44
Configuring Cisco 7500 Series Router....................................................................................................45
Configuring Cisco 7600 Series Router....................................................................................................45
Configuring Cisco 3560 and 3750 Flexible NetFlow..............................................................................46
Configuring the Cisco Nexus 7000 Flexible NetFlow ............................................................................46
Configuring NetFlow Export for Cisco Nexus 1000V............................................................................47
Configuring IPFIX for Avaya (Nortel) 8300 and 8600 ...........................................................................48
Configuring sFlow for HP Procurve 3500, 5400, and 6200....................................................................49

Chapter 4 - Packet Collection for Cascade Appliances ........................................................................ 51

Cascade Appliances for Packet Collection .....................................................................................................51
SPAN and Port Mirroring ...............................................................................................................................52
Remote SPAN and Encapsulated Remote SPAN ....................................................................................53
Sample Port Mirror Configurations.........................................................................................................54
Network Tap Instrumentation .........................................................................................................................57
VACL Configuration Examples......................................................................................................................58
VACL Port Mirroring Configuration on Cisco 6500 Running CatOS ....................................................58
VACL Port Mirroring Configuration on Cisco Catalyst 6500 Running Cisco IOS ................................59
Shark Passthru ................................................................................................................................................60
Packet Deduplication ......................................................................................................................................60

iv Riverbed Cascade Product Family Deployment Guide

 Contents

Snaplen ...........................................................................................................................................................61

Chapter 5 - Cascade Appliances and Steelhead Appliance Integration.............................................. 63

Steelhead Appliance and Cascade Appliances Overview ..............................................................................63
NetFlow Versus CascadeFlow.................................................................................................................64
SNMP Interface Persistence (ifindex) .....................................................................................................65
Cascade Appliance Deployment Considerations............................................................................................66
Enabling SNMP Polling ..........................................................................................................................66
In-Path Deployments ...............................................................................................................................66
Virtual In-Path Deployments...................................................................................................................67
Out-of-Path Deployments........................................................................................................................69
Configuring Steelhead Appliance for Flow Data Export ...............................................................................69

Chapter 6 - Additional Cascade Appliance Integration......................................................................... 73

SNMP Integration...........................................................................................................................................73
SNMP Integration for Flow Sources .......................................................................................................73
SNMP Integration for Device Management of Cascade Components....................................................74
SNMP Integration for Sending Traps......................................................................................................74
SNMP for Switch Port Discovery ..................................................................................................................75
Switch Port Discovery Supported Routers and Switches ...............................................................................76
Active Directory .............................................................................................................................................78
Integration for Active Directory 2008.....................................................................................................79
Integration for Active Directory 2003.....................................................................................................80
REST API .......................................................................................................................................................80

Chapter 7 - Cascade Profiler Analytics and Service Monitoring.......................................................... 81

Analytic License Limits per Profiler Platform ...............................................................................................81
Understanding the Analytics License Limits..................................................................................................82
Reducing the Number of Metrics ...................................................................................................................83
Conditions Required for Baseline Establishment ...........................................................................................85
Determining Which Metrics to Use................................................................................................................85

Chapter 8 - Troubleshooting.................................................................................................................... 87
RTT Values Not Available ..............................................................................................................................87
Not Receiving Reports by Email ....................................................................................................................88
DNS Names Not Being Resolved in Reports .................................................................................................88
Reports Are Not DNS-Resolving All Addresses............................................................................................88
Data in Reports Seems Inconsistent ...............................................................................................................89
Sensor Protocol Violations .............................................................................................................................90
Communication Issues....................................................................................................................................90

Riverbed Cascade Product Family Deployment Guide v

Contents

Switch Port Discovery Troubleshooting.........................................................................................................91

Appendix A - Licensing............................................................................................................................ 93
Licensing Overview........................................................................................................................................93
Runtime Licenses ....................................................................................................................................93
Capacity Licenses....................................................................................................................................94
Option Licenses .......................................................................................................................................94
License Installation.........................................................................................................................................94
Current-Generation Hardware License Installation.................................................................................94
Previous-Generation Hardware License Installation...............................................................................95
Other Device License Installation ...........................................................................................................95
Assigning Licenses .........................................................................................................................................95
Manual License Installation............................................................................................................................95
Automatic License Upgrades..........................................................................................................................96
Evaluation Licenses ........................................................................................................................................96
Licenses Available ..........................................................................................................................................96
Licensing the Express..............................................................................................................................96
Licensing the Profiler and Profiler-VE....................................................................................................96
Licensing the Enterprise Profiler Cluster ................................................................................................97
Licensing Gateway and Gateway-VE .....................................................................................................97
Licensing the Sensor and Shark ..............................................................................................................98

Index .......................................................................................................................................................... 99

vi Riverbed Cascade Product Family Deployment Guide

Preface

Welcome to the Riverbed Cascade Product Family Deployment Guide. Read this preface for an overview of the
information provided in this guide, the documentation conventions used throughout, and contact information. This
preface includes the following sections:
 “About This Guide” on page 1
 “Additional Resources” on page 3
 “Contacting Riverbed” on page 3
 “What Is New” on page 4

About This Guide

The Riverbed Cascade Product Family Deployment Guide describes best practices for configuring deployment the
Cascade Product Suite (Cascade appliances) deployment, including Steelhead appliances, Interceptor appliances, and
third-party appliances. The guide includes information about flow collection, packet collection, metrics, analysis,
troubleshooting, and licensing.
This guide includes information relevant to the following products:
 Riverbed Cascade Profiler appliance (Profiler)
 Riverbed Cascade Enterprise Profiler appliance (Enterprise Profiler)
 Riverbed Cascade Profiler Virtual Edition (Profiler-VE)
 Riverbed Cascade Express appliance (Express)
 Riverbed Cascade Gateway appliance (Gateway)
 Riverbed Cascade Gateway Virtual Edition (Gateway-VE)
 Riverbed Cascade Sensor appliance (Sensor)
 Riverbed Cascade Shark appliance (Shark), including Embedded Cascade Shark, which is Shark functionality
embedded in a Riverbed Steelhead appliance.
 Riverbed Cascade Shark Virtual Edition (Shark-VE)
 Riverbed Cascade Pilot software (Pilot)
 Riverbed Steelhead appliance (Steelhead appliance)

Riverbed Cascade Product Family Deployment Guide 1

Preface About This Guide

For information about the Riverbed Cascade Sensor Virtual Edition (Sensor-VE), consult the appropriate
documentation for that software release.

Audience
This guide is written for network administrators, operators, and engineers familiar with WANs, LANs, and the data
center environment.
You must also be familiar with the information in the following documents:
 Cascade Profiler and Cascade Express User’s Guide
 Cascade Sensor and Cascade Gateway User’s Guide
 Cascade Shark Appliance User’s Guide
 Virtual Cascade Shark Appliance Quick Start Guide
 Cascade Pilot Reference Manual
 Cascade Profiler, Express, Sensor, and Gateway Appliance Installation Guide
 Cascade Profiler and Cascade Express User’s Guide
 Cascade Sensor and Cascade Gateway User’s Guide
 RSP User’s Guide (RSP) or Steelhead EX Management Console User’s Guide (VSP)
 Steelhead Appliance Deployment Guide

Document Conventions
This guide uses the following standard set of typographical conventions.

Convention Meaning

italics Within text, new terms and emphasized words appear in italic typeface.

boldface Within text, CLI commands and GUI controls appear in bold typeface.

Courier Code examples appear in Courier font:

amnesiac > enable
amnesiac # configure terminal

<> Values that you specify appear in angle brackets:

interface <ipaddress>

[] Optional keywords or variables appear in brackets:

ntp peer <addr> [version <number>]

{} Required keywords or variables appear in braces:

{delete <filename> | upload <filename>}

| The pipe symbol represents a choice between the keyword or variable to the left or right of the symbol
(the keyword or variable can be either optional or required):
{delete <filename> | upload <filename>}

2 Riverbed Cascade Product Family Deployment Guide

Additional Resources Preface

Additional Resources
This section describes resources that supplement the information in this guide. It includes the following information:
 “Release Notes” on page 3
 “Riverbed Documentation and Support Knowledge Base” on page 3
 “Online Documentation” on page 3

Release Notes
The following online file supplements the information in this guide. It is available on the Riverbed Support site at
https://support.riverbed.com.

Online File Purpose

<product>_<version_number> Describes the product release and identifies fixed problems, known problems, and
<build_number>.pdf work-arounds. This file also provides documentation information not covered in the
guides or that has been modified since publication.

Examine the release notes file before you begin installation and configuration. It contains important information.

Riverbed Documentation and Support Knowledge Base

For a complete list and the most current version of Riverbed documentation, log in to the Riverbed Support site at
https://support.riverbed.com.
The Riverbed Knowledge Base is a database of known issues, how-to documents, system requirements, and common
error messages. You can browse titles or search for keywords and strings.
To access the Riverbed Knowledge Base, log in to the Riverbed Support site at
https://support.riverbed.com.

Online Documentation
The Riverbed documentation set is periodically updated with new information. To access the most current version of
Riverbed documentation and other technical information, consult the Riverbed Support site located at
https://support.riverbed.com.

Contacting Riverbed
This section describes how to contact departments within Riverbed.

Internet
You can learn about Riverbed products at
http://riverbed.com.

Riverbed Cascade Product Family Deployment Guide 3

Preface What Is New

Technical Support
If you have problems installing, using, or replacing Riverbed products, contact Riverbed Support or your channel
partner who provides support. To contact Riverbed Support, open a trouble ticket by calling 1-888-RVBD-TAC (1-
888-782-3822) in the United States and Canada or +1 415 247 7381 outside the United States. You can also go to
https://support.riverbed.com.

Professional Services
Riverbed has a staff of professionals who can help you with installation, provisioning, network redesign, project
management, custom designs, consolidation project design, and custom coded solutions. To contact Riverbed
Professional Services, email [email protected] or go to
http://www.riverbed.com/us/products/professional_services/.

Documentation
The Riverbed Technical Publications team continually strives to improve the quality and usability of Riverbed
documentation. Riverbed appreciates any suggestions you might have about its online documentation or printed
materials. Send documentation comments to [email protected].

What Is New
The following changes were made since the December 2012 release of the Riverbed Cascade Product Family
Deployment Guide:
 Updated - “Profiler and Profiler-VE Overview” on page 6
 Updated - “Choosing the Right Equipment” on page 9
 Updated - “Deploying the Express” on page 18
 Updated - “Deploying the Express 360, Shark, and Pilot” on page 20
 New - “REST API” on page 80
 Updated - “Licensing Overview” on page 93
 Updated - “Licenses Available” on page 96
 Removed references to Sensor-VE. For information about Sensor-VE, consult the appropriate documentation for
that software release.

Important: Flow per minute limits changed in February 2013. For information about the Cascade Product Suite prior to February
2013, consult documentation for that software release.

4 Riverbed Cascade Product Family Deployment Guide

CHAPTER 1 Cascade Product Family Overview

This chapter provides an overview of the Cascade appliances. It includes the following sections:
 “What Is the Cascade Product Family?” on page 5
 “The Cascade Appliances” on page 6

What Is the Cascade Product Family?

The Cascade Product Family is an enterprise-wide network performance management solution that provides visibility
into your data centers, offices, and users in remote offices. Cascade appliances use network flow data, supplemented
with packet-based performance metrics, to discover applications and monitor performance. Cascade appliances not
only use advanced behavioral analytics to track performance over time and alert you to any deviations from normal
behavior, but enable you to identify and resolve problems before there is an impact on end users.
When you deploy Cascade appliances in your network infrastructure, you gain the following advantages:
 Behavior analytics for proactive monitoring
 Dependency mapping for an always-accurate view of your applications and their dependencies
 Executive-level dashboards for a quick summary of service performance
 Cost-effective visibility into remote sites by leveraging Steelhead appliances
 Ability to plan for and understand optimized WANs
 Ability for true end-to-end coverage of the enterprise with dashboard-to-flow-to-packet analysis, providing
scalability and flexibility
Cascade appliances provide a full set of application-centric, site-centric, and business-centric views so that you can
discover your critical services, optimize performance, and continuously monitor service levels. Cascade appliancess
consistent view of the network breaks down the silos between network, application, infrastructure, and security teams
while shortening the time to resolve problems. In addition, built-in integration with the Steelhead appliance WAN
optimization products provides full visibility and control over application delivery.
For more details about Steelhead appliance integration, see “Cascade Appliances and Steelhead Appliance Integration”
on page 63.

Riverbed Cascade Product Family Deployment Guide 5

Cascade Product Family Overview The Cascade Appliances

The Cascade Appliances

This section describes the following Cascade appliances:
 “Profiler and Profiler-VE Overview” on page 6
 “Gateway and Gateway-VE Overview” on page 7
 “Sensor Overview” on page 7
 “Shark and Shark-VE Overview” on page 7
 “Pilot Overview” on page 8
Figure 1-1. Cascade Appliance Architecture

Profiler and Profiler-VE Overview

The Profiler and Profiler-VE provide, on a single interface, centralized reporting and analysis of the data collected by
other Cascade appliances, Steelhead appliances, and flow exporting routers and switches. The Profiler offers
performance analytics, security analytics, and proactive alerts for delivering application-aware monitoring and
troubleshooting to your network. It combines all network data into a single data set with in-depth views that support
flexible analysis of the information.
Members of the Profiler family of products include:
 Standard Profiler - Standard model designed for mid-level organizations supporting up to approximately 25,000
hosts.
 Enterprise Profiler - Designed to be expandable, supporting volumes larger than the Standard Profiler.
 Profiler-VE - Designed to allow easy deployment as part of a virtualized environment, supporting up to
approximately 4,500 hosts. You can deploy Profiler-VE on VMware ESXi v4.1 and v5.0.

6 Riverbed Cascade Product Family Deployment Guide

The Cascade Appliances Cascade Product Family Overview

 Express - An entry-level product designed for small organizations. It includes Profiler, Sensor and Gateway
functionality in one product and supports up to 75,000 flows per minute.
 Express 460 - An entry-level product designed for small organizations. It includes Profiler, Shark and Gateway
functionality in one product and supports up to 120,000 flows per minute.
For more information about the Profiler, see “Choosing the Profiler” on page 12.

Gateway and Gateway-VE Overview

The Gateway and Gateway-VE collect flow data from routers, switches, and other network devices. These appliances
support most standard flow types (NetFlow, sFlow, J-Flow, IPFIX, and so on). The Gateway aggregates the data,
deduplicates it, compresses it, encrypts it, and sends it to the Profiler. The Gateway can transmit data to up to two
Standard Profilers or Expresses.
You can deploy the Gateway in the same location as the Profiler, or regionally if you have multiple data centers. You
can deploy Gateway-VE on VMware ESXi v4.1 and v5.0.
For more information about the Gateway, see “Choosing the Gateway” on page 14.

Sensor Overview
The Sensor passively inspects packets from port mirroring ports or taps. It provides Layer-7 application classification
and sends performance information, including end-user experience metrics, to the Profiler.
For more information about the Sensor, see “Choosing the Sensor” on page 14.

Shark and Shark-VE Overview

The Shark appliance includes high-performance (1 GbE or 10 GbE) continuous packet capture, storage, and analysis.
You can use the Shark for fast indexing and in-depth analysis of multiterabyte network traffic recordings. You can drill
down to deliver micro-level flow resolution for analysis. The Shark sends performance reports to the Profiler. The
Shark delivers real-time or historical deep-packet inspection and analysis. You can access the Shark using Pilot. The
Shark uses Riverbed's unique XML-based protocol on top of an HTTPS connection for transferring data to Pilot.
Shark-VE is available in Cascade v9.5 and later. Shark-VE operates similarly to the Shark appliance, but it is intended
for use in virtual environments in which you want packet capture and continuous monitoring between virtual hosts.
For more information about the Shark, see “Choosing the Shark” on page 15.

Embedded Cascade Shark Overview

In RiOS v7.0 and later, the Steelhead appliance includes the Shark functionality as Embedded Cascade Shark.
Embedded Cascade Shark software enables on-demand packet capture on Steelhead appliances at remote sites, and
control and analysis of packet captures on remote Steelhead appliances directly from Pilot. As with the Shark, you can
use Embedded Cascade Shark to drill down to deliver microlevel flow resolution for analysis using Riverbed's XML-
based protocol on top of an HTTPS connection for transferring data to Pilot. You do not need to transfer full packets
until you need them.

Riverbed Cascade Product Family Deployment Guide 7

Cascade Product Family Overview The Cascade Appliances

Pilot Overview
Pilot seamlessly and securely integrates with a remote Shark to deliver a complete and feature-rich distributed network
analysis. Pilot is the only tool on the market to be fully integrated with Wireshark software, an open-source network
protocol analyzer. While the Profiler provides visibility across all flows across the network, Pilot provides an in-depth
view into problems requiring deep packet analysis.
For more information about Pilot, see “Choosing Pilot” on page 16.

8 Riverbed Cascade Product Family Deployment Guide

CHAPTER 2 Cascade Appliance Deployment
Scenarios

Deployment of Cascade appliances requires advanced planning to ensure that you install the appliances to capture
critical traffic. You must deploy your appliances efficiently, without wasting resources on unnecessary coverage. This
chapter includes the following deployment sections:
 “Choosing the Right Equipment” on page 9
 “Deployment Scenarios” on page 16
 “Port and Protocol Dependencies” on page 28
 “Profiler and Express Flow Storage” on page 31
 “Flow Redundancy with Cascade Appliances” on page 33

Choosing the Right Equipment

This section describes the equipment choices available to you. It includes the following sections:
 “Choosing the Profiler” on page 12
 “Choosing the Gateway” on page 14
 “Choosing the Sensor” on page 14
 “Choosing the Shark” on page 15
 “Choosing Shark-VE” on page 15
 “Choosing Pilot” on page 16

Important: Flow per minute licensing limits changed in February 2013. For information about the Cascade Product Family prior
to February 2013, consult documentation for that software release.

When determining what kind of equipment you need at each site—whether that site is a data center, a branch office, or
a large building on a corporate campus—answer the following questions:
 What kind of information do I want to know about this location? Do I need response-time information,
optimization information, WAN link bandwidth information, and application usage information?
 Do I have an extensive virtualized environment already in place?

Riverbed Cascade Product Family Deployment Guide 9

Cascade Appliance Deployment Scenarios Choosing the Right Equipment

 How many users and how much traffic am I expecting at this location, now and in the future? Am I considering
how virtual solutions can limit flow capacities?
 What kind of physical resources do I have at this location? Are there technicians that can help maintain the
hardware?
 What kind of network resources do I have at this location? Can my switch provide SPAN and mirror traffic? Can
my routers provide flow information?
 Do I have sufficient bandwidth to transfer flow data between this location and the Profiler?
 How much visibility do I need at this location?
 Do I need packet-level visibility to view objects calls and individual transactions within the application?
The following table shows additional Profiler solutions for several reporting attributes you might want to capture.

Environment Types Profiler Solution

Small environments needing a single appliance solution The Express 360

Small environments needing a single appliance solution The Express 460

with packet capture.

Medium- to large-size environments The Standard Profiler

Large- to enterprise-size environments The Enterprise Profiler

Virtualized environment with limited flow requirements Profiler-VE

The following table shows additional Cascade Product Family solutions for several reporting attributes you might want
to capture.

Tasks Cascade Appliance Solutions

Accurately calculate response-time information for non- The Shark, Sensor or Express
optimized flows

Accurately calculate response-time information for Steelhead appliance and the Shark, Sensor, or Express on the
optimized flows server-side

Report on and monitor link bandwidth information: for The Gateway or Express
example, to monitor percent use

Monitor types of Layer-7 applications on the network: The Sensor or Express 360
for example, to ensure that no one violates the network
usage policy

Obtain detailed packet information: for example, to The Shark or Express 460
analyze network traffic in case of a security violation

Gain visibility into virtualized environments Shark-VE

10 Riverbed Cascade Product Family Deployment Guide

Choosing the Right Equipment Cascade Appliance Deployment Scenarios

In choosing the right equipment, you want to make sure that the data you receive is the data you need. The following
table describes some of the different flow formats supported by Cascade appliances and specifies the features available
within these formats.

Flow Format NetFlow (all CascadeFlow Sensor Shark

variants)

Source of data for the Profiler x x x x

Source and destination IP number, IP protocol ingress interface, x x x x

IP type of service, number of bytes and packets, start and end
times of flow, and TCP flags

Exporting flow device is a Steelhead appliance x

Connection throughput for non-optimized connections x x

Monitor traffic from a SPAN or tap x x

Performance metrics x x

Throughput of 1 Gbps and 10 Gbps x x

Layer-7 application fingerprinting x

VoIP metrics x

Web transaction timing (object load times) x

Ability to forward traffic to multiple Profilers x x x x

Partial (first 256 byte) packet capture, GB storage x

Full and continuous packet capture, TB storage x

Remote management module x x

Different sites have varying numbers of users and volume of network traffic. A site with 10 users transferring large
files all day generates fewer flows than a site with 200 users viewing Web pages. For calculation purposes, Riverbed
recommends that you use 20-40 flows per minute as the estimated average flows per minute. Exact flows per minute
depend on the traffic characteristics in your environment.
Use multiplication to estimate the maximum number of flows per minute. For example, 100 users that each generate
20 flows per minute produce an approximate flow rate of 2,000 flows per minute. However, if the site has servers that
are accessed from remote locations, the overall flow rate is likely to increase, potentially by a large amount. If you have
used flow tools in the past, you might already have some flow estimates. You can also look at session counts on
firewalls or load balancers to obtain flow estimates.
You must have the appropriate number technical staff on site. In a small remote office of only a few non-technical
people, deploying a virtual version of an appliance might make more sense than installing a physical box.
Consider other network equipment that is already installed at a site when you decide which Cascade appliances to
install. If an office site contains multiple large switches capable of generating NetFlow or SPAN and port mirror traffic,
a Shark, Sensor, or Gateway might make sense. Conversely, if a small office contains only a wireless Internet router
with no other network infrastructure, you have limited options for deploying visibility solutions; it is possible you
might not need any Cascade appliances.
If you have a site that reports significant quantities of data across a WAN, consider the bandwidth used by Cascade
appliances for the transfers. Typical WAN bandwidth use is 1.5% of monitored traffic. Cascade appliances report flows
once per minute. If reporting multiple megabytes of traffic per minute seriously impedes the performance of WAN
links, you might need a different solution: for example, restricting the amount of data being monitored.

Riverbed Cascade Product Family Deployment Guide 11

Cascade Appliance Deployment Scenarios Choosing the Right Equipment

Choosing the Profiler

The Profiler is licensed on a flow-per-minute basis, after all flows have been deduplicated. You want to choose the
right Profiler model so that you do not receive inaccurate results and performance issues. Consider the following
factors when you are deciding which type and model of Profiler to install:
 The size of the current network you want to profile
 The planned expansion of coverage

The Express
There are two models of the Express: the Express 360 and the Express 460.
After deduplication, the Express 360 has flow rates ranging from 25,000 to 75,000 flows per minute. The Express 460
has flow rates ranging from 15,000 to 120,000. The Express is best suited for smaller organizations or a small
subsection of a larger network. Because the Express can forward the flows it receives directly to a different model of
Profiler, this deployment can make sense for sites where there is a need for local visibility and enterprise-wide
visibility.
The Express includes functionality similar to that of the Profiler and Gateway. The additional capabilities enable a
compact deployment.
The following table shows the Express model options. Most features are available in the base unit. The base unit is 2U
high, and you can upgrade in the field to the next-higher flow-rate version.

Base Unit and Deduplicate Flow Included Ports Optional Expansion Ports
Flow License Rate

CAX-000360 (L) Up to 25 K FPM Primary 10/100/1000 for Two port 10 Gbps SFP (LR and SR
management option for additional Sensor
CAX-000360 (M) Up to 50 K FPM monitoring)
Four ports 10/100/1000 to monitor
CAX-000360 (H) Up to 75 K FPM Sensor
Performs Layer 7 Application
Identification

CAX-000460 (U) Up to 15 K FPM Primary 10/100/1000 for

management
CAX-000460 (L) Up to 30 K FPM N/A
Four ports 10/100/1000 to monitor
CAX-000460 (M) Up to 60 K FPM Sensor

CAX-000460 (H) Up to 90 K FPM Provides up to 2T of packet storage

CAX-000460 (VH) Up to 120 K FPM

The Standard Profiler

The Standard Profiler has flow limits ranging from 150,000 to 600,000 flows per minute, making it ideally suited for
mid-size organizations with between 4,000 and 25,000 hosts assuming an average of 20 flows per minute per host.
The Standard Profiler cannot forward flows to other Profilers, nor can the Standard Profiler receive flows directly from
flow sources. Because the Sensors, Gateways, and Sharks can forward flows to two distinct Profilers, you can use the
Standard Profiler to monitor a small subset of a larger network. You can send the flows from the Sharks, Sensors, and
Gateways to the local Standard Profiler monitoring a network subset, and to another Profiler.

12 Riverbed Cascade Product Family Deployment Guide

Choosing the Right Equipment Cascade Appliance Deployment Scenarios

The following table shows the Standard Profiler model options. Most features are available in the base unit. Each unit
is 2U high. and you can upgrade in the field to the next-higher flow-rate version. You can add a platform analytics
license (LIC-CAP-2260-CAA) if you want analytics and service dashboards.

Base Unit and Deduplicate Flow Included Ports Optional Expansion Ports
Flow License Rate

CAP-02260 (L) Up to 150 K FPM

CAP-02260 (M) Up to 300 K FPM Primary 10/100/1000 for SAN card (two fiber HBA ports)
management
CAP-02260 (H) Up to 600 K FPM

The Enterprise Profiler

The Enterprise Profiler has a minimum flow limit of 800,000 flows per minute; you can expand the flow limit with
expansion modules up to a maximum of 4.8 million flows per minute. Each module provides support for an additional
400,000 flows per minute. In terms of hosts, a standard Enterprise Profiler can support at least 40,000 hosts, assuming
an average of 20 flows per minute per host, and potentially much more.
The following table shows the Enterprise Profiler model options. Most features are available in the base unit. The base
Enterprise Profiler is composed of two 1U units or one 2U unit. Each expansion module is an additional 2U unit. There
is an additional platform analytics license (LIC-CAP-4260-CAA) if you want analytics and service dashboards.

Base Unit and Deduplicate Flow Included Ports Optional Expansion Ports
Flow License Rate

CAP-04260-UI Part of base system N/A

(required)

CAP-04260-DB Part of base system N/A

(required)

CAP-04260-AN Base system, up to Primary 10/100/1000 for

(required) 800 K FPM management SAN card (two fiber HBA ports)
CAP 4260-EX Expansion unit, each
(optional; add 0-7) one adding 400 K
FPM

CAP 4260-DP N/A: used to balance N/A

(required only when flows on larger
you have five or systems
more 4260-EX)

To plan for future expansion, you must know the current estimated number of flows per minute and the expected flows
per minute in the timeframe being planned for. For example, if the network currently has 6,000 hosts and is expected
to grow to 9,000 hosts in the next two years, a Standard Profiler is sufficient to handle the growth. A software license
and hardware upgrade enables a Profiler licensed for 150,000 flows per minute to be upgraded to 600,000 flows.
Another example is a network currently providing service to 14,000 hosts and expected to grow to 25,000 in the next
year. In this situation, an Enterprise Profiler is a better choice. The Standard Profiler is sufficient in a 14,000-host
network, but it is unlikely to provide adequate performance and license limits to manage the network when it grows to
25,000 hosts.
To provide visibility into a subset of the network and a full overview of the entire network, you can install a smaller
capacity Profiler (such as Profiler-VE, a Express, or a Standard Profiler sufficient for local traffic flow) in combination
with a larger capacity Profiler (such as a Standard or Enterprise Profiler sufficient for overall traffic flow). This ensures
that the system meets both the immediate and planned growth needs.

Riverbed Cascade Product Family Deployment Guide 13

Cascade Appliance Deployment Scenarios Choosing the Right Equipment

Choosing the Gateway

The Gateway is available as both a physical and virtual appliance. The primary differences between the two appliances
are licensed flow capacities and ease of deployment. A single physical Gateway is often sufficient to manage a large
number of devices. The Gateway provides from 150,000 to 1.4 M flows per minute. Gateway-VE is usually sufficient
for smaller branch offices and other locations without significant numbers of flows. Gateway-VE provides 15,000 to
90,000 flows per minute.
The upgrade from the smallest to the largest Gateway within a class family (virtual or physical) requires only a license
change.
The following table shows the available Gateway models.

Gateway Model Deduplicate Flow Rate

CAG-02260-F1 Up to 150 K FPM

CAG-02260-F2 Up to 300 K FPM

CAG-02260-F3 Up to 600 K FPM

CAG-02260-F4 Up to 800 K FPM

CAG-02260-F5 Up to 1.4 M FPM

CAG-1060-VE-F1 Up to 15 K FPM

CAG-1060-VE-F2 Up to 30 K FPM

CAG-1060-VE-F3 Up to 60 K FPM

CAG-1060-VE-F4 Up to 90 K FPM

Consider the following requirements when you deploy the Gateway:

 The flow limits are within the licensed limits.
 The geographic coverage is appropriate.
 The flow capacity of the Profiler is not exceeded by the devices sending data.
 There are sufficient VMware resources available, if you choose to deploy Gateway-VE.
For a mid-size organization with multiple disparate geographic locations, a single 1.4 M flow Gateway can manage
the overall load—you must have the appropriate-sized Profiler because the Gateway can send more flow than the
standard Profiler can receive. However, this might not make sense if a significant amount of your corporate WAN
bandwidth is consumed with the transmission of flow data: for example, from remote sites to a centrally located
Gateway. Because flow data is usually transmitted through UDP, an unreliable protocol, the likelihood of packet loss
is increased the further a packet travels. In this sort of environment, it is a good idea to deploy multiple smaller
Gateways or Gateway-VEs at major locations.

Choosing the Sensor

When you choose Sensor options, consider the number and type of interfaces already installed and your overall system
capability. The Sensor can identify Layer -7 applications. The Sensor can process no more than 5 Gbps of packets. If
you must process more packets than a single Sensor can effectively handle, you need multiple Sensors.

14 Riverbed Cascade Product Family Deployment Guide

Choosing the Right Equipment Cascade Appliance Deployment Scenarios

The following table shows the Sensor option. Start with the base unit and add appropriate media cards.

Base Unit Included Ports Optional Expansion Ports

CAS-02260 Primary 10/100/1000 for management Two port 10 Gbps SFP (LR and SR option)
Four ports - 10/100/1000 copper for
monitoring Sensor

Choosing the Shark

You must choose the appropriate Shark to ensure that the appropriate quantity of packets are stored and are available
for analysis. Deploy the Shark with one or more capture cards.
Consider the number of interfaces possible and the amount of disk space available for storage. Capture cards are
separately ordered items. Model selection depends on the expected network packet rate and retention time you want.
The highest-capacity Shark includes approximately 32 TB of disk space for packet storage.
If you store every packet traversing a heavily loaded 10 Gbps network, the 32 TB of space allows approximately nine
hours of storage (assuming a throughput of 7 Gbps).
If you store every packet traversing a heavily loaded 1 Gbps network, the 32 TB of space allows approximately 65
hours of storage (assuming a throughput of 1 Gbps).
The following table shows the Shark models.

Shark Base Form Factor Storage Maximum Capture Cards

CSK-01100-BASE 1U 4 TB 1 (NIC-CSK-2TX, NIC-CSK-4TX-C, or

NIC-013-4SF)

CSK-02100-BASE 2U 8 TB 2 (supports all card types)

CSK-02200-BASE 2U 16 TB 2 (supports all card types)

CSK-03100-BASE 3U 16 TB 2 (supports all card types)

CSK-03200-BASE 3U 32 TB 2 (supports all card types)

Choosing Shark-VE
You must choose the appropriate Shark-VE model to ensure that the appropriate quantity of packets are stored and are
available for analysis. The available models are shown in the following table. You deploy Shark-VE in VMware ESXi
environments through connectivity to a standard or distributed vSwitch or to a Cisco Nexus 1000V Switch. You install
Shark-VE similarly to other virtual machine disk format (VMDKs), and you use a vSwitch promiscuous-mode port
channel to obtain packets. If you use the Cisco Nexus 1000V, Shark-VE uses a SPAN port. You can limit traffic to only
inter-VM traffic by using the appropriate capture Berkeley packet filter (BPF).
The following table shows the Shark-VE models.

Model Storage Capture Interfaces Export to the Profiler

VSK-00050 50 GB 4 50K FPM

VSK-00200 1 TB 4 50K FPM

VSK-00400 2 TB 4 50K FPM

Riverbed Cascade Product Family Deployment Guide 15

Cascade Appliance Deployment Scenarios Deployment Scenarios

For more information about installing and configuring Shark-VE, see the Virtual Cascade Shark Appliance Quick Start
Guide.

Choosing Pilot
When you deploy Pilot, you must consider only how many users must perform deep packet analysis. Pilot is Windows
client software and is licensed per installed machine. Pilot can analyze traffic from a Shark, an Embedded Cascade
Shark probe, and any standard packet capture files. You do not need Pilot for Profiler-level analysis and
troubleshooting.

Deployment Scenarios
This section describes the following deployment scenarios:
 “Deploying the Shark and Pilot” on page 16
 “Deploying the Express” on page 18
 “Deploying the Shark and Sensor” on page 20
 “Deploying the Express 360, Shark, and Pilot” on page 20
 “Deploying the Standard Profiler and Gateway” on page 23
 “Deploying the Enterprise Profiler and Gateway” on page 24
 “Deploying the Profiler, Gateway, Shark, and Pilot” on page 25
 “Deploying the Profiler, Gateway, Shark, Shark-VE, and Pilot” on page 26

Deploying the Shark and Pilot

You deploy the Shark and Pilot to look at extremely detailed views of network traffic and to improve troubleshooting
of your network-related issues. Because the Shark can only receive packets, you must install the Shark in a location
close to the source of the majority of the packets. Remember that this deployment does not include the higher-level
analysis and reporting that is included with the Profiler.
For most deployment scenarios, you can place the Shark in the data center. The majority of network traffic in most
corporate environments is centered on the data center, making it an ideal place to put a traffic-monitoring solution. You
can investigate and analyze the conversations flowing between end users and the servers in the data center—this is
invaluable when investigating a problem or security incident.
When you install the Shark in the data center, you do not always catch all traffic. However, this uncaptured traffic is
usually not of great interest or significant volume: for example, printing traffic. If you want to monitor traffic that does
not go through the data center, you can place additional Sharks at strategic wiring closets. Because there are many
Shark sizes, you can choose one solution that is appropriate for the data center and a smaller Shark for a remote wiring
closet. For available Shark models, see “Choosing the Shark” on page 15.
You can use other methods to obtain on-demand packet captures from remote locations, and then use Pilot to analyze
them. Although it is not feasible to analyze data center traffic without a packet capture appliance, you can analyze
occasional smaller packet captures without an appliance. Additionally, any Steelhead appliance running RiOS v7.0 and
later includes Embedded Cascade Shark. Embedded Cascade Shark enables you to control and analyze of packet
captures on the Steelhead appliance directly from Pilot.

16 Riverbed Cascade Product Family Deployment Guide

Deployment Scenarios Cascade Appliance Deployment Scenarios

If you connect the Shark to a network mirror port or TAP, you can detect network activity that you want to monitor but
not necessarily store: for example, you can create watches to detect certain conditions without capturing the traffic.
You can preserve only the desired information, thereby reducing the amount of disk space you use compared to a
capture job that captures all traffic.
For more details about watches, see the Cascade Pilot Reference Manual.
To store network traffic, you must define capture jobs on the Shark. Capture jobs tell the Shark what sort of packets to
store to disk. The capture job can store packets based on a wide variety of criteria, including physical source, medium,
or various aspects of packet-header information. You can define multiple capture jobs on each Shark to capture packets
that meet different, specific requirements. For example, you can define one capture job to capture all traffic on specific
VoIP ports, and define another capture job to capture all traffic destined for specific critical servers. The different
capture jobs operate independently of each other and can capture overlapping data.
Use Pilot to analyze the data stored on a Shark and to look at real-time and historical traffic. You can use a variety of
filters and drill-down techniques to analyze traffic. Pilot has numerous views available to assist with the analysis.
Because Pilot can connect to multiple Sharks, the location of Pilot is not as critical as the location of the Shark. You
can optimize the Pilot-to-Shark communication by applying views and filters to the data so that only the necessary
information is sent between the Shark and Pilot.
Pilot does not pull down all of the packets unless prompted to open the packets in Wireshark or save the packets to a
local file. Typically, when you save packets or pull packets into Wireshark, you have already filtered the data so that
you move only specific packets.
When looking at real-time (or near real-time) data, increasing the physical distance between Pilot and the Shark can
force more bandwidth travel on the WAN. Full-packet data sent from the Shark across a WAN, can use significant
amounts of bandwidth and possibly have an impact on other traffic.
To prevent these issues, place Pilot as close as possible to the Shark. If you place Pilot across a slower WAN from the
Shark, you must apply appropriate views and filtering before viewing raw packets or saving these packets locally.
Figure 2-1 shows an example Shark and Pilot deployment. Although this example shows only a single Shark appliance,
you might need additional Shark appliances for large data centers or to monitor additional locations.
Figure 2-1. Example Shark and Pilot Deployment

Riverbed Cascade Product Family Deployment Guide 17

Cascade Appliance Deployment Scenarios Deployment Scenarios

Deploying the Express

The Express is the ideal solution for a small enterprise looking for an advanced monitoring solution. The Express has
the following primary deployment scenarios:
 Acting as a standalone system for smaller network environments
 Integrated as part of a broader system that provides narrower views of portions of a larger network
When you deploy an Express you must choose between an Express 360 or 460. The significant difference between the
two models is that the Express 360 performs Layer-7 application identification, and the Express 460 has up to 2T of
packet storage. Riverbed recommends that if you need application identification you choose the Express 360, and if
you need packet capture and storage you choose the Express 460.
If you need application and packet storage, look at the “Deploying the Express 360, Shark, and Pilot” on page 20.

Standalone Deployment
The Express is designed to act as a standalone system capable of both receiving network traffic information and
providing all the reporting and monitoring functionality expected of Profiler solutions. You can achieve this with two
1-GB network monitor interfaces for receiving packets and flow data directly from the source.
The physical location of the Express is extremely important. Generally, you install the Express at a data center, close
to the core switching and routing infrastructure. This location creates the shortest connection paths between devices,
and provides the most flexible monitoring. Because there are only two monitoring interfaces for receiving and
analyzing SPAN, mirror, and TAP traffic, you must place the device as close to the sources of that data as possible.
Additionally, because the Express can receive flow data directly from flow sources, it is important to place it close to
sending devices so there is no impact on the WAN.
Figure 2-2 shows an example of a standalone Express deployment. Flow is collected locally at the data center from
routers and Steelhead appliances, and additional flow is collected from remote sites. There is port mirroring of traffic
for critical applications, sent directly to the Express monitoring ports.
Figure 2-2. Example Standalone Cascade Express Deployment

18 Riverbed Cascade Product Family Deployment Guide

Deployment Scenarios Cascade Appliance Deployment Scenarios

Integrated Deployment
When you deploy the Express as an integrated deployment, consider what portions of the network you want visible to
users of the different devices being deployed. There is no way to restrict visibility directly within the Profiler, but you
can use an Express to simulate limited visibility.
The Express receives data directly from the source—through built-in monitoring ports and direct flow reception—and
can forward the data to other Profilers. Because of this, you can use the Express to limit the view to a subset of the
network and still feed a larger system to provide a full view.
When you deploy an Express with a limited view, make sure that the Express supports the required flow rate for the
covered area. The largest Express has a limit of 90,000 deduplicated flows per minute, and it can be quickly
overwhelmed by larger deployments. For example, using an estimate of 20 flows per minute per host, 90,000
deduplicated flows should provide coverage for a network consisting of roughly 4,500 hosts; flows over the 90,000
are dropped.
Twenty flows per minute per host is an approximation of how much traffic a host can generate on a per-minute basis
for internal communications. If the host is also communicating with resources on the Internet, then the number of flows
generated is higher.
Consider physical location when you deploy the Express in an integrated environment. The Express provides all the
same advanced network metrics as other Profilers when installed correctly. For example, to provide accurate optimized
response time, you must deploy the monitoring interfaces between the server-side Steelhead appliance and the servers.
If you place the Express in the wrong location (for example, between the client-side Steelhead appliance and the
clients), you can prevent accurate collection and calculation of these values.
If you install the Express in close proximity to the data center, an do not plan appropriately, you can lose visibility into
other important areas of the network. If the data center is in one physical location but you are going to use the Express
to monitor a separate physical location, you have to choose between not having local visibility or not having certain
information available. You can avoid this issue by deploying an additional component, such as the Shark. For details,
see “Deploying the Express 360, Shark, and Pilot” on page 20.
Figure 2-3 shows the Express as part of a larger deployment that includes a Standard Profiler. This example shows that
the local network operator monitors all traffic on the Express and can configure local policies and local service
dashboards. The data received by the Express is also sent to a global Profiler. Collection from other sources by the
global Profiler is not shown.
Figure 2-3. Example Express in a Larger Deployment Environment

Riverbed Cascade Product Family Deployment Guide 19

Cascade Appliance Deployment Scenarios Deployment Scenarios

Deploying the Shark and Sensor

This deployment scenario presents one approach to using the Shark and Sensor in the data center. If you need Layer-
7 information exported to the Profiler, you must use the Sensor. Layer-7 identification is typically most important for
data entering and leaving the data center. Layer-7 application activity can identify certain client-to-server applications
that might use dynamic port assignments and detect potential security issues.
Because Layer-7 traffic detection is largely limited to traffic entering and leaving the data center, consider placing the
Sensor at the data center edge or in the data center core. You can selectively capture data entering and leaving the data
center through WAN-bound interfaces. However, traffic within the data center is typically more secure and better
defined, so Layer-7 identification within the data center becomes less of a requirement.
The Shark appliance provides high-rate packet capture and performance information to the Profiler. The Shark is
installed in the data center core or distribution tiers.
Alternatively, traffic within the data center is usually where packet capture becomes more of a requirement, because
you might want packet analysis for client-to-server traffic and server-to-server traffic.
Figure 2-4 shows the Sensor deployed in the data center core, and SPANs ports or VLANs for traffic going to and from
the WAN. This placement enables you to identify all Layer-7 traffic entering and leaving the data center, and you can
measure performance and retransmission at the core. One or more Sharks are deployed at the distribution tier. This
enables packet collection and analysis for data destined for the data center infrastructure and for data remaining within
the data center infrastructure. You can use Layer-4 mappings (naming of applications by IP and port) for all data
collected, in addition to standard port name definitions. All necessary data is collected for service monitoring and
analytics because both the Sensor and the Shark record and report response-time metrics, resets, retransmissions,
connection counts, and traffic volumes per conversation.
Figure 2-4. Example Shark and Sensor Deployment

Deploying the Express 360, Shark, and Pilot

This deployment is best for smaller environments where you want high visibility, dashboards, and examination, in
addition to Layer-7 application identification.

20 Riverbed Cascade Product Family Deployment Guide

Deployment Scenarios Cascade Appliance Deployment Scenarios

In addition to acting as a standalone device, you can configure the Express 360 to receive flow data from additional
sources and integrate with other products. To provide both high- and low-level views of the data on the network, you
can integrate the Express 360 with the packet capturing abilities of the Shark appliance and the detailed analysis
capabilities of the Pilot. This deployment has many advantages, including that the Express 360 provides a high-level
view of network activity but still allows easy drill-down access to the analysis engine on the Pilot.
When you use the Express 360 with the Shark and Pilot, the design considerations are similar to deploying the Express
independently or as part of an integrated solution. Primarily, consider the flow limits. The Express 360 is limited to no
more than 75,000 deduplicated flows, and the Shark can export up to 600,000 flows per minute (depending on the
incoming traffic rate). Excess flows from the Shark (if any) are dropped by the Express and are not used in analysis or
stored on the Express 360.

Riverbed Cascade Product Family Deployment Guide 21

Cascade Appliance Deployment Scenarios Deployment Scenarios

When you deploy the Express 360 with the Shark, it is not critical that you place the Express 360 in the data center,
for the following reasons:
 You can leverage the Shark interfaces (instead of the Express 360 interfaces) to monitor the traffic between the
server-side Steelhead appliance and the data center.
 You can use the Shark to augment reporting of non-optimized response-time information, rather than relying on
only the Express 360 to provide that data.
In an environment with multiple physical locations, you can remove the restriction of where you locate the Express
360. For example, you benefit from local visibility in a larger environment. You can place the Express 360 at a remote
office, monitoring local flow data and network traffic, and configure the server-side Steelhead appliance to forward
only traffic not already sent from the Express 360. A server-side Shark appliance can also detect optimized flow
response-time values.
One significant benefit of integrating the Express 360, the Shark, and Pilot is that you can use the Express 360 high-
level view of the network and automated monitoring functions, while retaining the ability to drill down to more detail
with Pilot, and even further, with Wireshark.
If you want to view packet data, you must define a capture job to capture the packets on the Shark. Riverbed
recommends that you set a capture job to store all packets being forwarded to the Express by the Shark, based on:
 your particular monitoring preferences.
 capacity.
 the network flow rate you want to monitor.
 other uses of the Shark.
Figure 2-5 shows an example deployment that includes the Express 360, the Shark, and Pilot. Routers and Steelhead
appliances send flows from the network to the Express 360, providing broad visibility into the network. A Shark
monitors traffic from switches in the data center, collecting packets for deeper visibility. The data from the Shark
merges with the flow data collected. You can log in to the Express and view applications flowing across the entire
network. When troubleshooting, if you need deeper packet-level analysis, the Profiler UI automatically launches
appropriate information into Pilot. This takes you from the Profiler view of network traffic directly into Pilot views of
packet data.
Figure 2-5. Example Express, Shark, and Pilot Deployment

22 Riverbed Cascade Product Family Deployment Guide

Deployment Scenarios Cascade Appliance Deployment Scenarios

Deploying the Standard Profiler and Gateway

This deployment is most common in an environment that encompasses a company with several thousand hosts and
several different physical locations. The primary objective of this deployment is to provide visibility into what is
happening on the network from a very high level.
Because the primary objective is to provide basic network visibility, you do not need the Shark or the Sensor to provide
network performance information and Layer-7 application identification information. You can deploy only the Profiler
and the Gateway to detect what hosts communicate with what other hosts during what time periods.
This scenario requires you to collect data from switches and routers. The Steelhead appliances at each location forward
the flows to a single Profiler for analysis and reporting. There is no restriction on the distance between physical
locations for this implementation.
If you have multiple data centers and want to use virtualization, you can use the Profiler and Gateway as tools to show
you which clients are using which servers, and where those clients are located. You can also monitor WAN links to
ensure that you have sufficient bandwidth for busy times by looking at real-time performance information and
historical information.
If your organization has multiple physical locations that are not connected with high amounts of bandwidth (such as a
gigabit MAN), Riverbed recommends that you deploy multiple Gateways throughout the enterprise, with one at each
location you want to monitor. Due to concerns about native flow data being sent across a WAN, placing a Gateway at
locations with sufficient traffic makes sense if the location warrants monitoring in the first place. A small Gateway can
support up to thousands of hosts and offers the same benefits of deduplication, compression, and reliable encrypted
transport of data to the Profiler.
If you have a small site that hosts a data center but has only a few employees, you might want to deploy the Gateway
even if the number of hosts is relatively small, because of the nature of the data. If you want to monitor only data that
is of low value, instead of deploying a Gateway at the site, you can flow data across the WAN.
When deciding where to deploy the Gateway, consider the location of the two sides of the conversations you want to
monitor. If the traffic is between remote clients and servers in a single data center, then you might not need to place
the Gateway at a remote office or send flow data from the remote office to a Gateway in the data center. Because all
critical traffic is in the data center, a single Gateway monitoring all the traffic in, out, and within the data center is
sufficient.
While you do not have to install the Profiler in the data center—the physical location of the Profiler is much less
important than the position of the Gateway—Riverbed recommends that you install the Profiler as close as possible to
the largest sources of flow data.

Riverbed Cascade Product Family Deployment Guide 23

Cascade Appliance Deployment Scenarios Deployment Scenarios

Figure 2-6 shows an example deployment that includes the Profiler and Gateway. All Steelhead appliances and routers
at remote sites, and routers within the data center, send flow data. There are no data flows from smaller sites (not shown
in Figure 2-6). Because these much smaller sites primarily communicate back to the data center, traffic detection is
based upon collection from the data center routers and Steelhead appliance.
Figure 2-6. Example Profiler and Gateway Deployment

Deploying the Enterprise Profiler and Gateway

For very large environments, the Enterprise Profiler provides an expandable solution that can process flows from tens
of thousands of hosts. The Enterprise Profiler provides a robust solution, allowing visibility ranging from a high-level
overview, down to the flow level. The Enterprise Profiler has a base flow rate of 800,000 flows per minute, and you
can add additional modules to support 400,000 flows per minute per module for a maximum supported flow capacity
of 4.8 million flows per minute, after deduplication. Because flows are stored across multiple expansion units, the
amount of disk space for flow storage is increased as capacity increases.
If you have a very large organization, the physical location of the Enterprise Profiler becomes less critical. When you
have enough traffic for this solution, you have multiple locations that are sending data. If there is a concentration of
flow in one area, it makes sense to install the Enterprise Profiler close to this source. In any case, you must locate the
Enterprise Profiler in an appropriate facility with sufficient bandwidth, power, and cooling.
The most important factor in this deployment is to install the Gateways in the correct locations. The Gateway needs to
collect data effectively. Depending on the size of your organization, there are several scenarios that make sense.
You can deploy fewer, larger-capacity Gateways if the number of data collection sites is relatively low and the flow
rate at those locations is very high. Fewer, larger-capacity Gateways is a good choice if your organization has one or
two large data centers to which all clients connect, and where all the collected information is concentrated. If you place
one or more large Gateways in a data center, you can collect all the data necessary. With proper placement, the Gateway
can detect conversations from the clients in the remote office to servers in the data centers.

24 Riverbed Cascade Product Family Deployment Guide

Deployment Scenarios Cascade Appliance Deployment Scenarios

Figure 2-7 shows a Gateway collecting and deduplicating the data flow, then forwarding the flow to the Enterprise
Profiler. Because this deployment does not require network performance and deep packet analysis, you do not need to
install the Shark or the Sensor. This solution enables you to report, analyze, and troubleshoot traffic across the entire
large enterprise network.
Figure 2-7. Example Enterprise Profiler and Gateway Deployment

Deploying the Profiler, Gateway, Shark, and Pilot

This scenario expands a standard Profiler and Gateway deployment to include a Shark and Pilot. The Shark and Pilot
enable you to:
 see network performance data (response time, server delay, and so on) and TCP health information (TCP
retransmission).
 drill-down from the high-level view provided by the Profiler to successively lower-level views until you reach the
packet-level view.
The physical location of the Shark is extremely important. The Shark provides extensive packet capture and analysis
capabilities. You must place the device in a location where it receives the maximum amount of critical traffic.
You must decide what information you want to monitor before you decide where to place the Shark. If you have a
single data center and the traffic to and from that data center is the most critical, you should place the Shark so it can
monitor the critical links or VLANs in the data center. However, if your servers contain critical data and are located in
a special area (outside the traditional corporate data center), then you might want to place the Shark in this area. For
more information about various methods of collecting packet data, see “Packet Collection for Cascade Appliances” on
page 51.
It is not a best practice to use a Shark to monitor a WAN, unless you want packet-level visibility into the WAN link.
The routers or Steelhead appliances on either end of the link are likely to provide flow data that includes link use
information down to the level of the individual conversations.
For performance reasons, you might need to limit the amount of data sent to a Shark. With a limit of eight 1-Gbps
interfaces or two 10-Gbps interfaces, the Shark has high capacities, but not unlimited packet-capture capacities. A data
center at a medium-sized organization can easily exceed 20 Gb of traffic per second.
The Shark provides the following ways to monitor the appropriate traffic:

Riverbed Cascade Product Family Deployment Guide 25

Cascade Appliance Deployment Scenarios Deployment Scenarios

 Physical - Collecting packets by using SPAN, port mirroring, and TAP on only the desired links
 Virtual - Selecting only those specific packets that you want to monitor using the built-in filtering capabilities
Figure 2-8 shows an example deployment that includes a Profiler, Gateway, a Shark, and Pilot. Routers and Steelhead
appliances send flows across the network to the Gateway and provide wide visibility into the network. A Shark sits off
of switches in the data center and collects packets for deeper visibility. Flow data from the Shark merges with all other
flow data collected by the Profiler. You can log in to the Profiler to view applications flowing across the entire network.
When troubleshooting, if you need deeper packet-level analysis, the Profiler UI automatically launches Pilot. This
takes you from the Profiler view of flow data directly into Pilot views of packet data.
Figure 2-8. Example Profiler, Gateway, Shark, and Pilot Deployment

Deploying the Profiler, Gateway, Shark, Shark-VE, and Pilot

This deployment expands upon the Profiler deployment described in “Deploying the Profiler, Gateway, Shark, and
Pilot” on page 25. By adding Shark-VE, you obtain visibility into the physical network, as well as visibility into the
relationship between virtual machines hosted on an ESXi platform in the virtual environment.

26 Riverbed Cascade Product Family Deployment Guide

Deployment Scenarios Cascade Appliance Deployment Scenarios

Figure 2-9 shows an example deployment that includes the Profiler, Gateway, the Shark, Shark-VE, and Pilot. Shark-
VEs are deployed on each ESXi platform in which you want visibility. Metrics are sent from within the virtual
environment to the Profiler. Using Pilot, you can also perform packet analysis.
Figure 2-9. Example Profiler, Gateway, Shark, Shark-VE, and Pilot Deployment

Riverbed Cascade Product Family Deployment Guide 27

Cascade Appliance Deployment Scenarios Port and Protocol Dependencies

Port and Protocol Dependencies

To assure that the Cascade appliances communicate, you must open up ports across any existing firewalls. The figures
in this section show which ports and protocols are necessary for different deployment scenarios. The figures also show
external port dependencies for various integrations. For more details about external integrations, see “Additional
Cascade Appliance Integration” on page 73.
This section describes the following:
 “Shark and Pilot Port Dependencies” on page 28
 “Profiler and Gateway Port Dependencies” on page 29
 “Cascade Appliance Full-Solution Port Dependencies” on page 30
 “Cascade Appliance Enterprise Solution Port Dependencies” on page 31

Shark and Pilot Port Dependencies

Figure 2-10 shows which ports and protocols you must have open for communications within a Pilot and a Shark
deployment. External connections are optional, depending on which integrations you use.
Figure 2-10. Pilot and Shark Dependencies

28 Riverbed Cascade Product Family Deployment Guide

Port and Protocol Dependencies Cascade Appliance Deployment Scenarios

Profiler and Gateway Port Dependencies

Figure 2-11 shows which ports and protocols you must have open for communication within a Profiler and Gateway
deployment. External connections are optional, depending on which integrations you use.
Figure 2-11. Profiler and Port Dependencies

Riverbed Cascade Product Family Deployment Guide 29

Cascade Appliance Deployment Scenarios Port and Protocol Dependencies

Cascade Appliance Full-Solution Port Dependencies

Figure 2-12 shows which ports and protocols you must have open for communications within a Profiler, Gateway,
Sensor, and Shark deployment. External connections are optional, depending on which integrations you use.
Figure 2-12. Full Solution Port Dependencies

30 Riverbed Cascade Product Family Deployment Guide

Profiler and Express Flow Storage Cascade Appliance Deployment Scenarios

Cascade Appliance Enterprise Solution Port Dependencies

Figure 2-13 shows which ports and protocols you must have for communications within an Enterprise Profiler,
Gateway, Sensor, and Shark deployment. Many external connections are optional, depending upon integrations you
use. You must have all components that compose the Enterprise Profiler on the same network subnet, using 1-Gb ports
and preferably connected to the same switch.
Figure 2-13. Enterprise Solution Port Dependencies

Profiler and Express Flow Storage

This section describes estimating and sizing your flow rate and flow storage requirements. It includes the following
sections:
 “Types of Flow Storage” on page 31
 “Flow Rate Estimation” on page 32
 “Flow Storage Size Estimation” on page 33

Types of Flow Storage

The Profiler and Express supports the following types of flow storage:
 Local storage - Disk internal to the Cascade appliance.
 Storage area network (SAN) - Near-line storage.
These storage mechanisms operate differently and provide their own benefits and disadvantages.

Riverbed Cascade Product Family Deployment Guide 31

Cascade Appliance Deployment Scenarios Profiler and Express Flow Storage

Local Storage
Every Express, Profiler, and Enterprise cluster includes some amount of local storage. The following table lists the
system storage type.

System Type Physical Disks Flow Storage (Raw)

CAX-360 (L/M/H) 2 3.7 TB

CAP-2260 (L/M/H) 12 11.8 TB

CAP-4260 16 23.6 TB

Systems with a single partition use the partition to store the flow information, and the boot and configuration
information for the Profiler. On systems with two partitions, one partition is used for the system software and the other
is dedicated to flow storage.
The primary advantages to local storage are performance and cost. Because the storage is located within the system, it
is extremely fast when performing reads and writes, limited only by the physical disks and bus connecting the disks to
the core system. The Profiler includes storage and no external storage is required.
The biggest disadvantage of local storage is that is has expansion limits. There is no option for internal disk expansion.

Storage Area Network (SAN)

SAN provides the most robust external solution. SAN is a reliable solution that enables the Profiler to very quickly
access four or more petabytes of storage (limited by the JFS/2 file system used for the device).
When you connect a SAN to the Profiler, the SAN functions as a new disk partition. The Profiler treats the SAN as
another disk that is part of itself, enabling it to easily access the data. You can also offload much of the processing to
the add-in card you use to connect the SAN.
You must use a separate logical unit number (LUN) and fiber connection between each analyzer and the SAN when
you connect a SAN to an Enterprise Profiler cluster. This increases the performance drastically, because each analyzer
is given its own dedicated channel to the SAN, and potentially has its own set of drives on the SAN (depending on the
SAN configuration).

Flow Rate Estimation

The most important aspect of system sizing is the number of flows your system receives every minute. To accurately
estimate the number of flows, you must know if the host accesses internal hosts, or internal and external hosts.
You must take into account the following when estimating flows:
 Estimate your flows per minute.
In general, you can expect a minimum of 20 flows per host per minute across all internal systems. Because
different system send different numbers of flows, it is probable that some hosts send much more then 20 flows per
minute, while some systems send fewer.
Riverbed recommends that you give serious thought when the estimate approaches the limits of a system; for
example, if the estimate indicates 49,000 flows per minute you might want to consider if a system capable of
supporting 50,000 flows is too restrictive.

32 Riverbed Cascade Product Family Deployment Guide

Flow Redundancy with Cascade Appliances Cascade Appliance Deployment Scenarios

 Consider how much of the traffic the Profiler detects.

A network with 10,000 hosts, only 25% of which forwards traffic to the Profiler, has a very different requirement
then a network with 5,000 hosts, that all forward traffic. The first example has an estimated 50,000 flows per
minute, and the second example has an estimated 100,000 flows per minute. You must know the total number of
hosts on a network and the number of hosts that have their flows forwarded.
 Estimate the current traffic level and what is expected to occur in the near future.
Sizing a system for 2,000 hosts when an additional 500 seats will be added over the next year (or conversely 500
seats will be removed) might result in a significant mis-sizing of the system. Additionally, if you are performing a
proof of concept (PoC) on a small portion of a network, ensure that the final solution is sized for the actual
implementation and not the PoC implementation.
 Know which hosts are talking to each other.
Knowing which hosts are talking to each other is the most difficult factor to quantify into an estimate. The 20
flows per host per minute number is based on conversations between only internal hosts. The flow processing
requirement changes drastically if a host is also accessing external systems (for example, Web sites), and those
flows are being processed by the Profiler. An increase of 100%, to 40 flows per minute per host, is not
unreasonable for hosts accessing Internet sites.

Flow Storage Size Estimation

Another primary sizing decision factor is how long you want to store flows. Because each flow currently occupies
approximately 300 bytes, it is theoretically possible, depending on the platform, to store up to tens of billions of
individual unique flows.

Note: This estimation assumes that you have 50% of available storage dedicated to storing flow records. On normal
implementations, half of the storage is dedicated to storing the minute-by-minute flows and half is used to store pre-aggregated data,
called rollups.

If an xx60 Express Profiler receives one flow per minute, it can store flows covering a time span of approximately
6,450 years. On a more realistic network sending 50,000 flows per minute, the same Express can store approximately
47 days worth of flow data.
If you must store flows for 180 days, then the storage included on the Express is inadequate. You must upgrade to a
larger system with additional storage capacity. When you determine the upgrade path to take, ensure that the path you
choose is the most logical approach. The following upgrades are available:
 Standard Profiler (with 11.8 TB of storage built in)
 Enterprise cluster (with 11.8 TB of storage in a base cluster and the ability to add an additional 82.6 TB of
storage)
When you choose the appropriate upgrade, take into account the desired flow storage capacity and the cost of the
system. While the Enterprise Profiler cluster might meet the needs of the network, unless you expect significant growth
in the near future, it is unlikely to be cost effective.

Flow Redundancy with Cascade Appliances

This section describes how to send and store redundant flow using Cascade appliances. Cascade appliances support
the following options to obtain redundant flows:

Riverbed Cascade Product Family Deployment Guide 33

Cascade Appliance Deployment Scenarios Flow Redundancy with Cascade Appliances

 Option 1 - all unique flows go to one Gateway, and the Gateway send the flows two Profilers
 Option 2 - all unique flows go to two Gateways, and each Gateway send the flows to one Profiler
 Option 3 - a blend of option one and option two.
Figure 2-14 shows Option 1: all unique flows go to one Gateway, and the Gateway sends the flows to two Profilers.
Figure 2-14. Option 1: Sending Flows to a Single Gateway

Figure 2-15 shows how you can scale Option 1 to the size of your organization. You can use this design if you run out
of flow volume and have multiple Gateways. Keep in mind that each flow source can send to only a single Gateway.
If the Gateway goes offline, you lose the flows for that Gateway for the period of time it is not operational.
Figure 2-15. Scalable Option 1: Sending Flows to a Single Gateway

34 Riverbed Cascade Product Family Deployment Guide

Flow Redundancy with Cascade Appliances Cascade Appliance Deployment Scenarios

Figure 2-16 shows Option 2: all flows go to two Gateways. Each Gateway sends its flows to only one Profiler. This
design provides Gateway redundancy as well as Profiler redundancy.
Figure 2-16. Option 2: Sending Flows to Two Gateways

Figure 2-17 shows how you can scale Option 2 to the size of your organization. You can scale this design as long as
each Gateway sends to only one Profiler.
Figure 2-17. Scalable Option Two: Sending Flows to Two Gateways

Riverbed Cascade Product Family Deployment Guide 35

Cascade Appliance Deployment Scenarios Flow Redundancy with Cascade Appliances

Figure 2-18 shows Option 3: flow sources that go to only one Gateway are sent to two Profilers, and the flows that go
to both Gateways are sent to a single Profiler. Riverbed recommends this option for cases that absolutely require this
level of complexity. Gateways must have a per-flow-source configuration, which means that you must specify the flow
sources on the Gateway for export.
Figure 2-18. Option 3: Blending Options 1 and 2

The Shark and Sensors can send flow directly to two Profilers. When configuring flow redundancy, you want to plan
how the configuration is synchronized between the two Profilers. You can synchronize your configuration by using the
Cascade backup and restore mechanism. One of the Profilers acts as the primary appliance on which configuration
changes are made. The second Profiler acts as the secondary appliance, which receives a copy of the configuration
from the primary appliance.
For more information about backup and restore, see the Cascade Profiler and Cascade Express User’s Guide.

36 Riverbed Cascade Product Family Deployment Guide

CHAPTER 3 Flow Collection for Cascade
Appliances

This chapter describes the flow collection for Cascade appliances. It includes the following sections:
 “Base Requirements” on page 37
 “Flow Data Fields Consumed by Cascade Appliances” on page 39
 “Flow Type Considerations” on page 40
 “Flow Collection Considerations” on page 40
 “Flow Collection in Virtual Environments” on page 41
 “Validating Flow Collection” on page 41
 “Sample Third-Party Configurations” on page 43
In this chapter, the term flow refers to standard flow types, including NetFlow, sFlow, Netstream, IPFIX, and jFlow.
The term device refers to a router, switch, or another networking device that supports standard flow export.
A Bluecoat Packeteer shaper supports flow-detail-records v2 (FDR) for application identifier collection. If the device
is a Steelhead appliance, Riverbed recommends that you use CascadeFlow instead of NetFlow to collect retransmit and
response time information. If the device supports multiple types of flow formats, NetFlow is preferable to sFlow.
For details about collecting data from Sharks and Sensors, see “Packet Collection for Cascade Appliances” on page 51.
For details about collecting CascadeFlow from Steelhead appliances, see “Cascade Appliances and Steelhead
Appliance Integration” on page 63.

Base Requirements
You must meet the following requirements to set up your router:
 Configure devices that support NetFlow for NetFlow v1, v5, v7, or v9 with no aggregation and no sampling.
Riverbed recommends that you use v5 or v9.
 Configure devices that support sFlow for sFlow versions v2, v4, or v5 with the lowest possible sampling rate.
Riverbed recommends that you use v5.
 Configure devices to export flow to the Express or Gateway management interface. Use default-configured
destination port on the Express or Gateway.
 Synchronize devices with an NTP server. Riverbed recommends that you synchronize devices with the same NTP
server used by the Profiler. For proper operation and reporting, you must synchronize the timestamps on the
network equipment and the Express or Gateway.

Riverbed Cascade Product Family Deployment Guide 37

Flow Collection for Cascade Appliances Base Requirements

 Set the active timeout setting for flows to 60 seconds.

Note: Cisco IOS versions show this timeout in either minutes or seconds.

 Do not adjust the inactive timeout setting from the default setting of 15 seconds. If you must, the time must be
less than 60 seconds.
 When you use NetFlow v5, make sure to add the ip route-cache flow (or appropriate) command for all active
interfaces and VLANs in addition to the ones you actively use. Because NetFlow v5 is typically ingress only, you
can calculate egress only by aggregating ingress from the other interfaces.
 If NetFlow v9 is available, you can selectively control which interfaces to use, and specify both ingress and
egress. Additionally, with NetFlow v9, you can configure the TTL command. This enables ordered-path reporting
in the Profiler. To enable TTL export, enter one of the following commands:
– If using standard NetFlow configuration, the command syntax from global configuration mode is ip flow-
capture ttl.
– If using flexible NetFlow configuration, the command syntax within the flow record template is match ipv4
ttl maximum.
 You must configure SNMP access to any devices sending flow to the Profiler. Standard flow export provides
information with only SNMP ifindex values. By enabling SNMP on these devices, the Express or Gateway can
look up the actual names, descriptions, speeds, and other information about the interfaces. For more information
about SNMP integration, see “SNMP Integration for Flow Sources” on page 73.
Additional requirements and considerations for Cisco equipment include:
 If you use NetFlow on a Cisco 4500 switch, the Supervisor Engine IV or V must be equipped with a NetFlow
Services daughter card and the required software versions.
 If you use NetFlow on a Cisco 6500 switch equipped with both MSFC and SUP1 modules, you must enable
NetFlow on the router level and the switch level. The route once, switch many concept applies to this hardware
configuration. A new flow is first routed by the MSFC module before it is placed in the MLS cache and is
switched. The Profiler must receive NetFlow data from both modules to avoid missing any data. A similar
concept applies to a chassis with SUP2 or 720 modules.
 If you use NetFlow with the Cisco Nexus 7000 series, and you are using NX-OS v4, you must have a minimum
version NX-OS v4.2(8). If you are using NX-OS v5, you must have a minimum version of NX-OS v5.2(1).
Earlier NX-OS releases have incorrect packets-per-second and bits-per-second statistics.
 NetFlow export from the Cisco ASA is does not include standard NetFlow records. Cisco ASA exports NetFlow
Event Log (NSEL) in a NetFlow wrapper. NSEL is event driven, exporting bytes only for the first and last packet
in the flow. There is no concept of an active timer, so you do not get regular updates. NSEL is considered a
replacement for syslog that enables the ASA to scale. The Profiler, Express and Gateway do not currently support
reporting from NSEL.
 Some Cisco devices support NetFlow export for Layer-2 switched traffic in addition to Layer-3 traffic. Generally,
Layer-2 switched NetFlow is available for forwarding ASICs PFC3B, PFC3BXL, or PFC3C. For verification on
whether your hardware or software supports Layer-2 NetFlow, see Cisco documentation. Use the following
command to enable NetFlow export for Layer-2 (if your hardware or software supports Layer-2 traffic export):
Router (config)# ip flow export layer2-switched vlan <vlanlist>

38 Riverbed Cascade Product Family Deployment Guide

Flow Data Fields Consumed by Cascade Appliances Flow Collection for Cascade Appliances

Flow Data Fields Consumed by Cascade Appliances

The following table shows the flow fields that are consumed by Cascade appliances. When you configure third-party
devices to export flow, you must include as many of the following fields as supported by the third-party device.

Field Name Description Flow Versions with Support

Connection initiator Indicates which host initiated the NetFlow v9

conversation. Used for proper client/
server determination.

Source IP address Source IP address of conversation All standard versions

Destination IP address Destination IP address of conversation All standard versions

Inbound SNMP ifindex SNMP ifindex that identifies the All standard versions
interface through which the conversation
is received for the device

Outbound SNMP ifindex SNMP ifindex that identifies the All standard versions
interface through which the conversation
is transmitted out of the device

Packet count Number of packets sent during the All standard versions
conversation

Byte count Number of bytes sent during the All standard versions
conversation

Timestamps Timestamps for the beginning and end of All standard versions
the conversation

Source port Source port being used All standard versions

Destination port Destination port being used All standard versions

TCP flags Set TCP flags NetFlow v5 and v9 on most devices, sFlow
v5

Layer-4 protocol Layer-4 protocol identifier All standard versions

QoS information Type of service (TOS), differentiated All standard versions

services code point (DSCP)

Time-to-live (TTL) Time-to-live value observed when the NetFlow v9

packet traversed the reporting device

Application identifier Layer-7 application identifier NBAR through NetFlow v9 with specific
hardware, also available from Sensors and
Packeteer through FDR records

Retransmitted bytes and TCP transmission counters CascadeFlow from Steelhead

retransmitted packets appliances, Sensors, and Sharks
Network round-trip time Measurement of round-trip time across CascadeFlow from Steelhead
the network appliances, Sensors, and Sharks

Riverbed Cascade Product Family Deployment Guide 39

Flow Collection for Cascade Appliances Flow Type Considerations

Field Name Description Flow Versions with Support

Total response time, server delay, Measurement of response time metrics CascadeFlow from Sensors and Sharks
client delay across the network

VoIP metrics: Voice-over-IP-metrics computed by the Shark v9.5 and later export
Shark
• MOS score
• R-Factor score
• Jitter
• RTP packet loss

Flow Type Considerations

If you use Steelhead appliances, Riverbed recommends that you always use CascadeFlow. For all other devices, if you
have a choice between getting data from devices supporting NetFlow versus devices supporting sFlow, Riverbed
recommends that you use NetFlow. sFlow is typically supported only with sampling and is less accurate than NetFlow.
Riverbed recommends that you use NetFlow v9 with TTL, and that you export the TTL field so that network segment
graphs are available in the Profiler. If version 9 is not available, use NetFlow v5.
Flow coalescing is the process in which flows reported by different devices merge together into a single record. The
Profiler and Express merge records from the Sharks, Sensors, Steelhead appliances, NetFlow sources (routers,
switches, and so on), and IPFIX sources and then tracks all the interfaces the flow crosses, and any changes in traffic
volume.
Flow coalescing does not occur between sFlow sources and other flow types. sFlow records do not merge with other
record types (CascadeFlow, IPFIX, NetFlow, and so on) because this data is typically sampled. You can have data
inconsistencies if sFlow is merged with non-sampled sources. When you want to see sFlow or NetFlow reported
conversations, make sure that your traffic query includes the source device of the flow.

Flow Collection Considerations

The Profiler deduplicates all flows that it receives for the same five pieces of information: source IP address,
destination IP address, source port, destination port, and protocol. The Profiler and Express track every source device
that sent the flow, for up to a total of five in-path devices (up to 10 total interfaces) per flow. You should gather flow
from as many sources as possible, because the Profiler and Express licenses are based on total flow volume, versus
number of devices or number of network interfaces for which flow is received. Do not forget the five-device limit.
If you exceed the five-device limit, the Profiler and Express prioritize the devices kept per flow in the following order:
 Steelhead appliances
 Sharks and Sensors
 All other flow sources
If you exceed the five-device limit within the same category, the device with the lowest IP address is included.

40 Riverbed Cascade Product Family Deployment Guide

Flow Collection in Virtual Environments Flow Collection for Cascade Appliances

Flow Collection in Virtual Environments

The vSphere v5 vSwitch and the Cisco Nexus 1000V support flow export in a virtual environment. Either solution
provides visibility into the virtualized environment, including:
 intrahost virtual machine traffic (virtual machine-to-virtual machine traffic on the same host).
 interhost virtual machine traffic (virtual machine-to-virtual machine traffic on different hosts).
 virtual machine-physical infrastructure traffic.

Validating Flow Collection

You can validate flow collection from the Profiler activity displayed on the Devices/Interface page.

To validate flow data

1. Choose System > Devices/Interface.

2. Select the Devices & Interfaces (Tree) tab to view Cascade appliances that send data the Profiler.

3. Expand the display for each Gateway to view which devices the Gateway is receiving flow data from.

4. Further expand the display for each flow-sending device listed in the tree to see specific interfaces.

5. Hover your mouse over the name of each interface to see details about the interface.

Riverbed Cascade Product Family Deployment Guide 41

Flow Collection for Cascade Appliances Validating Flow Collection

Figure 3-1 shows an expanded gateway-DataCenter, with further expansion of WAN-RTR-Hartford. The pop-up
window shows the details about the interface WAN-RTR-Hartford:wan, including inbound and outbound speed and
utilization.
Figure 3-1. Profiler Device/Interfaces Page

When you validate flow collection on the Devices/Interface page you might have the following display issues:
 If you do not see interface names and speeds it is likely because you have not configured SNMP polling to the
devices.
For details about how to configure SNMP polling for the flow-sending devices, see the Cascade Profiler and
Cascade Express User’s Guide.
 If you only see outbound traffic, it can be that you are not exporting traffic for that particular interface.
All interfaces for which a flow record is received are in the list, even though you might not be exporting flow for
that interface. You might see the data if the device is exporting data for the opposing interface and the flow
outbound interface is the one in question. For example, you are exporting flows for Interface 1, but the flow is
destined for Interface 2. When the flow is received on Interface 1, the record indicates that it is destined for
Interface 2. Therefore, Interface 2 is in the list, even though you might not be exporting for Interface 2.

42 Riverbed Cascade Product Family Deployment Guide

Sample Third-Party Configurations Flow Collection for Cascade Appliances

Sample Third-Party Configurations

This section has several third-party configuration examples that show you how to enable NetFlow export to the
Express or Profiler. Refer to vendor documentation specific to your device and version software. Commands complete
various actions, depending upon device software version.
This section includes the following:
 “Configuring Cisco 6500 Series Switches Running Native Cisco IOS” on page 43
 “Configuring Cisco 6500 Series Switches in Hybrid Mode” on page 44
 “Configuring Cisco 7500 Series Router” on page 45
 “Configuring Cisco 7600 Series Router” on page 45
 “Configuring Cisco 3560 and 3750 Flexible NetFlow” on page 46
 “Configuring the Cisco Nexus 7000 Flexible NetFlow” on page 46
 “Configuring NetFlow Export for Cisco Nexus 1000V” on page 47
 “Configuring IPFIX for Avaya (Nortel) 8300 and 8600” on page 48
 “Configuring sFlow for HP Procurve 3500, 5400, and 6200” on page 49

Configuring Cisco 6500 Series Switches Running Native Cisco IOS

The following example uses the native Cisco IOS CLI to configure the SUP and MSFC modules of a 6500 series
switch. The following commands generally work with Cisco IOS v12.2 or later, except where specified. For further
information, refer to the documentation for your version of Cisco IOS.

To configure the SUP and MSFC modules of a 6500 series switch

1. At the switch level (SUP2), enter the following commands to turn on NetFlow and set version, flow mask, and
timing:
Router(config)# mls netflow
Router(config)# mls nde sender version 5
Router(config)# mls flow ip interface-full
Router(config)# mls nde interface
Router(config)# mls aging normal 32
Router(config)# mls aging long 64

2. At the routing module (MSFC), enter the following commands to set the device source interface, version,
destination, and timeouts:
Router(config)# ip flow-export source loopback 0
Router(config)# ip flow-export version 9
Router(config)# ip flow-export destination <cascade-gateway-or-express_ip> <udp_port_number>
Router(config)# ip flow-cache timeout inactive 15 (this might be the default depending upon
code version)
Router(config)# ip flow-cache timeout active 1

Note: If you are running Cisco IOS v12.2(18) or later, use NetFlow v9. If NetFlow v9 is not available, use NetFlow v5.

Riverbed Cascade Product Family Deployment Guide 43

Flow Collection for Cascade Appliances Sample Third-Party Configurations

If you are running Cisco IOS v12.3(14) or later and are exporting NetFlow v9, you can include export of the TTL,
enabling the Profiler and Express to show network segment diagrams:
Router(config)# ip flow-capture ttl

If you are running Cisco IOS v12.3(14) or later, running NetFlow v9, and have hardware that supports export of
NBAR Layer-7 information, include the following command:
Router(config)# ip flow-capture nbar

Next, to enable NetFlow on your interfaces, enter the following commands, where applicable, for each interface
or interface grouping where you require NetFlow accounting (three types of interfaces):
interface <type> <slot>/<port>

For example:
Router(config)# interface fastethernet 0/1
Router(config-if)# ip route-cache flow

or
interface vlan <vlan_id>

For example:
Router(config)# interface vlan 3
Router(config-if)# ip route-cache flow

or
interface port-channel <channel_id>

For example:
Router(config)# interface port-channel 3
Router(config-if)# ip route-cache flow

3. Optionally, if you want to export Layer-2 switched flows (and your switch supports Layer-2 NetFlow export), enter
the following commands for the set of VLANs where you want the Layer-2 flows exported:
Router (config)# ip flow export layer2-switched vlan <vlanlist>

Configuring Cisco 6500 Series Switches in Hybrid Mode

The following example configures the SUP and MSFC modules of a Cisco 6500 series switch running in the hybrid
mode.

To configure the SUP and MSFC modules of a 6500 series switch in hybrid mode

1. At the switch level (SUP), enter the following commands to enable NetFlow data export (NDE) and to set
destination of flow, timers, and full flow:
Router(config)# set mls nde enable
Router(config)# set mls nde enable <gateway-or-express_ip> <udp_port_number>
Router(config)# set mls agingtime 16
Router(config)# set mls agingtime fast 32 0
Router(config)# set mls agingtime long-duration 64
Router(config)# set mls flow full

2. At the routing module (MSFC), enter the following commands to configure NDE and set the destination of flow:
Router(config)# ip flow-export <ip_address> < udp_port> <version>

44 Riverbed Cascade Product Family Deployment Guide

Sample Third-Party Configurations Flow Collection for Cascade Appliances

3. At the interface level, enter the following commands to enable NetFlow on each interface on which you want to
collect statistics and set timers:
Router(config)# interface <type> <slot>/<port-adapter>

For example:
Router(config)# interface fastethernet 0/1
Router(config-if)# ip route-cache flow
Router(config)# ip flow-cache timeout active 1
Router(config)# ip flow-cache timeout inactive 15

Configuring Cisco 7500 Series Router

The following example uses the Cisco IOS CLI to configure a Cisco 7500 series router.

To configure a Cisco 7500 series router using the Cisco IOS CLI

1. Enter the following commands to configure NDE (NetFlow Data Export):

Router# confg t
Router(config) # ip flow-export <cascade-gateway-or-express_ip> <udp_port_number> <version>

2. Enter the following commands to enable NetFlow at the interface level on each interface on which you want to
collect statistics:
Router(config) # interface <type> <slot>/<port-adapter>

For example:
Router(config)# interface fastethernet 0/1

For 7500:
Router(config-if)# ip route-cache flow

3. Enter the following commands to set the NetFlow timers:

Router(config)# ip flow-cache timeout active 1
Router(config)# ip flow-cache timeout inactive 15

Configuring Cisco 7600 Series Router

The following example uses the Cisco IOS CLI to configure a Cisco 7600 series router.

To configure a Cisco 7600 series router using the Cisco IOS CLI

1. Enter the following commands to configure NDE (NetFlow Data Export):

Router(config)# ip flow-export <cascade-gateway-or-express_ip> <udp_port_number>
Router(config)# ip flow-export <version>
Router(config)# mls nde sender <version>

2. Enter the following commands to enable NetFlow at the interface level on each interface on which you want to
collect statistics:
interface <type> <slot>/<port-adapter>

For example:
Router(config)# interface fastethernet 0/1
Router(config-if)# ip flow ingress

Riverbed Cascade Product Family Deployment Guide 45

Flow Collection for Cascade Appliances Sample Third-Party Configurations

3. Enter the following commands to set the NetFlow timers:

Router(config)# ip flow-cache timeout active 1
Router(config)# ip flow-cache timeout inactive 15

Configuring Cisco 3560 and 3750 Flexible NetFlow

The following example shows an example Flexible NetFlow configuration for the Cisco 3750 and 3560 series switches
with NetFlow service module C3KX-SM-10G.

To configure Flexible NetFlow for a Cisco 3750 or 3560 switch

1. Enter the following commands to create the flow record:

Switch# flow record Cascade-record
Switch# match ipv4 tos
Switch# match ipv4 protocol
Switch# match ipv4 source address
Switch# match ipv4 destination address
Switch# match ipv4 ttl
Switch# match transport source-port
Switch# match transport destination-port
Switch# collect counter bytes
Switch# collect counter packets
Switch# collect timestamp sys-uptime first
Switch# collect timestamp sys-uptime last

2. Enter the following commands to create the flow exporter and monitor:
Switch# flow exporter Cascade
Switch# destination <ip address of Gateway or Express>
Switch# transport udp <udp port of Gateway or Express>
Switch# flow monitor Cascade
Switch# record Cascade-record
Switch# exporter Cascade
Switch# cache timeout active 60
Switch# cache timeout inactive 60

3. Enter the following commands to enable export on a specific port:

Switch# interface TenGigabitEthernet1/1/1
Switch# ip flow monitor Cascade input
Switch# ip flow monitor Cascade output

Configuring the Cisco Nexus 7000 Flexible NetFlow

The following example uses Cisco Nexus OS v5.2.1 to configure NetFlow export. You must complete the set of
commands in Step 5 for each Layer-3 interface.

To configure a NetFlow export using a Cisco Nexus 7000 Flexible NetFlow

1. Enter the following commands to configure a record to include all necessary fields for the Profiler, Express, or
Gateway:
Switch# config t
Switch(config)# flow record Cascade-record
Switch(config-flow-record)# match interface input
Switch(config-flow-record)# match interface output
Switch(config-flow-record)# match ipv4 source address
Switch(config-flow-record)# match ipv4 destination address

46 Riverbed Cascade Product Family Deployment Guide

Sample Third-Party Configurations Flow Collection for Cascade Appliances

Switch(config-flow-record)# match protocol

Switch(config-flow-record)# match transport source-port
Switch(config-flow-record)# match transport destination-port
Switch(config-flow-record)# collect flow direction
Switch(config-flow-record)# collect ipv4 tos
Switch(config-flow-record)# collect ipv4 ttl max
Switch(config-flow-record)# collect transport tcp flags
Switch(config-flow-record)# collect counter bytes
Switch(config-flow-record)# collect counter packets
Switch(config-flow-record)# collect routing next-hop address ipv4
Switch(config-flow-record)# collect timestamp sys-uptime first
Switch(config-flow-record)# collect timestamp sys-uptime last

2. At the global level, enter the following commands to configure required timeout settings:
Switch# conf t
Switch(config)# feature netflow
Switch(config-netflow)# flow timeout active 60
Switch(config-netflow)# flow timeout inactive 15
Switch(config-netflow)# flow timeout session

3. Enter the following commands to configure NetFlow export:

Switch# config t
Switch(config)# flow exporter cascade-export
Switch(config-flow-exporter)# destination <cascade-gateway-or-express_ip>
Switch(config-flow-exporter)# source ethernet 2/1
Switch(config-flow-exporter)# transport udp 2055
!--- Listening port configured on Gateway
Switch(config-flow-exporter)# version 9

4. Enter the following commands to configure flow monitor:

Switch# config t
Switch(config)# flow monitor cascade-monitor
Switch(config-flow-monitor)# record netflow ipv4 cascade-record
Switch(config-flow-monitor)# exporter cascade-export

5. Enter the following commands to apply a flow monitor to a VLAN or interface (one time for each Layer-3
interface):
Switch# config t
Switch(config)# vlan 30
Switch(config-vlan)# ip flow monitor cascade-monitor input

Configuring NetFlow Export for Cisco Nexus 1000V

Configuring NetFlow export of the Cisco 1000V is similar to the physical Nexus switches running NX-OS (for
example, Cisco Nexus 7000), with some variation in commands. The primary difference is that the Riverbed
recommended configuration parameters are for the Cisco Nexus 7000 TTL export. Use the template shown in this
example (TTL export is not an option on the Cisco Nexus 1000V).

To configure NetFlow export for a Cisco Nexus 1000V

1. Enter the following commands to configure NetFlow Exporter and timing parameters:
n1000v# config t
n1000v(config)# flow exporter Cascade-export
n1000v(config-flow-exporter)# destination <cascade-gateway-or-express_ip>
n1000v(config-flow-exporter)# source mgmt 0
n1000v(config-flow-exporter)# transport udp 2055
!--- Listening port configured on Gateway

Riverbed Cascade Product Family Deployment Guide 47

Flow Collection for Cascade Appliances Sample Third-Party Configurations

n1000v(config-flow-exporter)# version 9
n1000v(config-flow-exporter-version-9)# option exporter-stats timeout 60
n1000v(config-flow-exporter-version-9)# template data timeout 1200
n1000v(config-flow-exporter-version-9)# option interface-table timeout 3600

2. Enter the following commands to configure flow monitor:

n1000v(config)# flow monitor Cascade-monitor
n1000v(config-flow-monitor)# record netflow-original
n1000v(config-flow-monitor)# exporter Cascade-export
n1000v(config-flow-monitor)# timeout active 60
n1000v(config-flow-monitor)# timeout inactive 15

3. Enter the following commands to apply the flow monitor to either each virtual interface or each port profile:
 For an interface:
n1000v(config)# interface veth 2
n1000v(config-if)# ip flow monitor Cascade-monitor input
n1000v(config-if)# ip flow monitor Cascade-monitor output

 For a port profile (the port profile must be configured with other appropriate parameters and inherited on the
appropriate interfaces or port groups):
n1000v(config)# port-profile type vethernet <Profile-Name>
n1000v(config-port-prof)# ip flow monitor Cascade-monitor input
n1000v(config-port-prof)# ip flow monitor Cascade-monitor output

Configuring IPFIX for Avaya (Nortel) 8300 and 8600

The following example uses Nortel ERS 8300 and ERS 8600 to configure flow export. You use similar commands to
configure other Nortel routers.

To configure IPFIX for Avaya (Nortel) 8300 and 8600

1. Enter the following command to enable IPFIX globally:

ERS# config ip ipfix state enable

2. Enter the following command to enable IPFIX at a port level, for each port where you want each export:
ERS# config ip ipfix port 5/2, 5/3, 5/4, 5/5, 5/6 all-traffic enable

3. Enter the following commands to set the timing parameters for the Cascade appliance compatibility (active timeout
is in minutes, export interval in seconds):
ERS# config ip ipfix active-timeout 1
ERS# config ip ipfix aging-interval 15
ERS# config ip ipfix export-interval 60

Depending on your router and software version, you might need to specify slot numbers in the previous
commands. The following example shows the commands with slot numbers:
ERS# config ip ipfix slot 5 active-timeout 1
ERS# config ip ipfix slot 5 aging-interval 15
ERS# config ip ipfix slot 5 export-interval 60

4. Enter the following commands to enable export and to export to the Express and Gateway:
ERS# config ip ipfix exporter-state enable
ERS# config ip ipfix collector add <gateway-or-express-ip-address> dest-port <gateway-or-
express-ip-address-listening-port> enable true

48 Riverbed Cascade Product Family Deployment Guide

Sample Third-Party Configurations Flow Collection for Cascade Appliances

or
ERS# config ip ipfix slot 5 exporter-state enable
ERS# config ip ipfix slot 5 collector add <gateway-or-express-ip-address> dest-port <gateway-
or-express -listening-port> enable true

Configuring sFlow for HP Procurve 3500, 5400, and 6200

The following example uses Procurve 3500, 5400, and 6200 to configure flow export. You use similar commands to
configure other HP Procurve devices.

To configure sFlow for HP Procurve 3500, 5400, and 6200

1. Enter configuration mode to configure the Express or Gateway as a flow destination:

ProCurve# configure
ProCurve(config)# sflow 1 destination <gateway-or-express-ip-address> <gateway-or-express -
listening-port>

In this example, 1 is the sflow instance. If this instance ID is already in use, then enter either 2 or 3 in the previous
and the following commands.

2. Enter the following command to activate sampling:

ProCurve(config)# sflow 1 sampling all 500

The example shows a sampling rate of one out of every 500 packets. Riverbed recommends that you set the
sampling rate to the lowest value recommended by HP; the lowest value recommended depends on device and
link speed. In the example, all results in using this sampling rate for all ports.

3. Enter the following commands to activate polling:

ProCurve(config)# sflow 1 polling all 60

In the example, all results in using this polling rate for all ports, and 60 indicates the polling and export interval.

4. Enter the following command to save the configuration:

ProCurve(config)# write mem

Riverbed Cascade Product Family Deployment Guide 49

Flow Collection for Cascade Appliances Sample Third-Party Configurations

50 Riverbed Cascade Product Family Deployment Guide

CHAPTER 4 Packet Collection for Cascade
Appliances

This chapter describes the different methods for Cascade appliance packet capture. You use packet capture to monitor
traffic monitoring and analyze packets. This chapter includes the following sections:
 “Cascade Appliances for Packet Collection” on page 51
 “SPAN and Port Mirroring” on page 52
 “Network Tap Instrumentation” on page 57
 “VACL Configuration Examples” on page 58
 “Shark Passthru” on page 60
 “Packet Deduplication” on page 60
 “Snaplen” on page 61

Cascade Appliances for Packet Collection

Cascade appliances collect packets using one of the following components:
 The Shark - Traffic capture and monitoring for high-rate packet capture and analysis
 The Sensor - Traffic monitoring with Layer-7 application identification
 The Express 360 with built-in Sensor capability - Port traffic monitoring with Layer-7 application
identification
 The Express 460 with built-in Shark capability - Port traffic monitoring with packet capture
You can forward flows from these appliances to the Profiler based upon the packets you collect. In addition to standard
NetFlow-type fields, these components send TCP health information and TCP performance information. For more
information about these components, see “Cascade Product Family Overview” on page 5 and “Cascade Appliance
Deployment Scenarios” on page 9.

Riverbed Cascade Product Family Deployment Guide 51

Packet Collection for Cascade Appliances SPAN and Port Mirroring

SPAN and Port Mirroring

Port mirroring is the most popular method for collecting packets. Port mirroring is commonly referred to as switched
port analyzer (SPAN). You can use the terms SPAN and port mirroring interchangeably. When you configure port
mirroring, depending upon your hardware, you can mirror:
 select ports or select VLANs from a device to a monitoring port.
 all ports or all VLANs from the device to a monitoring port.
You can also, depending upon your hardware, configure capture on ingress, egress, or both, on the interface or VLAN
you are monitoring.
Figure 4-1 shows a monitoring configuration in which you detect traffic among all local servers. By monitoring an
uplink port or VLAN, in addition to the local ports or VLANs, you can also detect traffic between all external hosts to
the local hosts. The Shark and Sensor have two or more monitoring ports that enable you to duplicate this configuration
multiple times using the same Shark or Sensor.
Figure 4-1. SPAN Connectivity

Best practices for port mirroring:

 For most monitoring and troubleshooting, you must collect both sides of the conversation. This means that if you
are capturing only one port, you must mirror both directions—ingress and egress. If you are monitoring all ports
or all communicating VLANs, you can capture ingress only. Capturing ingress and egress on all ports or all
VLANs is redundant, and the duplicate traffic is deduplicated on the Sensor or at the Profiler level.
 When you set up port mirroring, you must follow best practices according to your switch vendor. Because many
architectures use non-blocking methods that drop overages if you overrun a port mirror (for example, by sending
multiple gigabits per second worth of packets from a single gigabit port), depending on the switch you use, there
might be an adverse effect on traffic or switch performance.
 For large applications across numerous switches, you can use third-party port monitor aggregators for flexible
configurations. Vendors that supply port monitor aggregators include Anue Systems, NetOptics, Gigamon,
cPacket Networks, and VSS Monitoring.

52 Riverbed Cascade Product Family Deployment Guide

SPAN and Port Mirroring Packet Collection for Cascade Appliances

 Many switches have a limit on the number of monitoring ports that you can configure. This limit is often two
monitoring ports. If the limit is a problem in your environment, you can add a tap to an existing monitoring port
(essentially making a copy of the traffic already being monitored by another device), or you can use VLAN
access control lists (VACLs) to configure what amounts to an additional SPAN port, provided that your
equipment supports VACLs. For more information, see the following sections in this chapter.

Remote SPAN and Encapsulated Remote SPAN

This section describes the following SPAN variations:
 “RSPAN” on page 53
 “ERSPAN” on page 54
Riverbed recommends Remote SPAN (RSPAN) and Encapsulated Remote SPAN (ERSPAN) techniques in special
circumstances. With some routers and switches, an adverse impact on performance can occur with configuration of
RSPAN or ERSPAN. Read the appropriate documentation and release notes for the hardware and software of your
switch or router.

RSPAN
RSPAN enables an extension of a SPAN over the network to another switch on a Layer-2 nonroutable RSPAN VLAN.
You can use RSPAN when you have one or more access switches, and you want to configure a SPAN to a single Sensor,
Shark, or Express monitoring port at a distribution switch. To ensure that network traffic is not impeded, dedicate a
trunk port to carry the traffic from the access switches to the distribution switch.
Figure 4-2 shows a monitoring configuration in which you detect traffic to and from local servers on two different
switches. The monitoring port is on an upstream switch. The Shark and Sensor have two or more monitoring ports that
enable you to duplicate this configuration multiple times using the same Shark or Sensor.
Figure 4-2. RSPAN Connectivity

Riverbed Cascade Product Family Deployment Guide 53

Packet Collection for Cascade Appliances SPAN and Port Mirroring

ERSPAN
ERSPAN enables an extension of a SPAN over the network to another switch through a routed GRE-encapsulated
tunnel. You can use ERSPAN when a Sensor, Shark, or Express is monitoring from a distant switch. In this case, you
must have adequate bandwidth over the routed path that carries the mirrored traffic so that mirroring does not adversely
affect production network traffic.
Figure 4-3 shows a monitoring configuration that enables you to detect traffic to and from local servers on two different
switches when the monitoring port is on an upstream switch over a routed network. The Shark and Sensor have two or
more monitoring ports that enable you to duplicate this configuration multiple times using the same Shark or Sensor.
Figure 4-3. ERSPAN Connectivity

You must use ERSPAN in an virtualized environment that uses the Cisco Nexus 1000V. The Cisco Nexus 1000V
mirrors traffic sent between virtual machines by sending ERSPAN to an external Cisco Catalyst 6500 switch.

Sample Port Mirror Configurations

This section includes the following SPAN port configuration examples:
 “Cisco Catalyst 6500 SPAN” on page 54
 “Cisco Nexus 5000 SPAN” on page 55
 “Cisco Nexus 1000V ERSPAN to Cisco Catalyst 6500” on page 56
SPAN port configurations vary depending upon device and software version. For more information, see the
documentation that came with your device.
For details about Cisco switch configuration examples, go to:
http://www.cisco.com/en/US/products/hw/switches/ps708/products_tech_note09186a008015c612.shtml

Cisco Catalyst 6500 SPAN

The following steps describe how to configure a SPAN for all traffic for VLANs 1 through 100 using a Cisco Catalyst
6500 SPAN. You must only capture ingress on the VLANs to monitor all traffic.

54 Riverbed Cascade Product Family Deployment Guide

SPAN and Port Mirroring Packet Collection for Cascade Appliances

To configure a SPAN for all traffic for VLANs 1 through 100 using a Cisco Catalyst 6500 SPAN

1. From the switch CLI, enter configuration mode to set up a monitor session and configure the source traffic you
want to monitor:
Switch# conf t
Switch (config)# monitor session 1 source vlan 1-100 rx

2. Enter the following command to configure the destination port where the Sensor, Shark, or Express monitoring
port is connected:
Switch (config)# monitor session 1 destination gigabitethernet 4/3

The following example shows capturing all traffic to and from sources on the downstream port 5/1 and sending the
collected traffic to port 5/3.

To configure a SPAN for all traffic to and from a downstream switch on port 5/1 using a Cisco
Catalyst 6500 SPAN

1. From the switch CLI, enter configuration mode to set up a monitor session and configure the source traffic you
want to monitor:
Switch# conf t
Switch (config)# monitor session 1 source gigabitethernet 5/1 both

2. Enter the following command to configure the destination port where the Sensor, Shark, or Express monitoring
port is connected:
Switch (config)# monitor session 1 destination gigabitethernet 5/3

Cisco Nexus 5000 SPAN

The following steps illustrate how to configure a SPAN for all traffic for VLANs 1 through 100. The Cisco Nexus 5000
collects all traffic ingress to the VLANs. The example shows that using a SPAN on ingress works as well as VLANs
1 through 100.

To configure a SPAN for all traffic for VLANs 1 through 100 using a Cisco Nexus 5000 SPAN

1. From the switch CLI, enter configuration mode to set up a monitor session:
Switch# conf t
Switch (config)# monitor session 1
Switch (config-monitor)# exit
Switch (config)#

2. Enter the following commands to configure the destination port to which the Sensor, Shark, or Express monitoring
port is connected (first set the port as a monitoring port, and next place it into the created session):
Switch (config)# interface ethernet 5/4
Switch (config-if)# switchport monitor
Switch (config-if)# exit
Switch (config-if)# monitor session 1
Switch (config-monitor)# destination interface ethernet 5/4

3. While still in configuration mode, enter the following command to configure the source traffic you want to
monitor:
Switch (config-monitor)# source vlan 1-100

Riverbed Cascade Product Family Deployment Guide 55

Packet Collection for Cascade Appliances SPAN and Port Mirroring

The following example shows all traffic SPANing to and from a downstream switch on port 5/2. You want to make
sure that you are capturing all traffic to and from sources on the downstream port. Capture traffic in both directions on
the port (default if unspecified).

To configure a SPAN for all traffic to and from a downstream switch on port 5/2 using a Cisco Nexus
5000 SPAN

1. From the switch CLI, enter configuration mode to set up a monitor session:
Switch# conf t
Switch (config)# monitor session 1
Switch (config-monitor)# exit
Switch (config)#

2. Enter the following commands to configure the destination port to which the Sensor, Shark, or Express monitoring
port is connected (first, mark the port as a monitoring port, and next place it into the created session):
Switch (config)# interface ethernet 5/5
Switch (config-if)# switchport monitor
Switch (config-if)# exit
Switch (config-if)# monitor session 1
Switch (config-monitor)# destination interface ethernet 5/5

3. While still in configuration mode, enter the following command to configure the source traffic you want to
monitor:
Switch (config-monitor)# source interface ethernet 5/2 both

Cisco Nexus 1000V ERSPAN to Cisco Catalyst 6500

The following steps illustrate how to configure an ERSPAN for Cisco's virtual switch, Cisco Nexus 1000V, to a
Catalyst 6500. You must configure both the Cisco Nexus 1000V and the Catalyst 6500. This example shows data
collection from VLANs 1 through 10 on the Cisco Nexus 1000V switch. The example uses a ERSPAN identifier of
100 for the configuration.

To configure the Cisco Nexus 1000V to collect data on VLANs 1 through 10

1. From the switch CLI, enter configuration mode to set up a monitor session and provide a description:
Switch# conf t
Switch (config)# monitor session 1 type erspan-source
Switch (config-monitor)# desc CascadeErspanSource

2. Enter the following command to select which ports or VLANs to monitor:

switch (config-monitor)# Source vlan 1-10

3. Enter the following commands to provide the destination IP address of the 6500 switch (use any reachable IP
address on the 6500) and an identifier:
Switch (config-monitor)# Destination ip [6500 IP address]
Switch (config-monitor)# erspan-id 100
Switch (config-monitor)# no shut

To configure the Cisco Catalyst 6500 to ERSPAN

1. From the switch CLI, enter configuration mode to set up a monitor session, and provide a description:
Switch# conf t
Switch (config)# monitor session 1 type erspan-destination

56 Riverbed Cascade Product Family Deployment Guide

Network Tap Instrumentation Packet Collection for Cascade Appliances

Switch (config-monitor)# desc CascadeErspanDest

2. Enter the following commands to configure the specific destination interface, identifier, and receiving IP address:
Switch (config-monitor)# destination interface gix/y/z
Switch (config-monitor)# source
Switch (config-monitor)# erspan-id 100
Switch (config-monitor)# ip address [6500 IP address]
Switch (config-monitor)# no shut

Network Tap Instrumentation

You can insert passive network taps as another method for collecting packet data. This device sits inline on a physical
link and makes a copy of all traffic passing through to a monitoring device. You can classify taps as follows:
 Basic taps - Make a copy of the signal on the wire to a secondary port for monitoring. When you use a passive
tap, you must use two monitoring ports on the Shark or the Sensor for one link that you monitor, because the tap
uses a separate port to copy the traffic in each direction.
Figure 4-4 shows a tap on a link between Device A and Device B. The tap copies traffic in the direction from
Device A to Device B on one port, and the direction from Device B back to Device A on a second port.
Figure 4-4. Basic Tap Connectivity

 Regeneration taps - Send the same traffic for the same monitored link to multiple devices. These taps are useful
if you want to send traffic from link to both the Sensor, Shark, or Express and another device, for example, an
IDS.
 Aggregation taps - Enable you to aggregate both directions of traffic on a monitored link through a single port so
that you need only a single port on the Shark or Sensor for a link you want to monitor. If you use this method, you
can potentially miss some packets if the full-duplex link is running close to line rate in both directions.
Some aggregation taps can regenerate and send traffic from a monitored link to multiple monitoring devices
(sometimes referred to as port aggregation). Some aggregation taps can combine multiple monitored links to one
or more monitoring devices, sometimes referred to as link aggregation.

Riverbed Cascade Product Family Deployment Guide 57

Packet Collection for Cascade Appliances VACL Configuration Examples

 Advanced/Intelligent taps - Many of the same vendors that offer intelligent SPAN or port-mirror solutions also
offer solutions you can use for taps.
Best practices for tap deployment:
 Ensure that you understand which type of tap you are using, keeping in mind that basic taps require two
monitoring ports per monitored link.
 You can use taps on existing SPAN and port-monitoring ports. Using taps is useful if there are no longer SPAN
and monitoring ports available on the switch you want to monitor.
 You can chain taps. For example, if you already have a tap deployed to a monitoring device such as an IDS, you
can tap into the feed to the IDS for monitoring with the Sensor, Shark, or Express.

VACL Configuration Examples

You can use a VACLs, which is used to mirror ports, for cases when your switch supports only a limited number of in-
use SPAN ports. This section includes the following examples:
 “VACL Port Mirroring Configuration on Cisco 6500 Running CatOS” on page 58
 “VACL Port Mirroring Configuration on Cisco Catalyst 6500 Running Cisco IOS” on page 59
VACL configuration varies based upon device and software version number. For details, see the documentation
specific to your device and software version.

VACL Port Mirroring Configuration on Cisco 6500 Running CatOS

The following example shows VACL port mirroring configuration for a Cisco Catalyst 6500 running CatOs. Apply the
configuration to the switch only; there is no MSFC component. Connect the capture port where the Shark or the Sensor
are monitoring interfaces to trunk ports.

To configure VACL port mirroring on a Cisco Catalyst 6500 running CatOs

1. Enter the following commands to create the VACL and specify it as a capture VACL:
> set security acl ip CascadeMonitor permit any any capture
> show security acl info CascadeMonitor editbuffer

2. Enter the following commands to commit the VACL to NVRAM:

> commit security acl CascadeMonitor|all

58 Riverbed Cascade Product Family Deployment Guide

VACL Configuration Examples Packet Collection for Cascade Appliances

3. Enter the following commands to map the VACL to all VLANs you want to monitor:
> set security acl map CascadeMonitor vlan1,vlan2, vlan3

4. Enter the following commands to specify the capture port on which you have connected the Sensor, Shark, or
Express monitoring port (enables for normal switching and creates a copy on the capture port):
> set security acl capture-ports 5/3
> show security acl capture-ports

VACL Port Mirroring Configuration on Cisco Catalyst 6500 Running

Cisco IOS
The following example shows VACL port mirroring configuration for Cisco Catalyst 6500 running Cisco IOS. Apply
the configuration to the switch only; there is no MSFC component.

To configure VACL port mirroring on a Cisco Catalyst 6500 running Cisco IOS

1. From the switch CLI, enter the following commands to create the VACL:
Switch# conf t
Switch(config)# ip access-list CascadeMon
Switch(config-access-list)# permit ip any any
Switch(config-access-list)# exit
Switch(config)#

2. Enter the following commands to configure the assigned capture or monitoring port as a trunk port (Interface 5/3):
Switch(config)# interface GE5/3
Switch(config-if)# no ip address
Switch(config-if)# switchport
Switch(config-if)# switchport mode trunk
Switch(config-if)# switchport trunk encapsulation dot1q

3. Enter the following commands to define the VLAN access map:

Switch(conf)# vlan access-map map_name seq#
Switch (conf-map_name)#

4. Enter the following commands to configure the action clause as capture for the access map:
Switch (conf-map_name)# match ip address CascadeMon
Switch (conf-map_name)# action forward

or
Switch (conf-map_name)# action forward capture // Depending on Cisco IOS rev.
Switch (conf-map_name)# exit

5. Enter the following commands to apply the access map to all VLANs that you want to monitor:
Switch (conf)# vlan filter map_name vlan-list 1-10,15,16...

6. Enter the following commands to specify the capture port (previously configured trunk port):
Switch (conf)# interface GE5/3
Switch (config-if)# switchport capture

Riverbed Cascade Product Family Deployment Guide 59

Packet Collection for Cascade Appliances Shark Passthru

Shark Passthru
The only reason to use passthru mode is to tap a SPAN port that another device is already using: for example, a Sensor.
In passthru mode, you do not have to configure an additional SPAN on the device. With this solution, you are using
two ports on the Shark to monitor a single SPAN port.
When you place the Shark in passthru mode, it acts as a tap on a live interface. In passthru mode, Shark passes traffic
between two physical interfaces on the same card.

Note: The passthru mode does not fail-to-wire. If the Shark loses power or stops operating for whatever reason, the link does not
pass traffic.

Packet Deduplication
Depending upon the packet capture method you use, you might send multiple copies of the same packet to the Shark
or Sensor. This can occur when you are:
 port mirroring multiple VLANs from the same monitoring port and the packet is routed on the device from one
VLAN to another. Even if you are mirroring only ingress to the VLAN, the switch can mirror a copy of the packet
when it enters the first VLAN, and mirror a second copy when it enters the second VLAN.
 port mirroring both ingress and egress on the port or VLAN and the packet is routed into and out of the same port
or VLAN.
 using an aggregating tap and the packets are detected on both ports being aggregated.
 using an intelligent monitoring solution that is capturing the same packet from multiple ports and is not
performing deduplication.
If any of these actions apply to your environment, Riverbed recommends that you use port deduplication on the Sensor,
Shark, and Express. The Shark or Sensor can deduplicate packets when necessary. You must enable this feature on the
Shark, but it is enabled by default on the Sensor. Both appliances deduplicate packets by evaluating the packet
identifier along with other information in the packet. Deduplicated packets can capture TCP retransmissions; and
duplicate packets, due to instrumentation, are dropped. The Shark performs duplication on a per-port basis.
Some network devices might retransmit TCP packets as part of their normal operation. If you are collecting packets
on both sides of such devices to the same port on the Shark or the Sensor, enabling packet deduplication does not
remove retransmission counts as these are true retransmits. The following are two examples:
 If you capture traffic between an Interceptor appliance and Steelhead appliance and are using full transparency,
the packets appear as retransmissions to the Profiler and Express if the packets are captured is on the same port as
the originating packets. To avoid this, do not capture traffic between the Interceptor appliance and Steelhead
appliance if configured with full transparency.
 If you capture traffic on either side of a CISCO ASA Firewall to the same port on a the Shark or Sensor, the ASA
has a security feature that is enabled by default to help protect against TCP session hijacking. This feature causes
the ASA to rewrite sequence numbers for packets traversing it, resulting in observed retransmitted packets if the
packets are captured on the same Shark or Sensor monitoring port. To avoid this, you can disable the connection
random-sequence-number feature on ASA, or you can change your instrumentation so that you do not capture
traffic from both sides of the firewall to the same monitoring port.

60 Riverbed Cascade Product Family Deployment Guide

Snaplen Packet Collection for Cascade Appliances

Snaplen
Snaplen is an abbreviation for snapshot length. Snaplen equals the number of bytes captured for each packet. Having
a snaplen smaller than the maximum packet size on the network enables you to store of more packets, but you might
not be able to inspect the full packet content.
In most cases, you want to configure the Shark to capture full packets. If this is your case, do not adjust the default
snaplen parameter. Do not adjust this parameter. This enables full-packet analysis within Pilot. If you adjust the snaplen
to be smaller, some Pilot views cannot fully analyze the data. For example, volumes of data represented can appear to
be smaller than they actually are, and more detailed views do not contain all the data necessary for full analysis.
Because you do not typically use the Sensor for full-packet analysis, the snaplen is set to 250 bytes by default. This
enables TCP/IP headers and some application-level headers to be captured and viewed for a larger number of packets.
However, because the Sensor performs packet sampling, it is not useful to view a full sequence of packets. In either
case, setting snaplen on the Shark or Sensor does not affect the accuracy of the flow data sent to the Profiler.

Riverbed Cascade Product Family Deployment Guide 61

Packet Collection for Cascade Appliances Snaplen

62 Riverbed Cascade Product Family Deployment Guide

CHAPTER 5 Cascade Appliances and Steelhead
Appliance Integration

This chapter describes how to configure Cascade appliances and the Steelhead appliance into an integrated solution for
traffic monitoring and troubleshooting. When you integrate Cascade appliances and the Steelhead appliance into your
environment, you can successfully analyze the traffic on your network for capacity planning and troubleshooting, and
thus realize the benefits of optimization. This chapter includes the following sections:
 “Steelhead Appliance and Cascade Appliances Overview” on page 63
 “Cascade Appliance Deployment Considerations” on page 66
 “Configuring Steelhead Appliance for Flow Data Export” on page 69

Steelhead Appliance and Cascade Appliances Overview

This section describes a summary of Cascade appliances and includes the following sections:
 “NetFlow Versus CascadeFlow” on page 64
 “SNMP Interface Persistence (ifindex)” on page 65
The primary integration point of the Steelhead appliance and Cascade appliances is the CascadeFlow export from the
Steelhead appliance. The Steelhead appliance sends the Express or Gateway an enhanced version of NetFlow called
CascadeFlow. CascadeFlow includes:
 NetFlow v9 extensions for round-trip time measurements that enable you to understand volumes of traffic across
your WAN and end-to-end response time.
 extensions that enable the Profiler and Express to properly measure and report on the benefits of optimization.
You can deploy a SPAN or port mirror switch in proximity to the Steelhead appliance when you monitor traffic from
the Steelhead appliance's auxiliary port.

Note: In RiOS v7.0.1 and later, RSP was replaced with Virtual Services Platform (VSP). VSP comes preinstalled in the Steelhead
EX appliance. For more information about VSP, see the Steelhead EX Management Console User’s Guide.

Riverbed Cascade Product Family Deployment Guide 63

Cascade Appliances and Steelhead Appliance Integration Steelhead Appliance and Cascade Appliances Overview

NetFlow Versus CascadeFlow

NetFlow provides detailed records of conversations in the network. A basic NetFlow record includes the IP addresses
of the endpoints (workstations, servers, printers, and so on); the port or protocol used; traffic volume in bits and
packets; TCP flags ingress and egress interface; and so on. These records are sent to a flow collector such as the
Gateway. When the records are intelligently coalesced and stored, you can read them to understand traffic throughout
the network for reporting, troubleshooting, and automatic detection of network and application issues.
The Steelhead appliance supports standard NetFlow v5 and v9. Use only these standard versions when exporting flow
to non-Cascade flow collectors or to an earlier version of the Profiler, Express, or Gateway. CascadeFlow was created
by Riverbed. It extends NetFlow v9 with 12 custom extensions that provide additional details specific to flows passing
through the Steelhead appliance.
The following table shows Profiler feature compatibility in RiOS v6.0 and later.

Feature Flow Version Profiler Version

Basic traffic reporting NetFlow v5 and v9 Profiler v8.0 and later

Enhanced reporting: CascadeFlow compatible (v5-based) Profiler v8.3 and later
• Automatic identification of Steelhead
appliance pairs
• Automatic identification of LAN-WAN
interfaces
• WAN optimization reporting

Enhanced reporting: CascadeFlow (v9-based) Profiler v8.4 and later

• Automatic identification of Steelhead
appliance pairs
• Automatic identification of LAN-WAN
interfaces
• WAN optimization reporting
• End-to-end response time measures for
optimized session

64 Riverbed Cascade Product Family Deployment Guide

Steelhead Appliance and Cascade Appliances Overview Cascade Appliances and Steelhead Appliance Integration

When you use CascadeFlow, the Steelhead appliance sends four flow records for each optimized TCP session to the
NetFlow collector: ingress and egress for the inner channel connection, and ingress and egress for the outer channel.
A pass-through connection still sends four flow records even though there are no separate inner and outer channel
connections. In either case, the Profiler, Express, and Gateway merges these flow records together with flow data
collected for the same flow from other devices.
Figure 5-1. NetFlow v9

SNMP Interface Persistence (ifindex)

One of the more common Profiler and Express reporting issues happen when interface names change across reboots
of the Steelhead appliance, which is a direct result of nonpersistent SNMP interface index names.
SNMP uses index values to reference interface names. Only index values are included in the flow record when the
Steelhead appliance sends flow data to the Express or Gateway. An SNMP poll from the Express and Gateway to the
Steelhead appliance is used to map the index number to an interface name. The SNMP interface index value-to-name
mapping can potentially change across reboot and upgrades, causing the Profiler to incorrectly map the Steelhead
appliance interfaces. Use the ifindex-persistence command on the Steelhead appliance to permanently pin SNMP
interface names to SNMP index values.

To enable ifindex persistence

 On the Steelhead appliance, connect to the CLI and enter the following commands:
sh > en
sh # config t
sh (config) # snmp ifindex-persist
#--- You must restart the optimization service, so that netflow can use the configured ifindex
sh (config) # exit
sh # write mem
sh # restart
sh # show service

To verify that SNMP persistence is enabled

 On the Steelhead appliance, connect to the CLI, enter the following command, and look for Persistent ifindex:
yes:
sh # show snmp
SNMP enabled: yes
System location:
System contact:
Engine ID: 0x8000430b805dc6257f4b328d15
Read-only community: riverbed

Riverbed Cascade Product Family Deployment Guide 65

Cascade Appliances and Steelhead Appliance Integration Cascade Appliance Deployment Considerations

Traps enabled: yes

Interface listen enabled: no
Trap interface: primary
Persistent ifindex: yes
No Listen Interfaces.
No trap sinks configured.

For additional details about ifindex values, see “SNMP Integration for Flow Sources” on page 73.

Cascade Appliance Deployment Considerations

This section provides recommendations on how to improve the accuracy of the exported flow data when you deploy
Cascade appliances in specific Steelhead appliance deployment scenarios, and describes certain Steelhead appliance
features. It includes the following sections:
 “Enabling SNMP Polling” on page 66
 “In-Path Deployments” on page 66
 “Virtual In-Path Deployments” on page 67
 “Out-of-Path Deployments” on page 69

Enabling SNMP Polling

The Gateway appliance uses SNMP v3 to poll the Steelhead appliance. This requires the Steelhead Access Control List
to allow the Gateway to poll OID 1.3.6.1.2. One of the steps in creating an ACL on the Steelhead appliance is creating
a View that lists OIDs that are included or excluded from access. If OID 1.3.6.1.2 is not in the “included” section, the
Gateway will not be able to get data when it polls the Steelhead appliance.

In-Path Deployments
In an in-path configuration, you deploy Steelhead appliances in the physical path of the client and server, where they
detect all traffic. You can source flow export from either the Steelhead appliance primary or auxiliary interface to the
Gateway or the Express. Enable flow export for all traffic on all Steelhead appliance interfaces that have traffic
traversing them: for example, lan0_0 and wan0_0 for a single in-path Steelhead appliance deployment.
Riverbed recommends that you enable simplified routing when using Cascade appliances and Steelhead appliance in-
path configurations. Simplified routing avoids situations where traffic can potentially run through the Steelhead
appliance more than once—commonly known as ricochet. When packet ricochet occurs, the same traffic is reported
by the Steelhead appliance multiple times, which causes an unexpected increase in bandwidth, packets, and other
traffic statistics in various Profiler reports. Ricochet can happen when you install the Steelhead appliance in a different
subnet from the client or server, and you do not configure the appropriate static routes to avoid traffic passing through
the Steelhead appliance multiple times on the way to and from the default gateway.
For more details about simplified routing and in-path deployments, see the Steelhead Appliance Management Console
User’s Guide and the Steelhead Appliance Deployment Guide.

66 Riverbed Cascade Product Family Deployment Guide

Cascade Appliance Deployment Considerations Cascade Appliances and Steelhead Appliance Integration

Virtual In-Path Deployments

A virtual in-path deployment happens when you direct traffic to the Steelhead appliance through WCCP, PBR, the
Interceptor appliance, or a Layer-4 load balancer. All traffic might not traverse the Steelhead appliance because of
router redirection access control lists (ACLs). You must rely on NetFlow from the router, as well as the Steelhead
appliance, to ensure that you are sending complete flow data to a Gateway or Express. You must only select optimized
flows to export from the Steelhead appliance to the Express or Gateway, but you must be sure to export NetFlow from
the router for Cascade appliances to detect unoptimized and optimized traffic.
If you configure the virtual in-path environment to use correct addressing, when you export all flow from routers in-
path between the Steelhead appliances, additional flows are exported (original IP and port numbering and inner-
channel IP and port numbering). Selectively export NetFlow to avoid exporting the inner-channel traffic, unless you
wan to track this traffic.
In a virtual in-path deployment, you want to use only the WAN interfaces or traffic interception. The Profiler
optimization-benefit reporting relies on the separation of preoptimization traffic (known as LAN-side traffic), and
post-optimization traffic (known as WAN-side traffic). To separately report LAN-side and WAN-side traffic with only
the physical WAN interface connected, you must enable a pseudo-SNMP interface that represents the LAN-side traffic.
Use the SNMP fake index command to enable the pseudo-SNMP interface. This command causes the Steelhead
appliance to report LAN-side traffic as if it is on a LAN interface during the NetFlow export, even though you are not
using the physical LAN interface.
You can source flow export the primary or the auxiliary interface to export flow data to the Profiler.

To enable fake index

 If you have RiOS v6.0 or later and you want to export CascadeFlow or NetFlow v9, the fake index automatically
starts with flow export, and no additional steps are required.
or,
If you have RiOS prior to v6.0, or when you want to export a flow type other than CascadeFlow or NetFlow v9,
connect to the CLI and enter the following commands:
configure terminal
ip flow-export destination <ip address> <port> interface wan0_0 fakeindex on

For more details about virtual in-path deployments, see the Steelhead Appliance Deployment Guide.

Interceptor Appliance Considerations

When you deploy Steelhead appliances virtually in-path and load-balanced by an Interceptor appliance, in addition to
ensuring that you have fake index enabled, make sure that you avoid capturing inner-channel traffic. When you deploy
a Shark or Sensor, do not place either appliance so it capture traffic between the Steelhead appliances and Interceptor
appliance, because this configuration does not detect any information about the non-optimized network traffic. You
only detect communication between the Steelhead appliances and Interceptor appliance.
If you configure the Steelhead appliance with full-transparency, you must exclude traffic between the Steelhead
appliances and Interceptor appliance when you configure port mirroring for collection by the Shark or Sensor;
otherwise, you might detect excessive retransmission and incorrect or missing RTT metrics.

Subnet Side Rules

In virtual in-path deployments, configure VLAN subnets to enable the Steelhead appliance to properly determine the
direction of non-optimized traffic that passes through the Steelhead appliance. If no LAN subnets are configured, the
Steelhead appliance does not discern whether the traffic is passing from the WAN to the LAN or in the opposite
direction. This can result in over-reporting traffic in a particular direction or for a particular interface.

Riverbed Cascade Product Family Deployment Guide 67

Cascade Appliances and Steelhead Appliance Integration Cascade Appliance Deployment Considerations

In a virtual in-path deployment you must configure LAN subnets on each, because the LAN subnets behind each
Steelhead appliance are typically unique.

To configure subnet side rules on a Steelhead appliance

1. On the Steelhead appliance, choose Configure > Networking > Subnet Side Rules.

2. Select Add a Subnet Side Rule.

3. From the drop-down list, select one of the following:

 Start
 End
 Rule number
The rules are evaluated in order; evaluation stops when a rule is matched. Rules must be in proper order for
your network.

4. Specify a subnet using a valid CIDR notation.

5. Select whether the subnet is on the LAN or WAN side of the appliance.

6. Click Add to save the rule.

7. Continue to add rules until you have mapped all subnets.

8. Click Save to save your changes permanently.

Figure 5-2. Subnet Side Rules Page

68 Riverbed Cascade Product Family Deployment Guide

Configuring Steelhead Appliance for Flow Data Export Cascade Appliances and Steelhead Appliance Integration

Out-of-Path Deployments
In an out-of-path deployment, you do not install the Steelhead appliance directly in the data flow between the client
and server. Because non-optimized traffic does not always pass through the Steelhead appliance, you must export
NetFlow from the router for the Profiler or Express to detect non-optimized and optimized traffic. Export only
optimized traffic from the Steelhead appliance. To export flow data to the Express or Gateway, you can export flow
from the primary or auxiliary interface.
As with virtual in-path configuration, you must enable SNMP fake index to properly report the direction of the
optimized traffic through the Steelhead appliance. You do not need to configure subnet side rules, because the out-of-
path Steelhead appliance detects only optimized traffic and never passes through any flows. For more details about
SNMP fake index, see “Virtual In-Path Deployments” on page 67.
For more details about out-of-path deployments, see the Steelhead Appliance Deployment Guide.

Configuring Steelhead Appliance for Flow Data Export

This section explains how to configure the Steelhead appliance for flow data export.

To enable flow data export on the Steelhead appliance

1. On the Steelhead appliance, choose Configure > Networking > Flow Export.

2. Select the Configure tab to display the Configuration menu.

3. Select Networking to expand the Networking menu.

4. Select Flow Export to display the Networking > Flow Export page.

5. Select Enable Flow Export.

6. Select Disable Top Talkers.

Riverbed recommends that you disable top talkers unless you have it enabled for reporting on the Steelhead
appliance. Enabling top talkers forces the active flow timeout to 30 seconds. This results in flow updates being
sent to the collector twice per minute. While Riverbed recommends a timeout of 60 seconds in most cases, there
are no issues with sending updates more frequently. The collector correctly collates the incoming information and
sends a single, complete, update to the Profiler once per minute.

7. Specify 60 seconds in the Active Flow Timeout field.

8. Specify 15 seconds in the Inactive Flow Timeout field.

9. Click Apply.

Riverbed Cascade Product Family Deployment Guide 69

Cascade Appliances and Steelhead Appliance Integration Configuring Steelhead Appliance for Flow Data Export

10. Click Save to save your changes permanently.

Figure 5-3. Flow Export Page

To configure the Flow Collector and Exporting interfaces

1. On the Flow Export page, select the Add a New Flow Collector tab.

2. Enter the IP address and listening UDP port number of the Express or the Gateway.

3. Select the version of flow data to be exported:

 For the Profiler v8.4 and later, use CascadeFlow.
 For the Profiler v8.3.2, select CascadeFlow compatible and select the LAN Address check box.

4. For the desired interfaces, select All to export both optimized and non-optimized flow information.

5. Click Add to add the Express or Gateway to the collector list.

70 Riverbed Cascade Product Family Deployment Guide

Configuring Steelhead Appliance for Flow Data Export Cascade Appliances and Steelhead Appliance Integration

6. Click Save to save your changes permanently.

Figure 5-4. Flow Collector and Exporting Interfaces

When you click Apply and Add, the Management Console updates the running configuration. Your changes are
written to disk only when you save your configuration.

Riverbed Cascade Product Family Deployment Guide 71

Cascade Appliances and Steelhead Appliance Integration Configuring Steelhead Appliance for Flow Data Export

72 Riverbed Cascade Product Family Deployment Guide

CHAPTER 6 Additional Cascade Appliance
Integration

The Cascade Product Family includes a number of additional integrations that enable you to complete your deployment
by evaluating additional data or integrating with other management and reporting systems. This chapter includes
information about the most commonly used Cascade integrations:
 “SNMP Integration” on page 73
 “SNMP for Switch Port Discovery” on page 75
 “Switch Port Discovery Supported Routers and Switches” on page 76
 “Active Directory” on page 78
 “REST API” on page 80
For additional assistance, contact Riverbed Professional Services.

SNMP Integration
This section describes the following SNMP integrations. It includes the following sections:
 “SNMP Integration for Flow Sources” on page 73
 “SNMP Integration for Device Management of Cascade Components” on page 74
 “SNMP Integration for Sending Traps” on page 74

SNMP Integration for Flow Sources

When devices send flow, standard SNMP interface identifiers (ifindex values) indicate which interfaces the flow
traverses. You must map the interface identifiers to names and descriptions to identify them on the Profiler. You must
also obtain the link speed information so you can convert raw bandwidth numbers to link utilization percentages. The
Profiler, Express, and Gateway gathers the following information using standard SNMP from all devices sending
standard flow, or Steelhead CascadeFlow:
 Device name
 Interface names
 Interface descriptions
 Interface capacities

Riverbed Cascade Product Family Deployment Guide 73

Additional Cascade Appliance Integration SNMP Integration

Ensure that you configure firewalls between the Express or Gateway and flow-reporting devices to enable SNMP
access between the Express or Gateway and remote device. If there are any access rules on the flow-reporting devices,
you must enable these access lists to allow SNMP access from the Express or Gateway.
For more information about configuring of the Profiler and Express for SNMP collection of these items, see the
Cascade Profiler and Cascade Express User’s Guide.

SNMP Integration for Device Management of Cascade Components

You can monitor the status of Cascade appliances through SNMP from an external SNMP manager. Currently, the
Profiler, Gateway, and Sensor generally support the industry-standard UCD-SNMP-MIB. For more information, see
http://www.oidview.com/mibs/2021/UCD-SNMP-MIB.html.
When you use SNMP, it is normal to detect high CPU and memory use. This does not mean that the appliance is
experiencing a problem. Because Cascade appliances are appliances and not standard servers, processes tend to hold
the entire CPU for normal use (100% CPU utilization is normal) and make efficient use of available memory resources.
For the Enterprise Profiler, you can poll each physical component separately.
Because the Profiler reports a broken connection with a Sensor, Gateway, or Shark, a best practice is to configure your
SNMP manager to send health traps and only poll the Profiler Event Manager module.

SNMP Integration for Sending Traps

The Profiler can send traps through SNMPv1 or SNMPv3 to third-party trap receivers. You can customize which types
of traps to send to which devices within the notifications pages of the Profiler UI. Some of the use cases for sending
SNMP traps are as follows:
 Sending the Profiler or Express health messages to a third-party network manager or SNMP device manager
 Sending service and performance and availability events to a third-party network manager
 Sending security events to a security event manager (SEM)
You must configure the third-party device receiving the trap with the Profiler MIB (labeled Mazu-MIB). This MIB is
available from either the Riverbed Support site or Profiler help downloads page.
You can route the appropriate events to the appropriate devices by first configuring recipient groups within the Profiler
and then configure which events are sent to which recipients. Recipient groups can contain email recipients and
SNMPv1 or SNMPv3 recipients. For more information about how to configure these notifications, see the Cascade
Profiler and Cascade Express User’s Guide.

74 Riverbed Cascade Product Family Deployment Guide

SNMP for Switch Port Discovery Additional Cascade Appliance Integration

SNMP for Switch Port Discovery

Standard flow records identify hosts by their IP addresses. The Profiler supports discovery of MAC addresses and
switch port locations of individual hosts based upon their IP addresses. This enables the Profiler to display complete
information, as shown in Figure 6-1. This information appears in multiple places throughout the Profiler UI.
Figure 6-1. MAC and Host Switch Information Displayed as a Result of Switch Port Discovery

Riverbed Cascade Product Family Deployment Guide 75

Additional Cascade Appliance Integration Switch Port Discovery Supported Routers and Switches

Switch Port Discovery Supported Routers and Switches

The following table shows a partial list of supported switches and routers for switch port discovery in the Profiler v9.0
and later.

Vendor Model Lookup Router Switch Comments

Airespace wireless 3500, 4101, 4102 No Yes APs appear as switch ports
controllers

Allied Telesyn AT-8000 No Yes

Aruba wireless 5000, 6000 Yes Yes APs appear as switch ports
controllers

Cisco 2500 series 2501, 2503, 2511, 2514, Yes Yes

AS2509RJ, AS2511RJ

Cisco 2600 series 2610, 2610XM, 2611, 2620, Yes Yes

2620XM, 2621, 2621XM,
2651XM, 2691

Cisco 2800 series 2811, 2821, 2851 Yes Yes

Cisco Catalyst 2900 2908xl, 2912MfXL, 2924CXL, No Yes

series 2924CXLv, 2924 MXL

Cisco 2940 and 2950 2940-8TT, 29500t24 No Yes

series

Cisco 2970 series 2960, 2970G-24T-E No Yes

Cisco Catalyst 3500 3508GXL, 3524XL, 3548SL No Yes Layer-2 devices—only when
series you run Cisco IOS

Cisco 3550 (partial) 3400 with MetroBase, 3550- No Yes Running Cisco IOS
12T

Cisco Catalyst 3550 3550, 3560, 3550-24, 3550-48 Yes Yes Running Cisco IOS
(partial)

Cisco Catalyst 4000 4006, 4503, 4506, 4507, 4510, No Yes

series wsc4003, wsc4006, wsc4503,
wsc4506, wsc4912g

Cisco Catalyst 5000 Wsx5302 Yes No Most models not supported

series

Cisco Catalyst 6500 6503, 6509, sp72033, s3223, Yes Yes

series s32p3, s222, 6kMsfc, 6kMsfc2,
wsc6509

Cisco wireless 2006, 4112, 4124, 4136, 4402, No Yes APs appear as switch ports
controllers 4404

Dell PowerConnect 3348, 3448P, 3424, 3424P, 5324 No Yes

3000 and 5000 series

Dell PowerConnect 6024F, 6224, 6248 Yes Yes

6000 series

IBM BladeCenter All No Yes

Ethernet switch family

76 Riverbed Cascade Product Family Deployment Guide

Switch Port Discovery Supported Routers and Switches Additional Cascade Appliance Integration

Vendor Model Lookup Router Switch Comments

Linksys 2048 family All No Yes

Enterasys Networks Matrix N-series DFE Yes Yes

Matrix series

Enterasys Networks C3G124-24, C3G124-48, Yes Yes

SuperStack C-series C2G124-24, C2G124-48

Extreme Network Alpine 3808, Summit 7i, 48si Yes Yes

Alpine and Summit

Foundry EdgeIron EdgeIron 24G No Yes

series

Foundry IronWare FLS624, FLS648, FWSX424, Yes Yes

family ServerIronGT

HP ProCurve 2312, 2324, 2510, 2512, 2524, No Yes ProCurve devices are widely
2600, 2610, 2626, 2650, 2800, supported (newer devices not
2810,2900, 2910al, 3124, in this list are likely
3324XL, 3400cl, 3500, 3500yl, supported)
4000, 4100GL, 4104GL,
4108GL, 4200vl, 5300XL,
5400yy, 5400zl, 6108, 6200yl,
6400cl, 6410cl, 6600, 6600ml,
8000, 8200zl
Juniper M-series Router All Yes No
series

Juniper Netscreen series All Yes No

Nortel Alteon AD series 180, 183, 184, AD2, AD3, AD4 Yes Yes

Nortel AP222x series AP-2220, AP-2221 No Yes

Nortel BayStack Hub 102, System 5000 No Yes Requires Advanced NMM
series
Nortel Business Switch -50, 110, 120, 210, 220, 1010, Yes Yes
series 1020

Nortel Centillion series 5000BH, 5005BH, C50, C100 No Yes v4.x/5.x or later
Nortel Ethernet 2526, 2550, 3510, 4524, 4526, Yes Yes
Routing/BaystackYes- 4548, 4550, 5510, 5520, 5530
Yes-

Nortel Passport 1600 1612, 1624, 1648 Yes Yes

series

Nortel Routing, Accelar 1050, 1100, 1150, 1200, 8106, Yes Yes For 8600, code for switch
Family 8110, 8603, 8606, 8610, 8610co support must be v3.2 and
later

Nortel Baystack Switch 303, 304, 350, 380, 410, 420, No Yes
series 425, 450, 460, 470, BPS

Nortel Multiprotocol 2430, 5430, AN, ARN, ASN, Yes Yes

Router/BayRS BLN, BCN

Nortel Synoptics 281X, System3000 No Yes

Riverbed Cascade Product Family Deployment Guide 77

Additional Cascade Appliance Integration Active Directory

Vendor Model Lookup Router Switch Comments

Nortel VPN Router/ 100, 400, 600, 1000, 1010, Yes No

Contivity 1050, 1500, 1600, 1700, 1740,
1750, 2500, 2600, 2700, 4500,
4600, 5000

Nortel Wireless 2270 2270 No Yes APs appear as switch ports

Active Directory
The Profiler provides a user identity feature that maps active directory (AD) user names with IP addresses. This feature
enables you to view:
 the users associated with an end station on the network (Figure 6-2).
 all the end stations that a user has logged into (Figure 6-3).
Figure 6-2. Users Logged into a Given Host for a Selected Time Period

78 Riverbed Cascade Product Family Deployment Guide

Active Directory Additional Cascade Appliance Integration

Figure 6-3. Hosts That a Given User Logged into for a Selected Time Period

This feature relies on the security audit events obtained from one or more Microsoft active directory domain
controllers. You can send this event data directly to the Profiler from a domain controller, or for AD-2008, an event
collector host. Riverbed provides a service application named Cascade AD Connector that forwards the appropriate
events from a domain controller, or event collector, to the Profiler.

Integration for Active Directory 2008

For AD-2008, you must use the Cascade AD Connector v2.0. You can install the connector on either the domain
controllers or an event collector, but Riverbed recommends that you install the AD Connector on the event collector.
Even though installation on a domain controller is easier, you must install it as many times as the number of domain
controllers. The installation on an event collector requires a few more steps, including planning of event-collecting
topology (if not already implemented in the environment), but it requires no additional product installed on domain
controllers and provides more flexible delivery paths.
For more information about configuring integration with AD-2008, see Cascade AD Connector 2008 for Windows
Server 2008: Riverbed Cascade AD 2.0 Release Notes.
You can download the connector and the document from either the Riverbed support site or directly from the Profiler
help downloads page.

Riverbed Cascade Product Family Deployment Guide 79

Additional Cascade Appliance Integration REST API

Integration for Active Directory 2003

For AD-2003, you must use the Cascade AD Connector v1.5. You can install the connector on the domain controllers
or another Windows server acting as a collector within the same domain controller, but Riverbed recommends that you
install the connector on the domain controller. Installing the connector directly on the domain controller requires no
messaging between the domain controllers and an external collector, whereas if you use the external collector, you need
significant inter-system communications.
For more information about configuration integration with AD-2000 and AD-2003 environments, see Technical Note
#29: Microsoft AD Integration for User Identity.
You can download the connector and the document from Riverbed Support or directly from the Profiler help
downloads page.

REST API
Cascade v10.0.5 and later includes a REST API for the Shark and Profiler. Using REST API, you can perform such
activities as:
 Packet captures on Shark
 Traffic reports on Profiler
 WAN optimization reports on Profiler
 Appliance configuration
You can see more information about specific functions of the REST API directly on your appliance at
https://{Profiler or Shark address}/rest/api/index.php.

80 Riverbed Cascade Product Family Deployment Guide

CHAPTER 7 Cascade Profiler Analytics and
Service Monitoring

Cascade appliance service monitoring simplifies discovering, modeling, and configuring monitoring for enterprise
applications. This chapter describes how you can account for the analytic license limit when mapping services. This
chapter also provides best practices for how to stay within these limits and which specific metrics to use.
This chapter includes the following sections:
 “Analytic License Limits per Profiler Platform” on page 81
 “Understanding the Analytics License Limits” on page 82
 “Reducing the Number of Metrics” on page 83
 “Conditions Required for Baseline Establishment” on page 85
 “Determining Which Metrics to Use” on page 85
This chapter assumes you are familiar with the Cascade Profiler and Cascade Express User’s Guide.

Analytic License Limits per Profiler Platform

If you are running the Profiler v9.0.5 or later with a deployed analytics license, the following table shows the number
of analytics policy metrics allowed system wide.

Profiler Model Available Analytic Policy Metrics

Express Up to 5000

Standard Profiler Up to 7500

Enterprise Profiler Up to 10,000

The available analytic policy metrics number includes the total number of analytic policy metrics to be tracked system
wide. It includes the following:
 All metrics tracked through service configuration, up to eight metrics PER segment monitored
 All metrics tracked through performance and availability policy configuration:
– Link congestion, up to two metrics per policy configured
– Link availability, up to two metrics per policy configured

Riverbed Cascade Product Family Deployment Guide 81

Cascade Profiler Analytics and Service Monitoring Understanding the Analytics License Limits

– Application performance, up to seven metrics per policy configured

– Application availability, up to two metrics per policy configured
The following policy types do not count toward the analytics limit:
 Security policies
 User-defined policies

Understanding the Analytics License Limits

You can configure hundreds of metrics on the Profiler by clicking a few checkboxes. If you understand the concept
behind the check boxes, you can make intelligent decisions about how to configure services for optimal performance.
Figure 7-1 shows a simple SharePoint service configuration. The SharePoint Web servers are connected to three end-
user locations (Atlanta, Boston, and New York), and are supported by LDAP and MS-SQL on the back end.
Figure 7-1. Simple SharePoint Service Configuration

Using the example shown in Figure 7-1, five metrics are selected (Figure 7-2) for each of the segments during service
discovery for the SharePoint service.
Figure 7-2. Five Metrics for SharePoint Service

For the backend segments, DBTransactions and AuthenticationService, there are a total of 10 metrics used (five
metrics times two segments). However, for SharePoint-Web, there are three end-user locations. Because it is necessary
to monitor each location individually, this yields a total of fifteen metrics (five metrics time three locations). The total
number of metrics used to monitor the SharePoint service is twenty-five.
A location with a 100 end users is more representative of a larger enterprise network. If there are 100 end-user
locations, the SharePoint service requires over 500 metrics (510 metrics to be exact) for monitoring.
You can view the number of metrics on a running system:

82 Riverbed Cascade Product Family Deployment Guide

Reducing the Number of Metrics Cascade Profiler Analytics and Service Monitoring

 during the process of service discovery using the wizard.

On the Commit Changes to Service page of the wizard, you have to commit changes to the service before the
analytics are configured. The page shows a count of how many metrics to be added when you save the service.
Figure 7-3. Metrics Shown on the Commit Changes to Service Page

Note: The Commit Changes to Service Page only shows how many metrics are added when the service is committed. It does
not display the total used or how many are still available.

 on the Information page.

The System > Information page includes the Policies Usage display, which shows the total number of metrics
currently in use and the overall limit for the system. The difference between the number of metrics currently used
and the limit (5000, 7500, or 10,000, depending upon the platform) is the number remaining. The remaining
number is available for you to configure.
Figure 7-4. Metrics Shown on the Information Page

Reducing the Number of Metrics

You can reduce the number of metrics used with services in the Profiler in the following ways:
 Reducing locations - You can reduce the number of locations you monitor to reduce the number of metrics in
use. The following methods reduce the number of locations:
– You can reduce the number of groups in the group type used for identifying end-user locations, for a
reduction across all services.

Riverbed Cascade Product Family Deployment Guide 83

Cascade Profiler Analytics and Service Monitoring Reducing the Number of Metrics

– You can reduce the number of locations you use for a specific service. Figure 7-5 shows how to reduce the
number of locations during discovery for any service by choosing a subset of the end-user locations to
monitor.
Figure 7-5. Reducing the Number of Locations on the Customize Monitoring on the Locations Page

Note: Both methods cause a reduction in the number of configured policies configured.

 Not tracking by groups - You can turn off by-group tracking and track all front-end metrics as one item instead
of a detailed breakdown by groups. This might make sense if all sites or groups have similar characteristics (for
example, response time) when attaching to the front end of the application. To turn off by-group tracking, edit
and deselect the Track By End-User checkbox on the front-end end-user component.

Note: You continue to receive a list of clients that have an impact if an analytic has an error.

 Reducing monitoring - You can reduce the number of metrics being monitored by eliminating a metric per end-
user location and a metric for each back-end segment in the service. Figure 7-6 shows four metrics configured by
default and four additional metrics you can configure. Turning off only one of these metrics, particularly if there
is a large number of end-user locations, can result in a large reduction in metrics you use.
Figure 7-6. Default and Configurable Metrics on the Customize Monitoring on the Metrics Page

As mentioned earlier, you can aggregate locations into regions. You do not reduce the number of metrics you have, but
you can more easily manage the Services Dashboard. If you have hundreds of locations in the ByLocation group, you
can create a ByRegion group type instead. Specify the group type to represent the end-user locations on the General
Settings page. The default setting specifies the ByLocation group type.

84 Riverbed Cascade Product Family Deployment Guide

Conditions Required for Baseline Establishment Cascade Profiler Analytics and Service Monitoring

Conditions Required for Baseline Establishment

A segment must collect a certain amount of data before it can establish daily and weekly baselines.The daily baseline
requires three days and one hour of collected data. Until the system has been running and collecting data for a segment
for this period of time, the analytic metrics related to the segment metrics remains in an initializing state.
The weekly baseline requires three weeks and one day of data. Additionally, for each metric to initialize, there must
be some data for the metric in 50% of each 15-minute time period. This means that if the segment is a backup that runs
only once per day or is active for only a few hours a day, the segment does not initialize.

Determining Which Metrics to Use

When you configure service monitoring within the Profiler, Step 4 of the configuration wizard enables you to decide
which metrics to monitor for each segment within the service. The metrics are organized according to three different
categories. The following table summarizes all metrics.

Category Metric Name Enabled Dips Spikes

Connectivity New connections Yes

(Conn)
Active connections Yes

Bandwidth Yes

User experience Response time Yes Yes

(UserExp)
Average connection duration Yes

Average application throughput per connection Yes

Efficiency TCP retransmissions Yes Yes

(Effncy)
Number of TCP resets Yes Yes

The following metrics are enabled by default:

 Active connection rate (detecting spikes)
 Response time (detecting spikes)
 TCP resets (detecting spikes)
 TCP retransmissions (detecting spikes)
When you consider which metrics to use for an application, take into account what each segment is responsive for
versus the service as a whole. Front-end segments might have very different characteristics than back-end segments.
For example, if you have a Web-connected front end, you might detect numerous brief connections versus a back-end
segment for the same service, which might have only continuous database interactions over only a handful of
connections. For another service, you might have a front-end segment that uses Citrix, which might keep connections
open throughout very long periods of time, while back-end connections to application servers might be shorter in
duration but greater in number. For details about characteristics to consider per metric, see “Reducing the Number of
Metrics” on page 83.
The four default, enabled metrics satisfy a majority of TCP-based application segments, although for segments with a
low number of connections, you might want to disable or change the settings on the active connections metric.
When you choose which metrics to include, use the following best practice guidelines:

Riverbed Cascade Product Family Deployment Guide 85

Cascade Profiler Analytics and Service Monitoring Determining Which Metrics to Use

 Applications with low connectivity rates - For back-end segments for which applications have connectivity
between only a few servers, or front-end segments for which only a few clients are connected at a time, the active
connection rate is very low, and the tolerance band might be very tight. You might detect alerts when there is a
very minor change in connectivity (one new session connects longer than what is normal).
For these situations, you can disable this metric, or you can increase the tolerance band and add an appropriate
noise floor to the metric. The noise floor can help control minor fluctuations. Figure 7-7 shows a segment that has
only a few connections active per second, with a raised tolerance to 5 for low and 6 for high, and an added noise
floor of four connections per second.
Figure 7-7. Metric for a Low Connectivity Rate Segment

 Active connection rate metric consideration without weekly seasonality - If you are trying to keep the number
of alerts low, Riverbed recommends that you not disable the active connections metric until after the weekly
baseline is set. The baseline is three weeks and one day of data.
 UDP applications - For UDP applications, the TCP health and TCP performance measurement-based metrics do
not work. You can disable TCP resets, TCP retransmissions, and response time. For UDP segments that have
periodic bandwidth, you can enable the bandwidth metric.
 Back-end segments with continuous communications - For many back-end segments, you can enable the
average application throughput per connection metric. This metric tracks the bandwidth that is consumed during
the active parts of the session. You are alerted when the baselined value dips below the threshold. This dip can
indicate that less data is transferred, which can indicate that the application efficiency has dropped, and this can
have an impact on user experience.
 Single-transaction-oriented TCP sessions - For application segments that tend to set up a new TCP session for
each transaction, you can enable the average connection duration metric. This metric tracks the duration of the
connections and alerts you if it dips below that baseline. For this type of segment, tracking new connections in
addition to active connections can also be beneficial.
 Revisit metrics and tuning after three weeks of data - Although three days and one hour of data are required
for the analytic metrics to initialize, it takes three weeks and one day for the analytics to begin using a weekly
baseline. This baseline becomes more predictable when you monitor weekly seasonality is monitored (for
example, lower traffic volumes on the weekend). Tuning and final decisions on which metrics might not be best
for the segment are made after this time period.
 Understanding the characteristics of your application - To better understand the characteristics of the
segments on your application, you can run service-level-objective (SLO) reports after the segments have
initialized. The SLO reports enable you to see the baselined periodicity of each metric. If the segment has not yet
initialized, you can run reports to gain a better understanding of the segment characteristics. Running reports in
this manner helps you to choose which metrics to use per segment and to fine-tune after initialization.

86 Riverbed Cascade Product Family Deployment Guide

CHAPTER 8 Troubleshooting

Troubleshooting Cascade appliances can be a complex process involving looking at multiple different areas of the
product.
In its simplest form, the Profiler is divided into two distinct areas: the Web-based UI and the supporting infrastructure
(system processes, scripts, and so on). Problems can occur in one area only or across multiple areas. For example,
identifying why specific data is missing from a report might reveal a problem with the way the report is being run (a
UI-based problem), a problem with the underlying query processing engine (an infrastructure-related problem), or a
combination of the two.
This chapter includes troubleshooting information in the following areas:
 “RTT Values Not Available” on page 87
 “Not Receiving Reports by Email” on page 88
 “DNS Names Not Being Resolved in Reports” on page 88
 “Reports Are Not DNS-Resolving All Addresses” on page 88
 “Data in Reports Seems Inconsistent” on page 89
 “Sensor Protocol Violations” on page 90
 “Communication Issues” on page 90
 “Switch Port Discovery Troubleshooting” on page 91

RTT Values Not Available

When you run a report with RTT columns, sometimes the included rows do not show data. Some of the reasons data
might not be available include the following:
 Non-TCP flow - The Profiler calculates RTT information only for TCP-based flows. Any flow that is not TCP-
based (ICMP, UDP, and so on) does display RTT information.
 Flow not seen by the Shark, or Sensor - The flow must be detected by a Shark, Sensor, or Express to calculate
RTT information. If a flow is reported only from a network router through NetFlow, there is not any RTT
information available.
If the flow is only partially seen (for example, due to asymmetric routing) by the Shark, or Sensor, the RTT
information is not valid and is discarded.

Riverbed Cascade Product Family Deployment Guide 87

Troubleshooting Not Receiving Reports by Email

 Retransmits during the initial flow setup - The Profiler does not show RTT information if there is retransmitted
information during the initial flow setup. If the Profiler does not know which packet is the correct packet to use,
there is no way to accurately calculate RTT information.
 Delay between the packets - If the delay between the SYN, SYN-ACK, ACK, and DATA packets is too great
(multiple minutes), the RTT timer might expire and RTT information is not calculated.
 Drops questionable values - The Profiler takes a conservative approach and drops values that are questionable.
 Flows that start before the initial startup time - Because RTT information is calculated only during the initial
flow setup, any flows that started before the beginning of the report time frame do not include RTT information:
for example, a report from 14:00:00-15:00:00 that includes flows that started prior to 14:00:00 do not show RTT
information.

Not Receiving Reports by Email

Information and errors related to the emailing of reports are stored in the Profiler audit log. Anytime an email is sent,
an entry is placed in the audit log, as are any error messages that are received during the transmission process.
When you configure the SMTP server you must use an appropriate host and required account information. You can
choose to use a username and password for authentication. Ensure that you configure the SMTP server to enable
connections from the Profiler and that it is able to relay messages to the appropriate destinations.

Note: The Profiler does not currently support encrypted SMTP.

DNS Names Not Being Resolved in Reports

You must complete the following steps for the Profiler to resolve IP addresses to DNS names:

1. Configure the DNS servers in the General Settings page.

2. In its UI Preferences settings, you must configure each user account to resolve IP addresses to DNS names.
The following are options for DNS resolution in UI Preferences:
 Resolve host names using DNS - Turns on the use of DNS to resolve an IP address into a host name.
 Resolve host names for hosts managed by DHCP - Uses the optional Cascade DHCP integration to use the
information provided during DHCP imports for name resolution.
 Suppress DHCP/DNS search domains from resolved host names - Suppresses the display of the listed search
domains (for example, riverbed.com) when showing names.

Reports Are Not DNS-Resolving All Addresses

To prevent the Profiler from overwhelming DNS servers, you can place the following limits on resolving hosts. You
can control the:

88 Riverbed Cascade Product Family Deployment Guide

Data in Reports Seems Inconsistent Troubleshooting

 number of DNS resolution requests the Profiler sends to a DNS server at one time.
 maximum number of addresses the Profiler attempts to resolve addresses for.
Riverbed recommends that you set these values to 500 each to prevent overwhelming the DNS server. This mean that
the Profiler does not attempt to resolve more then 500 rows from any one table and sends up to 500 requests at one
time. You can increase these values to allow more addresses to be processed.
Be aware that the Profiler waits only one second for responses from DNS servers, to prevent slow DNS servers from
delaying the return of reports. Any addresses that are not resolved before the time expires do not display the associated
DNS name.

Data in Reports Seems Inconsistent

This issue is one of the hardest issues to troubleshoot. Consider the following possible causes:
 Mismatched report types - If you run a host-centric Profiler report versus an interface-based report.
A host centric-report is any report you run from the Reports > Traffic Reports page, and select the Host,
Application, or Advanced tab. These reports always query the database based upon the perspective of the hosts in
the hosts field. When you run these queries, you look for all hosts matching the query conditions, and all output is
from the perspective of the hosts.
An interface-centric report is any report you run from the Reports >Traffic Reports page, and select the Interface
tab. These reports always run from the perspective of the interfaces in the interface field. When you run these
queries, you are querying for all interfaces matching the query conditions, and all output is from the perspective
of the interfaces.
 Missing data
The primary causes of missing data are as follows:
– No coverage of the desired data - With a large network, you can have pockets of data that are not covered
by devices reporting to the Profiler: for example, a branch office might not have a device reporting traffic
internal to the branch. You cannot report on traffic for which there is no coverage.
– Too many devices reporting data - The Profiler currently is limited to storing data from a five devices. If
flow is reported from more than five devices, the data is not retained. This results in the reported traffic rates
for those devices being inaccurately low.
– Missing directional data (ingress or egress) - When you export NetFlow, depending on the version and
device type, you might not be able to export both ingress and egress data for each interface. Consider a
common example in which only ingress data is received with NetFlow v5. When the flow records are sent,
the egress interface is indicated in the record, even though the statics are counted on an ingress interface.
When Profiler receives these ingress records, it assumes that the data is preserved as it passes through the
device. This means that if the record is received on Interface 1, and the record indicates the egress interface is
Interface 2, Profiler assumes that this amount of data leaves the device on Interface 2.
In some cases where this assumption is not valid: for example, on a router with a 10-Mbps interface on one
side, and a 1.5-Mbps interface on the other side, and the router is forced to dropped data due to the 1.5-Mbps
interface being oversubscribed. Because the Profiler has no way of knowing how much data actually went out
the other interface, the numbers can be incorrect. If the 10Mbps-interface is receiving 5 Mbps, the Profiler
reports 5 Mbps leaving the 1.5Mbp-interface because this is the best data Profiler received.

Riverbed Cascade Product Family Deployment Guide 89

Troubleshooting Sensor Protocol Violations

– Routing issues - The Profiler must detect all the details of each flow that it reports on to provide accurate
information. When you have asymmetric routing (where traffic takes one path from client-to-server and a
different path from server-to-client), the Profiler can miss one side of the conversation. This results in
inaccurate information from reports.
 Different data resolution - While the Profiler provides several different data resolutions many other devices do
not provide multiple resolutions or do not allow detailed control over which resolution you use. Comparing a
report from two different resolutions (for example, one hour on the Profiler and five minutes on a router) is very
likely to result in differences in reported values.

Sensor Protocol Violations

The Sensor protocol violation (SPV) error is one of the more common errors that you can receive when using the
Profiler. The error appears as a system event indicating that the violation occurred between the Profiler and one or more
reporting devices (Gateway, Shark, or Sensor). Two potential causes of the SPV error are slow transfers and lack of
time synchronization:
 Slow transfers - Because the Profiler must receive data from remote devices and analyze it all within one 60-
second period, there is very little leeway for slow transfers. The Profiler allows 48 seconds in each minute for all
remote devices to send their data (transfers happen in parallel). Any data sent to the Profiler, any data after the 48
seconds is ignored and SPV errors are generated.
If you detect SPV errors, send a file transfer between the Profiler and remote device reported in the SPV and
calculate how fast data is transferring. If the transfer rate is extremely low and the number of flows is high, this is
the likely issue.

Note: If the issue is time sensitive—it only occurs when backups are also running—you must perform the test around the same
time that the issue occurred.

 Lack of time synchronization - If one Shark, Sensor, or Gateway is unable to NTP synchronize with the
Profiler, the time on that device can drift sufficiently from the Profiler time. The Profiler then has difficulty
ensuring that the data being sent is for the same time slice the Profiler is currently processing. You must ensure
that no firewall or ACL are blocking NTP.

Communication Issues
The Cascade appliances communicate with each other on select ports. These ports must be open between the Profiler
and the remote device for all functions to work correctly. The following primary ports are used for communications
among devices:
 TCP/41017 - Used to pass data back and forth between the Profiler and remote devices such as the Shark and
Gateway. You must open this port bidirectionally between devices.
 TCP/8443 - Used to facilitate the exchange of SSL keys between the Profiler and remote devices such as the
Gateway and Sensor. You must open this port bidirectionally between devices.
 UDP/123 - Used for NTP synchronization between the Profiler and the remote devices such as the Shark and
Gateway. The Profiler acts as the NTP server for the remote devices.

90 Riverbed Cascade Product Family Deployment Guide

Switch Port Discovery Troubleshooting Troubleshooting

If anything is blocking communications on the specified ports, that portion of the Profiler system does not work
correctly. For example, if UDP/123 (NTP) is blocked between the Profiler and Shark, the time on the Shark is likely
to drift, resulting in inaccurate reports.
To ensure that a port is open, use the following telnet and CLI commands:
[mazu@cascade-gateway etc]$ telnet 10.38.7.8 41017
Trying 10.38.7.8...
Connected to 10.38.7.8.
Escape character is '^]'.
^]
telnet> quit
Connection closed.

In this example, the connection was successful, and the CLI output shows port TCP/41017 is open between the host
and 10.38.7.8. Had the port been closed, the conversation might have looked like this:
[mazu@cascade-gateway etc]$ telnet 10.38.7.2 41017
Trying 10.38.7.2...
telnet: connect to address 10.38.7.2: connection refused

The connection was rejected by 10.38.7.2 on port TCP/41017.

Switch Port Discovery Troubleshooting

If device polling fails, ensure that rules on the device allow polling from the Profiler or Express. Also check the device
is in the list of supported devices in “Switch Port Discovery Supported Routers and Switches” on page 76. If polling
still fails, you can enter the following CLI commands to verify SNMP communication and support.
For a lookup router, enter the command:
Profiler# snmpinfo <RouterIP> --dev --fw --version <1 or 2> --comm <READonlyCommunityString>

For a switch, enter the command:

Profiler# snmpinfo <RouterIP> --dev --fw --version <1 or 2> --comm <READonlyCommunityString>

Command output includes information about whether or not the device is reachable and supported. You can give
Riverbed Support the output of the command for additional information.
For more details, see “Additional Cascade Appliance Integration” on page 73.

Riverbed Cascade Product Family Deployment Guide 91

Troubleshooting Switch Port Discovery Troubleshooting

92 Riverbed Cascade Product Family Deployment Guide

APPENDIX A Licensing

This appendix explains various aspects of licensing the Cascade Product Family. It includes the following sections:
 “Licensing Overview” on page 93
 “License Installation” on page 94
 “Other Device License Installation” on page 95
 “Assigning Licenses” on page 95
 “Manual License Installation” on page 95
 “Automatic License Upgrades” on page 96
 “Evaluation Licenses” on page 96
 “Licenses Available” on page 96

Important: Flow per minute limits changed in February 2013. For information about the Cascade Product Family prior to February
2013, consult documentation for that software release.

Licensing Overview
The Profiler, Gateway, and Sensor systems have the following types of licenses:
 “Runtime Licenses” on page 93
 “Capacity Licenses” on page 94
 “Option Licenses” on page 94

Runtime Licenses
All devices require a runtime license. Runtime licenses permit basic system operation. If you do not have a runtime
license installed, your system operates in a basic mode, which enables basic UI interaction, but does not process new
flow data.

Riverbed Cascade Product Family Deployment Guide 93

Licensing License Installation

Capacity Licenses
Capacity licenses control the maximum amount of data the device can process. Capacities are set based on the
maximum number of flows that the device can process each minute (for example, an F1 Gateway can accept no more
than 100,000 flows per minute).
Different types of Cascade devices have different flow capacities:
 Gateway - 100,000, 200,000, 500,000, 800,000, and 1,400,000 flows per minute
 Gateway-VE - 90,000 flows per minute
 Express 360 - 15,000, 30,000, 60,000, and 90,000 flows per minute
 Express 460 - 15,000, 30,000, 60,000, 90,000, and 120,000 flows per minute
 Profiler - 150,000, 300,000, and 600,000 flows per minute
 Profiler-VE - 90,000 flows per minute
 Enterprise cluster - 800,000 or more, depending on the number of analyzer or expansion modules you have
installed
Some Cascade appliances do not require a capacity license. The Sensor and Shark enable data to be processed up to the
capacity of the system without requiring additional capacity licenses.

Option Licenses
Option licenses enable you to add functionality to the Express, Profiler, Profiler-VE, or Enterprise Profiler cluster.
Examples of functionality that requires an option license include:
 Profiler or Express analytics
 Profiler or Express security module
 SAN storage support
The security module and analytics licenses do not require you to purchase an additional part number. The Express,
Profiler, and Enterprise Profiler cluster appliances support the security module through the runtime license and include
the analytics license (as of February 2013).

License Installation
License installation varies depending on the type of hardware the license is installed on. Cascade appliances systems
based on current-model hardware use an automatic license server that provides license keys directly to the Cascade
appliance or through a Web portal that enables you to manually install the license. Older hardware relies on manually
installed licenses, using instructions provided with the license.

Current-Generation Hardware License Installation

When you purchase a license for your Cascade appliance system for xx60 hardware, a license key is generated and
automatically assigned to the system in the Riverbed licensing portal.
If the Cascade appliance system has Internet access, you can automatically download the new license to the device. If
the Cascade appliance system does not have Internet access, you can manually add the new license to the system by
entering the provided key into the Cascade appliance UI.

94 Riverbed Cascade Product Family Deployment Guide

Assigning Licenses Licensing

Previous-Generation Hardware License Installation

When you purchase a license for a Cascade appliance system for earlier hardware (typically hardware purchased prior
to February 2012), an email is sent from Riverbed that includes the command you run to enable the purchased license.
You must run this command exactly as specified, and only on the system the license is intended for.

Other Device License Installation

The Pilot and Shark have different license installations.
You purchase Pilot licenses one complete Pilot software installation per workstation. You do not need to purchase any
options for Pilot beyond the initial license. To have a fully functional software licence without expiration, you must
activate the license with Riverbed through the Internet (if Internet access is not available the license can be activated
by phone, fax, or mail). You can use Pilot for 30 days without activating the license.
There is no software-based upgrade or license required for the Shark. The Shark does not support an upgradeable
software license model (similar to the Sensor). Instead, the Shark is sold with a base chassis (1U, 2U, or 3U) that
supports as many packets per second as the NIC cards that the Shark can support.

Assigning Licenses
When you purchase Cascade appliance hardware, all appropriate licenses are assigned to each unit. The build-time
license (indicating what type of system the unit is: for example, the Gateway) is installed during the manufacturing
process and you can not change it. Additional licenses (capacity and option licenses) are also generated during the
manufacturing process but are not installed directly on the system at that time.
You can retrieve the additional licenses on the Riverbed licensing portal. For example, if you purchase a single CAG-
2260 with an F1 capacity license, a license is generated on the Gateway that enables it to support 100,000 flows per
minute. This license is not activated, but it is available to you in the Riverbed licensing portal. The license
automatically downloads to the Gateway when the Gateway licenses are synchronized with the portal.
Sometimes you must manually assign licenses to specific hardware. This helps ease the process of installing different
models of hardware in different locations.
For example, if you want to deploy 10 Gateways at sites around the world, each site requires a different capacity level.
You purchase four F4 Gateways, two F3 Gateways, and four F1 Gateways. If each Gateway is automatically assigned
an operational and a capacity license, you must ensure that the appropriate units are shipped to the appropriate
destinations. However, with the flexible licensing model, all you have to do is ship a Gateway to each location. When
the Gateway is at the location and installed, you can use the licensing portal to assign capacity licenses to the Gateways
installed at the other locations. You can deploy different capacity licenses (or option licenses) to different systems
without interacting directly with each system prior to deployment.

Manual License Installation

The Riverbed licensing portal provides an easy-to-use, automatic system to ensure that all your devices are up to date.
However, there are times when the Cascade appliance cannot access the portal: for example, on secure networks with
no Internet access. In these cases, you must manually access the licensing portal, retrieve the desired license keys, and
manually add them to the Cascade appliances.

Riverbed Cascade Product Family Deployment Guide 95

Licensing Automatic License Upgrades

Automatic License Upgrades

When you purchase an upgrade for a specific Cascade appliance, the licensing process is automatic. After the purchase
is approved, the licensing portal receives a new, upgraded license. The next time that appliance checks with the license
server, the new license is downloaded and installed. You do not need to remove old licenses: installed license with the
highest capacity always take precedence.

Evaluation Licenses
You can evaluate all of the licenses Cascade appliances support. Evaluation licenses are provided with specific
expiration dates, after which the license no longer works. You can test out some functionality, such as monitoring
analytics, for a specific period of time. The expiration date is displayed on the license in the licensing portion of the UI.
Because all licenses are installed with expirations, you can install the runtime license as a temporary license. When the
runtime license expires, the system stops functioning correctly. You then must install a license with a later expiration
date or no expiration.

Licenses Available
This section describes the licenses available for different Cascade appliances.

Licensing the Express

The Express (CAX-360) has three capacity licenses and one optional license. The capacity licenses provide:
 25,000 (L) flows per minute.
 50,000 (M) flows per minute.
 75,000 (H) flows per minute.
The Express 460 (CAX-460) has four capacity licenses and one optional license. The capacity licenses provide:
 15,000 (U) flows per minute
 30,000 (L) flows per minute
 60,000 (M) flows per minute
 90,000 (H) flows per minute
 120,000 (VL) flows per minute
The optional license provides 5,000 Cascade analytics and VXLAN support.

Licensing the Profiler and Profiler-VE

The Profiler (CAP-2260) has three capacity licenses and two optional licenses. The capacity licenses provide:
 150,000 (L) flows per minute.
 300,000 (M) flows per minute.

96 Riverbed Cascade Product Family Deployment Guide

Licenses Available Licensing

 600,000 (H) flows per minute.

The optional licenses provide:
 SAN support.
 7,500 Cascade analytics.
 VXLAN support.
Profiler-VE (CAP-1060-VE) has four capacity licenses and a single optional license. The capacity licenses provides:
 15,000 (VL) flows per minute.
 30,000 (L) flows per minute.
 60,000 (M) flows per minute.
 90,000 (H) flows per minute.
The optional license provide:
 3,750 Cascade analytics.
 VXLAN support.

Licensing the Enterprise Profiler Cluster

The Enterprise cluster (CAP-4260-UI, CAP-4260-DB, CAP-4260-AN) provides a base capacity license of 800,000
flows per minute. You can expand up to seven additional CAP-4260-EX modules, each providing 400,000 flows per
minute of capacity. No additional capacity licenses are required for the Enterprise cluster.
The optional licenses available provide:
 SAN support
 10,000 Cascade analytics

Licensing Gateway and Gateway-VE

The Gateway (CAG-2260) has five different capacity licenses and no optional licenses. The capacity licenses provide:
 150,000 (F1) flows per minute.
 300,000 (F2) flows per minute.
 600,000 (F3) flows per minute.
 800,000 (F4) flows per minute.
 1,400,000 (F5) flows per minute.
Gateway-VE (CAG-1060-VE) has four capacity licenses and no optional licenses. The capacity licenses provide:
 15,000 (VL) flows per minute.
 30,000 (L) flows per minute.
 60,000 (M) flows per minute.
 90,000 (H) flows per minute.

Riverbed Cascade Product Family Deployment Guide 97

Licensing Licenses Available

Licensing the Sensor and Shark

The Sensor and Shark do not have any capacity or optional licenses available. All Sensors ship with the ability to
forward their maximum number of flows per minute available.

98 Riverbed Cascade Product Family Deployment Guide

Index

A Cisco 3560 switch with Flexible

Active directory NetFlow 46
2003 80 Cisco 3750 switch with Flexible
2008 79 NetFlow 46
Additional document resources 3 Cisco 6500 series switches in hybrid
Analytics mode 44
license limits 82 Cisco 6500 series switches running native
license limits per Profiler platform 81 IOS 43
Automatic license upgrades 96 Cisco 7500 series router 45
Avaya 8300 and 5600 and IPFIX 48 Cisco 7600 series router 45
Fake index feature 67
B IPFIX for Avaya (Nortel) 8300 and
Best practices 8600 48
metrics 85 NetFlow export for Cisco Nexus 1000V 47
port mirroring 52 Nexus 7000 flexible netFlow 46
SNMP 74 sample port mirror configurations 54
tap deployment 58 sFlow for HP Procurve 3500, 5400, and
6200 49
C Steelhead appliance for flow data
Capacity licenses 94
export 69
Cascade appliances
VACL 58
flow data export 69
VACL port mirroring on Catalyst 6500 run-
in-path deployments with Steelhead
ning CatOS 58
appliances 66
VACL port mirroring on Catalyst 6500 run-
NetFlow 64
ning Cisco IOS 59
out-of-path deployments with Steelhead
Contacting Riverbed 3
appliances 69
overview 6 D
simplified routing 66 Deduplicate rate
virtual in-path deployments with Steelhead Enterprise Profiler 13
appliances 67 Express 12
Cisco Gateway 14
3560switch 46 Standard Profiler 13
375 switch 46 Deployment
6500 series switches in hybrid mode 44 considerations 66
6500 switches running native Cisco IOS 43 Express Profiler 12
7500 series router 45 Express, Shark, and Pilot 20
7600 series router 45 Gateway 14
Catalyst 6500 SPAN 54 Pilot 16
NetFlow export for Nexus 1000V 47 Profiler 12
Nexus 1000V ERSPAN to Cisco Catalyst Profiler, Gateway, Shark, and Pilot 25
6500 56 Scenarios 16
Nexus 5000 SPAN 55 Sensor 14
Nexus 7000 46 Shark 15
VACL port mirroring on Catalyst 6500 run- Shark and Pilot 16
ning CatOS 58 Shark and Sensor 20
VACL port mirroring on Catalyst 6500 run- Standard Profiler 12
ning Cisco IOS 59 Standard Profiler and Gateway 23
Configuring

Riverbed Cascade Product Family Deployment Guide 99

Index

Document conventions, overview of 2 capacity license 94

Documentation, contacting 4 current generation hardware 94
Documents evaluation 96
additional resources 3 installation 94
limits per Profiler platform 81
E manual installation 95
Embedded Cascade Shark 7 option 94
Enterprise other device license installation 95
model options 13 runtime licenses 93
ERSPAN 54 Limits, analytic license 82
Evaluation licenses 96 Local storage 32
Express
integrated deployment 19 M
model options 12 Metrics
standalone deployment 18 best practices 85
reducing the number 83
F which to use 85
Fake index, configuring 67 Model options
Flow collection Enterprise Profiler 13
base requirements 37 Express 12
considerations 40 Gateway 14
validation 41 Sensor 15
virtual environments 41 Shark 15
Flow data export Shark-VE 15
Cascade appliance 69 Standard Profiler 13
Flow data fields
consumed by Cascade appliances 39 N
Flow rate NetFlow
estimation 32 Cascade appliances 64
Flow source Network tap 57
SNMP integration 73
Flow storage 31 P
estimation 33 Packet collection
types 31 Cascade components 51
Flow type Packet deduplication 60
considerations 40 Pilot
Flows per minute overview 7
average 11 Port dependencies
estimated maximum 11 enterprise solution 31
full solution 30
G Pilot and Shark 28
Gateway Profiler and Gateway 29
model options 14 Port mirroring 52
overview 7 best practices 52
Gateway-VE, overview 7 Professional services, contacting 4
Profile
H REST API 80
HP Procurve Profiler
sFlow 49 compatibility with RiOS 64
Enterprise models 13
I Express models 12
ifindex 65 overview 6
In-path deployments with Steelhead Standard models 13
appliances 66 Profiler-VE
Installation overview 6
license 94
other device license 95 R
Interceptor appliance 60 Related reading 3
considerations 67 Release notes 3
virtual in-path deployments 67 REPAN 53
REST API 80
L RiOS compatibility with Profiler 64
License Riverbed, contacting 3
assigning 95
automatic upgrades 96
available 96

100 Riverbed Cascade Product Family Deployment Guide

 Index

S
SAN 32
Sensor
model options 15
overview 7
protocol violations 90
Sensor-VE
overview 7
Shark
model options 15
overview 7
passthru 60
REST API 80
virtual 7
Shark-VE
model options 15
overview 7
Simplified routing
Cascade appliances 66
Snaplen 61
SNMP
best practices 74
switch port discovery 75
SNMP integration
device management of Cascade
components 74
flow sources 73
sending traps 74
SNMP interface persistence 65
SPAN 52
Staffing needs 11
Standard
model options 13
Steelhead appliance 63
flow data export 69
in-path deployment 66
Storage
flow 31
local 32
Subnet side rules 67
Switch port discovery
SNMP 75
supported routers and switches 76
troubleshooting 91

T
Tap deployment
best practices 58
Technical publications, contacting 4
Technical staffing needs 11
Technical support, contacting 4

V
VACL
configuration examples 58
port mirroring configuration on Catalyst 6500
running Cisco IOS 59
port mirroring configuration on Cisco 6500
Running CatOS 58
VSP 63

Riverbed Cascade Product Family Deployment Guide 101

Index

102 Riverbed Cascade Product Family Deployment Guide