Network Deployments r7-3 Revm November2015 0
Network Deployments r7-3 Revm November2015 0
Copyright © 2015 Silver Peak Systems, Inc. All rights reserved. Information in this document is subject to change at any time. Use of
this documentation is restricted as specified in the End User License Agreement. No part of this documentation can be reproduced,
except as noted in the End User License Agreement, in whole or in part, without the written consent of Silver Peak Systems, Inc.
Trademark Notification
The following are trademarks of Silver Peak Systems, Inc.: Silver Peak SystemsTM, the Silver Peak logo, Network MemoryTM, Silver
Peak NX-SeriesTM, Silver Peak VX-SeriesTM, Silver Peak VRX-SeriesTM, Silver Peak Unity EdgeConnectTM, and Silver Peak
OrchestratorTM. All trademark rights reserved. All other brand or product names are trademarks or registered trademarks of their
respective companies or organizations.
THIS DOCUMENTATION IS PROVIDED “AS IS” WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED,
INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
PURPOSE, OR NON-INFRINGEMENT. SILVER PEAK SYSTEMS, INC. ASSUMES NO RESPONSIBILITY FOR ERRORS OR
OMISSIONS IN THIS DOCUMENTATION OR OTHER DOCUMENTS WHICH ARE REFERENCED BY OR LINKED TO THIS
DOCUMENTATION. REFERENCES TO CORPORATIONS, THEIR SERVICES AND PRODUCTS, ARE PROVIDED “AS IS”
WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESSED OR IMPLIED. IN NO EVENT SHALL SILVER PEAK SYSTEMS, INC.
BE LIABLE FOR ANY SPECIAL, INCIDENTAL, INDIRECT OR CONSEQUENTIAL DAMAGES OF ANY KIND, OR ANY DAMAGES
WHATSOEVER, INCLUDING, WITHOUT LIMITATION, THOSE RESULTING FROM LOSS OF USE, DATA OR PROFITS,
WHETHER OR NOT ADVISED OF THE POSSIBILITY OF DAMAGE, AND ON ANY THEORY OF LIABILITY, ARISING OUT OF OR
IN CONNECTION WITH THE USE OF THIS DOCUMENTATION. THIS DOCUMENTATION MAY INCLUDE TECHNICAL OR OTHER
INACCURACIES OR TYPOGRAPHICAL ERRORS. CHANGES ARE PERIODICALLY ADDED TO THE INFORMATION HEREIN;
THESE CHANGES WILL BE INCORPORATED IN NEW EDITIONS OF THE DOCUMENTATION. SILVER PEAK SYSTEMS, INC.
MAY MAKE IMPROVEMENTS AND/OR CHANGES IN THE PRODUCT(S) AND/OR THE PROGRAM(S) DESCRIBED IN THIS
DOCUMENTATION AT ANY TIME.
http://www.silver-peak.com/support
ii PN 200059-001 Rev M
Contents
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Who Should Read This Manual?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Manual Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
Related Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
Technical Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
Network Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Summary of Initial Configuration Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Collecting the Necessary Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
Using the Initial Config Wizard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Verifying Appliance Connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Enabling Subnet Sharing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Creating Tunnels and Updating the Subnet Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Configuring the Router to Redirect Traffic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Using a Cisco Router for Policy-Based Routing (PBR). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Using a Juniper Router for Filter-Based Forwarding (FBF). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Verifying Traffic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
iv PN 200059-001 Rev M
Contents
Chapter 6 Out-of-Path with PBR and VRRP Redundant Silver Peak Appliances . . . . . . . . 157
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Network Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Collecting the Necessary Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
Summary of Configuration Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Using the Initial Config Wizard for Site A. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
Configuring VRRP on A1 and A2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Using VRRP with two Silver Peaks acting as Master and Backup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
Configuring Flow Redirection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
Using the Initial Config Wizard with Site B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182
Verifying Appliance Connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
Enabling Subnet Sharing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Creating Tunnels and Updating the Subnet Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Configuring A1 and A2 to Advertise Non-Local Subnets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
Configuring the Cisco Router for Policy-Based Routing (PBR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
Verifying Traffic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
PN 200059-001 Rev M v
Silver Peak NX Series Appliances Network Deployment Guide
vi PN 200059-001 Rev M
Preface
Welcome to the Silver Peak Network Deployment Guide. Read the Preface to understand the target
audience, the manual’s organization, related documents, and how to contact Customer Support.
Most deployments in this guide focus on using subnet sharing as the auto-optimization method for
routing flows. A smaller number demonstrate the use of TCP-based and IP-based auto-optimization.
Manual Organization
This section outlines the chapters and summarizes their content.
To keep things simple, we illustrate the examples with the typical in-line deployment in Site B offices
and out-of-path deployment at Site A. However, Site B offices are not restricted to in-line deployment,
nor is Site A restricted to out-of-path deployments.
Chapter 1, “Fundamentals of Deploying WAN Optimization,” describes some of the fundamental
concepts of deploying WAN acceleration in enterprise networks. It provides an overview and
introduction to common installation models, pros and cons of each, and recommendations.
Chapter 2, “In-Line Deployment,” describes the procedures for an in-line deployment where the Silver
Peak Appliance sits between the WAN router and the Ethernet switch.
Chapter 3, “Out-of-Path with Policy-Based-Routing Redirection,”, describes the procedures for a
scenario that deploys the Site B location in-line and the Site A network out-of-path with an available
spare router port and uses Policy-Based Routing (PBR) on the WAN router to redirect traffic to the Silver
Peak appliance.
Chapter 4, “Out-of-Path with WCCP,” (Comparing Subnet Sharing & TCP/IP-based
Auto-Optimization), describes the procedures for setting up Web Cache Communications Protocol
(WCCP) service. The example uses a Cisco router paired with a single Silver Peak appliance deployed
out-of-path (Router mode). It also highlights the differences in traffic redirection required when using
subnet sharing, as opposed to TCP-based or IP-based auto-optimization.
Chapter 5, “Out-of-Path with VRRP Peering to a WAN Router,” describes the procedures for a scenario
where the Silver Peak appliance uses the Virtual Router Redundancy Protocol (VRRP) to peer with the
existing router, when no spare router port is available.
Chapter 6, “Out-of-Path with PBR and VRRP Redundant Silver Peak Appliances,” describes the
procedures for setting up high availability. In this example, Site A deploys a primary and a secondary
appliance out-of-path (Router mode), and the Site B location deploys the appliance in-line (Bridge
mode). Additionally, the peered Site A appliances use the Virtual Router Redundancy Protocol (VRRP)
to create and share a common IP address, called the Virtual IP address (VIP).
Chapter 7, “Out-of-Path with WCCP Redundant (Active/Active) Appliances,” describes the procedures
for setting up high availability by using Web Cache Communications Protocol with a Cisco router and
redundant Silver Peak appliances in an out-of-path deployment.
Related Publications
Release Notes provide information on new software features, system bugs, and software
compatibility.
All user documentation is also available for download from http://www.silver-peak.com.
Technical Support
For product and technical support, contact Silver Peak Systems at either of the following:
We’re dedicated to continually improving the usability of our products and documentation.
If you have suggestions or feedback for our documentation, please send an e-mail to
[email protected].
If you have comments or feedback about the GUI’s ease of use, please send an e-mail to
[email protected].
This chapter describes some of the fundamental concepts of deploying WAN acceleration in enterprise
networks.
In This Chapter
Introduction See page 2.
PN 200059-001 Rev M 1
Silver Peak NX Series Appliances Network Deployment Guide Introduction
Introduction
Deploying WAN optimization in an enterprise network is similar to deploying other enterprise
networking technologies (for example, firewalls). There are, however, a few tenets to keep in mind:
1 Silver Peak appliances need to have visibility into any traffic that requires optimization.
As such, all traffic to be optimized must flow though the appliances. There are three ways to
accomplish this:
Server mode In this default configuration, the management path and the
[default] datapath both use the same interface and the same IP address.
2 Silver Peak WAN acceleration is a “symmetric” solution. That is, to optimize the traffic on the link,
Silver Peak appliances are required on both ends of the WAN link.
3 Silver Peak’s Network Acceleration functions require that the appliances have visibility into both
the transmit and receive directions of a flow. If not, the flow is considered “asymmetric” and
Network Acceleration will be defeated although Network Integrity and Network Memory will
continue to provide benefit.
Definition of Terms
Following are the definitions for common terms used throughout the guide, listed alphabetically:
Term Definition
Acceleration Refers to techniques used to improve transmission of TCP protocols across a WAN. a
TCP Proxy session is created to reduce the impact of latency on a TCP flow.
Techniques such as local acknowledgements and window sizing are used to
accelerate TCP traffic.
Auto Optimized Traffic IP traffic that is automatically recognized by the Silver Peak appliances and
optimized accordingly, without the need for manually created Route Policies. This
is the default entry for the Route Policy if no entries are made, or for the last line in
the route map.
Bypass Bypass refers to hardware bypass. If there is a major problem with the appliance
hardware, software, or power, all traffic goes through the appliance without any
processing. Bypass mode can be enabled manually.
Silver Peak appliances can be installed in the data path (in-line) between an L2/L3
switch and the edge WAN router, with fail-to-wire in case of failure.
Bypass mode and Hardware Bypass both refer to the failover method, which is
Fail-to-Wire for copper interfaces, and Fail-to-Glass for fiber interfaces.
2 PN 200059-001 Rev M
Introduction Chapter 1 Fundamentals of Deploying WAN Optimization
Term Definition
Data Path IP Address Generally, an IP address of an interface through which end-device traffic flows or to
which it is redirected.
• If the device is out-of-path in Server mode, the data path and management path IPs
are the same.
• In Router mode, with an out-of-band management interface, the management IP
and Data Path address are different.
• In Bridge mode, the Data Path IP is separate from the management IP.
In some deployments—like DHRM (Dual Home Router Mode) or multiple VLANS —
there could be multiple data path IP addresses.
Network MemoryTM Silver Peak's innovative approach to data reduction that leverages advanced pattern
recognition and local information
Optimization A collection of techniques that accelerate, compress, and improve the efficiency of
transmission of data across a WAN. Optimization includes acceleration techniques,
data reduction, forward error correction, packet order correction, QoS, and other
techniques.
Pass-through Traffic By default, traffic that is not directed to a tunnel by the Route Policy passes
transparently through the Silver Peak appliance. Pass-through traffic can be either
shaped or unshaped.
Route Policy Uses MATCH criteria to delineate flows and SET actions to specify how to handle that
flow. For example, a Route Policy entry would direct a specific flow to a designated
tunnel.
Tunnel Provide virtual point-to-point links between two application acceleration devices. They
work by wrapping original packets of data inside an outer IP header, which is used to
specify the address of the device on the far end of the WAN link.
PN 200059-001 Rev M 3
Silver Peak NX Series Appliances Network Deployment Guide Using Physical and Virtual Appliances
lan0 This interface is intended for connection to the LAN side of the network.
lan1 This interface is intended for connection to the LAN side of the network.
wan0 This interface is intended for connection to the WAN side of the network.
wan1 This interface is intended for connection to the WAN side of the network.
tlan0 This fiber interface is intended for connection to the LAN side of the network.
twan0 This fiber interface is intended for connection to the WAN side of the network.
mgmt0 This interface is intended for network access to the appliance’s management
interfaces (the Web-based Appliance Manager and the Command Line Interface). It is
recommended that this interface is always connected to the network.
The mgmt0 next-hop IP address points to a Level 3 (L3) switch or router.
mgmt1 This interface is intended for local access to the appliance’s management interfaces
(the Web-based Appliance Manager and the Command Line Interface) with a laptop.
The mgmt1 interface may sometimes be used for flow redirection. For more
information, see the Silver Peak Appliance Manager Operator’s Guide.
If you are using out-of-band management with Router mode (as opposed to using Server mode), then
each physical Silver Peak appliance requires two IP addresses on the network. These IP addresses are
described in the following table.
Appliance IP Address The IP address originates and terminates the tunnels used to interconnect Silver Peak
appliances.
Management IP Address This IP address is used for management and configuration of the Silver Peak appliance
(mgmt0) via the web-based Appliance Manager.
4 PN 200059-001 Rev M
Using Physical and Virtual Appliances Chapter 1 Fundamentals of Deploying WAN Optimization
Although it isn’t a requirement, it’s considered a best practice to use different subnets for mgmt0 and the
Appliance data path IP.
The physical (NX) and virtual appliance Quick Start Guides each explain how to access and configure
the mgmt0 interface. Here, we offer a quick, generic review.
2 The next task is to determine the virtual appliance’s mgmt0 IP address. In a browser, this address
provides access to the Appliance Manager.
• If you’re using DHCP, the virtual appliance IP address displays in Silver Peak’s Console User
Interface.
• If you’re not using DHCP, then you must configure the static IP address and default gateway.
Continue with the following steps.
PN 200059-001 Rev M 5
Silver Peak NX Series Appliances Network Deployment Guide Using Physical and Virtual Appliances
3 In the virtual appliance console, press the function key, F4, select Static, and press Enter.
4 Enter the IP addresses for the mgmt0 interface and default gateway.
6 PN 200059-001 Rev M
Using Physical and Virtual Appliances Chapter 1 Fundamentals of Deploying WAN Optimization
7 To verify connectivity, press function key, F1, and enter the following command sequence:
PN 200059-001 Rev M 7
Silver Peak NX Series Appliances Network Deployment Guide Choosing an Optimization Strategy for the Traffic Path
Auto-optimization uses different mechanisms for TCP versus non-TCP traffic. Because both
mechanisms ultimately require an exchange of packets between two appliances, unidirectional IP
traffic will not trigger auto-optimization.
Auto-opt may not work with a firewall in the path. Some firewalls may be configured to strip out or
block the TCP options in the initial SYN packet, which will break auto-optimization. Subnet sharing
does not use the TCP options field, and thus avoids this issue. Therefore, use of subnet sharing is a
recommended best practice.
You can, if you choose, modify the default entry’s SET action of auto-optimized.
The Route Policy, then, only requires manual entries for flows that are to be:
• sent pass-through (shaped or unshaped)
• dropped
• configured for a specific high-availability deployment.
• routed based on application, VLAN, DSCP, or ACL (Access Control List)
You can, however, choose to forego auto-optimization and create any and all route policies manually.
Note IMPORTANT — A tunnel must exist before subnet sharing can proceed.
Using Appliance Manager, you can create tunnels in one of three ways:
If you enable auto-tunnel on the Configuration - System page, then the initial TCP-based or
IP-based handshaking creates the tunnel. That requires outbound and inbound redirection to be in
place.
You can let the Initial Configuration Wizard create the tunnel to the remote appliance.
You can create a tunnel manually on the Configuration - Tunnels page.
8 PN 200059-001 Rev M
Configuring Dynamic Path Control Chapter 1 Fundamentals of Deploying WAN Optimization
Although we’ve used CIFs, SSH, and FTP traffic to illustrate, the steps work with any application. Note
that these examples are not intended to be recommendations about how to handle that specific traffic in
your network.
PN 200059-001 Rev M 9
Silver Peak NX Series Appliances Network Deployment Guide Configuring Dynamic Path Control
After you’ve created your entries, click Apply, and then click Save Changes.
10 PN 200059-001 Rev M
Configuring Dynamic Path Control Chapter 1 Fundamentals of Deploying WAN Optimization
Note The example below uses CIFS, but the steps work for any application.
PN 200059-001 Rev M 11
Silver Peak NX Series Appliances Network Deployment Guide WAN Hardening
WAN Hardening
WAN hardening is an option that provides additional protection against unsafe connections from remote
sites. When WAN hardening is enabled, only traffic arriving from a Silver Peak IPsec tunnels is allowed
to enter.
When Silver Peak appliances are deployed in Router Mode, you have the option of hardening any
WAN–side interface. This means:
For traffic inbound from the WAN, the appliance only accepts encapsulated traffic arriving from
another Silver Peak appliance, via an IPsec tunnel. All other connections are rejected.
For traffic outbound to the WAN, the appliance only allows IPsec tunnel packets and management
traffic.
Any data from the internet that gets backhauled via a Silver Peak IPSec tunnel will reach its
destination at the hardened sites. This allows for integration with other security tools, such as
firewalls, at the data center.
Data sourced directly from the internet, or any other connection that doesn’t flow through a Silver
Peak IPsec tunnel, is discarded when it hits the hardened interface. Only data from authenticated
Silver Peak IPsec tunnels is allowed to pass across a hardened interface.
12 PN 200059-001 Rev M
Determining the Need for Traffic Redirection Chapter 1 Fundamentals of Deploying WAN Optimization
How you plan to optimize traffic affects whether or not you also need inbound redirection from the
WAN router (also known as WAN-side redirection):
• If you enable subnet sharing (which relies on advertising local subnets between Silver Peak
appliances) or route policies (which specify destination IP addresses), then you only need
outbound redirection.
If you have subnets that are not directly attached to the Silver Peak, you may need to manually
add those subnets so the local appliance can advertise them to its peers. If those subnets are not
reachable via the default LAN–side next-hop router, then you may also need to add a static route
to the local Silver Peak, specifying which next-hop router to use to reach a given subnet. For an
example, see “Configuring A1 and A2 to Advertise Non-Local Subnets” on page 196.
• If, instead, you default to TCP-based or IP-based auto-optimization (which relies on initial
handshaking outside a tunnel), then you must set up inbound and outbound redirection on the
WAN router.
• Additionally, for TCP flows to be optimized, both directions must travel through the same client
and server appliances. If the TCP flows are asymmetric —as could occur in a high-availability
deployment — you need to configure clusters for flow redirection among local appliances.
For more about flow redirection, refer to the Appliance Manager Operator’s Guide.
The following diagrams show where redirection is required and which methods you can use:
• when subnet sharing is enabled
• when using TCP-based or IP-based auto-optimization (that is, subnet sharing is not enabled)
• when directed to a specific tunnel by the Route Policy
PN 200059-001 Rev M 13
Silver Peak NX Series Appliances Network Deployment Guide Determining the Need for Traffic Redirection
Figure 1-1
14 PN 200059-001 Rev M
Determining the Need for Traffic Redirection Chapter 1 Fundamentals of Deploying WAN Optimization
Figure 1-2
PN 200059-001 Rev M 15
Silver Peak NX Series Appliances Network Deployment Guide Determining the Need for Traffic Redirection
Figure 1-3
16 PN 200059-001 Rev M
High Availability Chapter 1 Fundamentals of Deploying WAN Optimization
High Availability
In High Availability (HA) configurations, the redundant Silver Peak appliances are deployed in router
mode and either WCCP or PBR redirects flows from the routers to the appliances.
The redundant appliances may be configured Active/Active or Active/Backup. This is determined by
how the WCCP or PBR redirection is configured on the routers and the appliances.
For the purposes of discussion, we’ll assume that HA is configured in the same location as the servers
and we’ll refer to the HA (redundant) appliances as “server-side”. We’ll refer to the non-redundant
appliances as “client-side”. Of course, it doesn’t need to be this way—it’s possible to have redundant
Silver Peak appliances in offices without servers.
Asymmetry Mitigation
Flow redirection can prevent TCP asymmetry in high availability environments.
For the appliances, this requires configuring HA (or redundant) appliances as peers, and enabling flow
redirection. Both tasks are on the Configuration - Flow Redirection screen.
Where it’s an element of any deployment chapter in this guide, the instructions include the configuration
steps.
PN 200059-001 Rev M 17
Silver Peak NX Series Appliances Network Deployment Guide Considerations for Deployments
18 PN 200059-001 Rev M
Verifying Connectivity After Configuring Deployment Chapter 1 Fundamentals of Deploying WAN Optimization
ping
ping is a good general tool to verify reachability. However, it is not the best tool to use to verify correct
deployment of WAN optimization appliances because:
1 It doesn’t verify the path that traffic takes.
It’s important to verify the path, not just reachability, because the appliance must intercept traffic on
both sides of the WAN for optimization and acceleration to be effective.
2 It relies on ICMP, and some redirection methods (for example, WCCP) don’t support ICMP.
You need a tool that can verify paths by revealing all hops taken along a path. Some tools you can use to
verify the paths taken are ping -r and traceroute.
Linux/Silver Peak ping -R Record Route. Includes the RECORD_ROUTE option in the
ECHO_REQUEST packet and displays the route buffer on
returned packets.
Note that the IP header is only large enough for nine such
routes. Some hosts ignore or discard this option.
Pros
• Most (but not guaranteed all) network devices support it, whether they are routers or not.
• Shows the return path, too.
Cons
• Limited to nine devices in the traffic path, including the source and destination.
• ping -r may fail to verify connectivity with some WCCP deployments.
traceroute
Windows and Unix each have slightly different versions.
• Both are suitable for non-WCCP deployments, but because Windows traceroute uses ICMP, it
isn’t suitable for WCCP.
• For WCCP deployments, you need to use Unix traceroute or a 3rd-party Windows traceroute
that uses UDP instead of ICMP.
• The downside of traceroute is that only router hops display.
PN 200059-001 Rev M 19
Silver Peak NX Series Appliances Network Deployment Guide Verifying Connectivity After Configuring Deployment
Basic procedure
1 Verify connectivity for optimized traffic.
• In Router mode (out-of-path deployment), Silver Peak appliances look like router hops. They’ll
display in both ping -r and traceroute.
• In Bridge mode (in-line deployment), Silver Peak appliances look like bridges. They’ll display
in ping -r, but not in traceroute.
3 Test network connectivity by using your applications. For example, do a CIFS mount or an FTP
transfer.
20 PN 200059-001 Rev M
CHAPTER 2
In-Line Deployment
Using Subnet Sharing
In this deployment scenario, the Silver Peak Appliance sits between the WAN router and the Ethernet
switch.
In This Chapter
Overview See page 22.
PN 200059-001 Rev M 21
Silver Peak NX Series Appliances Network Deployment Guide Overview
Overview
In an in-line deployment, the Silver Peak appliance is inserted in-line between the WAN router and the
Ethernet switch on the LAN side of the network. In this mode, the appliance intercepts all packets
destined for the WAN. Based on the Route Policy’s MATCH criteria, or using Subnet Sharing–enabled
auto-optimization, the appliance optimizes all flows that are directed to a tunnel. All other traffic passes
through the appliance without optimization.
When the appliance fails, it behaves as if it were a crossover cable. Best practice is to use a crossover
cable between the appliance and the WAN–side router, and a standard ethernet cable between the
appliance and the LAN–side switch. Verify the physical layer connectivity between the L2 switch and
router with the appliance turned off. If you don’t receive link on the router and/or switch, you’ll need to
correct the cabling.
Network Diagram
Summary
Appliance Placement Appliance placed in-line between Ethernet LAN switch and WAN router
• Appliance lan0 interface connects to Ethernet LAN switch
• Appliance wan0 interface connects to WAN router
Fail-Safe Behavior • Fails-to-Wire: The appliance behaves as a crossover cable between the Ethernet
LAN switch and the WAN router in any failure scenario (hardware, software, power).
• IMPORTANT: Ensure that the Ethernet LAN’s switch and the WAN router have
compatible Ethernet interface physical configuration settings (speed and duplex
settings can be found on the Configuration > Interfaces page). This is to ensure
that traffic flows correctly if the Silver Peak appliance “Fails-to-wire”.
IP Addresses This deployment model requires two IP addresses (on the same or separate subnets)
• Silver Peak Appliance data path IP address (to originate and terminate tunnel)
• Silver Peak Management IP Address (for appliance configuration and
management)
22 PN 200059-001 Rev M
Overview Chapter 2 In-Line Deployment
1 Gather all the IP addresses Saves time and avoids mistakes. “Collecting the Necessary
needed for setup Information” on page 23.
2 Install the appliance into the Physical appliance: Connect each site’s Silver Peak Appliance Manager
network appliance between its WAN edge router and Operator’s Guide
Ethernet switch. Verify connectivity, connect power,
Quick Start Guides
and verify LEDs.
Virtual appliance: Configure the hypervisor, with
the required interfaces.
3 Configure the appliances In a browser, access and use the Initial “Using the Initial Config Wizard” on
Configuration Wizard to configure each appliance page 25.
in Bridge mode.
4 Verify appliance connectivity Tests data path connectivity. “Verifying Appliance Connectivity”
Do NOT proceed until you verify connectivity. on page 32.
5 Create a tunnel on each Specify the local and remote endpoints for the “Creating Tunnels” on page 33.
appliance tunnel.
6 Test the connectivity from both Verify that the tunnel is up and that flows are being “Verifying Traffic” on page 35.
ends optimized.
Hostname B C
Mode In-line (Bridge) In-line (Bridge)
Admin Password: Old admin admin
Admin Password: New / Confirm
mgmt1 IP Address / Mask --- ---
Time Zone
NTP Server IP Address
License (for virtual appliance only)
mgmt0 IP Address / Maska 192.168.1.9/24 192.168.1.5/24
mgmt0 Next-hop IP Address 192.168.1.1 192.168.1.1
LAN Next-hop IP Address (optional) b --- ---
PN 200059-001 Rev M 23
Silver Peak NX Series Appliances Network Deployment Guide Overview
Hostname B C
Appliance data path IP Address / Mask 10.110.11.100/24 10.110.21.100/24
Appliance data path Next-hop IP 10.110.11.1/24 10.110.21.1/24
a. In this example, all mgmt0 IP addresses are in the same subnet. In your actual network, it’s likely that mgmt0 IP
addresses are in different subnets.
b. LAN next-hop IP is only required when there are subnets for which the Silver Peak appliance does not have a
configured IP address.
24 PN 200059-001 Rev M
Using the Initial Config Wizard Chapter 2 In-Line Deployment
2 For the username and for the password, enter admin. The initial configuration page appears.
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
PN 200059-001 Rev M 25
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard
3 Read it, and click Next. Select the MAC addresses for lan0 and wan0. Make sure that the addresses
match the MAC addresses associated with the virtual interfaces of the Silver Peak virtual machine.
26 PN 200059-001 Rev M
Using the Initial Config Wizard Chapter 2 In-Line Deployment
For example, in the VMware client, you would check on the Virtual Machine Properties page.
PN 200059-001 Rev M 27
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard
4 Click Apply & Next. The License & Registration page appears. Enter the license details.
5 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
28 PN 200059-001 Rev M
Using the Initial Config Wizard Chapter 2 In-Line Deployment
6 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Bridge and
configure the appliance data path IP next-hop router address and max WAN bandwidth.
a Leave Auto Tunnel and Auto Subnet Sharing deselected. We’ll take care of these features later.
Although it’s not technically necessary to deselect either one, we have chosen to do so for
tutorial purposes later in this chapter.
b Do not add tunnels. We’ll manually add remote appliances and create tunnels later.
PN 200059-001 Rev M 29
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard
8 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
9 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
30 PN 200059-001 Rev M
Using the Initial Config Wizard Chapter 2 In-Line Deployment
Click Done. The appliance saves the settings and reboots automatically.
11 Repeat the installation process for Appliance B, following the same procedure as you did for
Appliance C.
PN 200059-001 Rev M 31
Silver Peak NX Series Appliances Network Deployment Guide Verifying Appliance Connectivity
Tip Prior to putting a bridge mode appliance in production, it is always a good practice to test
connectivity with the appliance in bypass to make sure that the network will function in the event
the Silver Peak device fails to wire.
32 PN 200059-001 Rev M
Creating Tunnels Chapter 2 In-Line Deployment
Creating Tunnels
Create a tunnel between appliances B and C. This involves accessing each appliance, in turn, and creating
a tunnel to the other (remote) appliance.
a In the Name field, assign a locally significant name for the tunnel.
b In the Admin field, accept the default value, Up.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager prefills the IP address for the local appliance.
e In the Remote IP address field, enter the data path IP address of the remote Silver Peak
appliance.
f Leave Auto Max BW selected, so the appliance uses the lower of the two system bandwidths.
If you wanted to configure this manually, then you would deselect Auto Max BW and, in the
Max BW field, enter the maximum bandwidth for this tunnel. The value must be less than or
equal to the upstream bandwidth of your WAN connection.
g Leave the Min BW at its default, 32 [Kbps].
h Click Apply.
i Save the changes.
PN 200059-001 Rev M 33
Silver Peak NX Series Appliances Network Deployment Guide Creating Tunnels
a In the Name field, assign a locally significant name for the tunnel.
b In the Admin field, accept the default value, Up.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager prefills the IP address for the local appliance.
e In the Remote IP address field, enter the data path IP address of the remote Silver Peak
appliance.
f Leave Auto Max BW selected, so the appliance uses the lower of the two system bandwidths.
g Leave the Min BW at its default, 32 [Kbps].
h Click Apply.
i Save the changes.
Within a few seconds, the tunnel Status changes to Up - active. Click Refresh, if required.
Be aware that a tunnel doesn’t come up unless it’s configured on both ends. Configuring a tunnel on a
single device will not cause a connection to come up.
34 PN 200059-001 Rev M
Verifying Traffic Chapter 2 In-Line Deployment
Verifying Traffic
Subnet sharing enables Silver Peak devices that are connected by tunnels to automatically share subnet
information and direct all IP traffic to the appropriate destinations.
1 Verify that each appliance is learning subnets from the other appliance.
a At each appliance, access Configuration > Subnets.
b Verify that local subnets are being advertised to peers.
c Verify that the subnet table lists subnets learned from the remote appliance.
The local appliance uses this learned subnet information. When auto optimization is enabled
(this is the default Route Policy, and it hasn’t been changed in this example), LAN-to-WAN
flows are examined for the destination address. If the destination address matches a subnet
learned by the local appliance, the flow is routed into the tunnel that terminates at the Silver
Peak advertising the subnet.
PN 200059-001 Rev M 35
Silver Peak NX Series Appliances Network Deployment Guide Verifying Traffic
In this example, the Outbound Tunnel is the one connecting the two Silver Peak appliances.
The Inbound and Outbound sections provide basic Clicking the icon in the Detail
statistical information associated with the flow. column provides additional
information for as long as the
flow is active.
36 PN 200059-001 Rev M
Verifying Traffic Chapter 2 In-Line Deployment
Note that the flow Status is OPTIMIZED. This is the desired status.
If the Status is ALERT, click on ALERT for a pop-up that provides a troubleshooting hint.
Note that in this case, one end of the tunnel was set administratively down, so packets could not
be properly routed.
4 Test network connectivity by using your applications. For example, do a CIFS mount or an FTP
transfer.
PN 200059-001 Rev M 37
Silver Peak NX Series Appliances Network Deployment Guide Verifying Traffic
38 PN 200059-001 Rev M
CHAPTER 3
Note If you’re using a Juniper router, their equivalent term for this redirection method is
Filter-Based Forwarding [FBF]. Check your router manufacturer’s documentation to verify
terminology.
In This Chapter
SECTION 1: Using Subnet Sharing See page 40.
PN 200059-001 Rev M 39
Silver Peak NX Series Appliances Network Deployment Guide
Creating Tunnels and Updating the Subnet Table See page 60.
40 PN 200059-001 Rev M
Overview Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
Overview
This scenario deploys Site B in-line and the Site A network out-of-path with an available spare router
port. It uses Policy-Based Routing (PBR) at the router to redirect traffic destined for the WAN to the
Silver Peak appliance.
Network Diagram
Figure 3-1 Out-of-Path Deployment with Policy-Based Routing (PBR): Router Mode [Spare Router Port Available]
In this example, the Silver Peak appliance optimizes traffic to/from 10.110.33.0/24 and 10.110.11.0/24.
Summary
IP Addresses This deployment model requires two IP addresses (on the same or separate subnets):
• Silver Peak Appliance data path IP address (to originate and terminate tunnel)
• Silver Peak Management IP Address (for appliance configuration and
management)
PN 200059-001 Rev M 41
Silver Peak NX Series Appliances Network Deployment Guide Overview
Fail-Safe Behavior
Fail-safe behavior should always be tested before production deployment by ensuring that traffic
continues to flow in each of the following cases:
1 With the appliance in bypass state
2 With the appliance powered off
3 With the tunnels administratively down.
1 Gather all the IP addresses Saves time and avoids mistakes. “Collecting the Necessary
needed for setup Information” on page 43.
2 Install the appliances Physical appliance: Connect the Site A appliance Silver Peak Appliance Manager
to the Site A router, and insert the Site B appliance Operator’s Guide
between its WAN edge router and the Ethernet
Quick Start Guides
switch. Verify connectivity, connect power, and
verify LEDs.
Virtual appliance: Configure the hypervisor, with
the required interfaces.
3 Configure the appliance In a browser, access and use the Initial “Using the Initial Config Wizard” on
Configuration Wizard to configure each appliance page 44.
— one in Bridge mode, the other in Router mode.
Reboot the appliance after finishing the
configuration.
4 Verify appliance connectivity Tests data path connectivity. “Verifying Appliance Connectivity”
on page 56.
Do NOT proceed until you verify connectivity.
5 Enable subnet sharing This prepares each appliance to share local “Enabling Subnet Sharing” on
subnets. page 58.
6 Create a tunnel on each Specify the local and remote endpoints for the “Creating Tunnels and Updating
appliance tunnel. the Subnet Table” on page 60.
7 Configure the router Access the router’s command line interface, and “Configuring the Router to Redirect
configure the router for policy-based routing. Traffic” on page 64.
8 Test the connectivity from both Verify that the tunnel is up and that flows are being “Verifying Traffic” on page 67.
ends optimized.
42 PN 200059-001 Rev M
Overview Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
Table 3-1 Out-of-Path Deployment with Policy-Based Routing (PBR): Router Mode [Spare Router Port Available]
Hostname A B
Mode Out-of-Path (Router) In-line (Bridge)
Admin Password: Old admin admin
Admin Password: New / Confirm
Time Zone
NTP Server IP Address
License (for virtual appliance only)
mgmt0 IP Address / Maska 192.168.1.7/24 192.168.1.9/24
mgmt0 Next-hop IP Address 192.168.1.1 192.168.1.1
Appliance data path IP Address / Mask 10.110.31.100/24 10.110.11.100/24
Appliance data path Next-hop IP 10.110.31.1/24 10.110.11.1/24
b
LAN Next-hop IP Address (optional) not applicable ---
a. In this example, all mgmt0 IP addresses are in the same subnet. In your actual network, it’s likely that mgmt0 IP
addresses are in different subnets.
b. LAN next-hop IP is only required when there are subnets for which the Silver Peak appliance does not have a
configured IP address.
PN 200059-001 Rev M 43
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard
Appliance A
2 For the username and for the password, enter admin. The initial configuration page appears.
Read it, and click Next.
44 PN 200059-001 Rev M
Using the Initial Config Wizard Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
3 Select the MAC address for wan0. Make sure that the addresses match the MAC addresses
associated with the virtual interfaces of the Silver Peak virtual machine (VM).
For example, in the VMware client, you would check on the Virtual Machine Properties page.
PN 200059-001 Rev M 45
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard
4 Click Apply & Next. The License & Registration page appears. Enter the license details.
5 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
46 PN 200059-001 Rev M
Using the Initial Config Wizard Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
6 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Router and
then click +Add to add a WAN interface.
Configure the appliance data path IP next-hop router address and max WAN bandwidth.
PN 200059-001 Rev M 47
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard
8 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
48 PN 200059-001 Rev M
Using the Initial Config Wizard Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
9 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
11 Click Done. The appliance saves the settings and reboots automatically.
PN 200059-001 Rev M 49
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard
Appliance B
12 For the username and for the password, enter admin. The initial configuration page appears.
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
13 Read it, and click Next. Select the MAC addresses for lan0 and wan0. Make sure that the addresses
match the MAC addresses associated with the virtual interfaces of the Silver Peak virtual machine.
50 PN 200059-001 Rev M
Using the Initial Config Wizard Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
For example, in the VMware client, you would check on the Virtual Machine Properties page.
PN 200059-001 Rev M 51
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard
14 Click Apply & Next. The License & Registration page appears. Enter the license details.
15 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
52 PN 200059-001 Rev M
Using the Initial Config Wizard Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
16 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Bridge and
configure the appliance data path IP next-hop router address and max WAN bandwidth.
a Leave Auto Tunnel and Auto Subnet Sharing deselected. We’ll take care of these features later.
Although it’s not technically necessary to deselect either one, we have chosen to do so for
tutorial purposes later in this chapter.
b Do not add tunnels. We’ll manually add remote appliances and create tunnels later.
PN 200059-001 Rev M 53
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard
18 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
19 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
54 PN 200059-001 Rev M
Using the Initial Config Wizard Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
21 Click Done. The appliance saves the settings and reboots automatically.
PN 200059-001 Rev M 55
Silver Peak NX Series Appliances Network Deployment Guide Verifying Appliance Connectivity
56 PN 200059-001 Rev M
Verifying Appliance Connectivity Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
3 To ensure that local routing is working correctly, ping an address on the subnet from which PBR will
be redirecting traffic.
To do that, use the same ping screen, specify either an address of a device or the router’s address in
that subnet, and ping with the -I option, as shown.
Tip Prior to putting a bridge mode appliance in production, it is always a good practice to test
connectivity with the appliance in bypass to make sure that the network will function in the event
the Silver Peak device fails to wire.
PN 200059-001 Rev M 57
Silver Peak NX Series Appliances Network Deployment Guide Enabling Subnet Sharing
Note You could have selected Auto Subnet Sharing in the Initial Config Wizard, instead of
doing this step. We do it here to highlight how the Subnet table changes after tunnels come up.
To enable subnets on A
1 Select Configuration > Subnets. The Subnets tab appears. Notice that no subnets are displayed.
58 PN 200059-001 Rev M
Enabling Subnet Sharing Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
To enable subnets on B
1 Select Configuration > Subnets. The Subnets tab appears. Set the configuration.
PN 200059-001 Rev M 59
Silver Peak NX Series Appliances Network Deployment Guide Creating Tunnels and Updating the Subnet Table
a In the Name field, assign a locally significant name for the tunnel.
b In the Admin field, accept the default value, Up.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager supplies the IP address for the local appliance.
e In the Remote IP address field, enter the data path IP address of the remote Silver Peak
appliance.
f Leave Auto Max BW selected, so the appliance uses the lower of the two system bandwidths.
If you wanted to configure this manually, then you would deselect Auto Max BW and, in the
Max BW field, enter the maximum bandwidth for this tunnel. The value must be less than or
equal to the upstream bandwidth of your WAN connection.
g Leave the Min BW at its default, 32 [Kbps].
h Click Apply.
i Save the changes.
The tunnel status won’t change to Up until a tunnel is configured at both ends. That is, until after
we configure a tunnel from B to A.
60 PN 200059-001 Rev M
Creating Tunnels and Updating the Subnet Table Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
a In the Name field, assign a locally significant name for the tunnel.
b In the Admin field, accept the default value, Up.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager prefills the IP address for the local appliance.
e In the Remote IP address field, enter the data path IP address of the remote Silver Peak
appliance.
f Leave Auto Max BW selected, so the appliance uses the lower of the two system bandwidths.
If you wanted to configure this manually, then you would deselect Auto Max BW and, in the
Max BW field, enter the maximum bandwidth for this tunnel. The value must be less than or
equal to the upstream bandwidth of your WAN connection.
g Leave the Min BW at its default, 32 [Kbps].
h Click Apply.
i Save the changes.
Within a few seconds, the tunnel Status changes to Up - active. Click Refresh, if required.
PN 200059-001 Rev M 61
Silver Peak NX Series Appliances Network Deployment Guide Creating Tunnels and Updating the Subnet Table
After Appliance B learns the Appliance A’s subnets, it automatically sends packets destined there
into the correct tunnels.
Notice that the subnet containing Site A’s end devices — the 10.110.33.0 subnet — does not appear
in the table.
This is because the Silver Peak at Site A doesn’t have an interface with an IP address in the
10.110.33.0 subnet. As a result, the local Silver Peak at Site A can’t advertise this subnet to
Appliance B.
So, we need to specifically configure Appliance A to advertise this subnet to the other Silver Peaks.
62 PN 200059-001 Rev M
Creating Tunnels and Updating the Subnet Table Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
PN 200059-001 Rev M 63
Silver Peak NX Series Appliances Network Deployment Guide Configuring the Router to Redirect Traffic
Using a Juniper Router for Filter-Based Forwarding (FBF) See page 65.
CAUTION Do not enable this PBR on the interface to which the Silver Peak appliance
connects.
To gain access to the CLI, access the router via the console port or a Telnet session.
Note If the Silver Peak appliances are using auto-optimization but not enabling subnet sharing,
then the route-map on the Cisco router also needs to be applied to the WAN interface to intercept
incoming traffic from the WAN that’s not in a tunnel between the Silver Peaks. Also, an
additional access-list entry would be required, with the source and destination subnets reversed
to match the traffic coming in on the WAN interface. This does not apply to the example as
implemented in this chapter.
If the Silver Peak appliance is not directly connected to the router/switch that is doing the redirection,
use an IP SLA statement to ensure that traffic is redirected only when the Silver Peak appliance is Up.
configure terminal
ip sla 1
icmp-echo 10.110.31.100
ip sla schedule 1 life forever start-time now
64 PN 200059-001 Rev M
Configuring the Router to Redirect Traffic Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
exit
interface gigabitEthernet 3
ip route-cache policy
ip policy route-map silverpeak
end
write mem
routing-options{
static {
route 0.0.0.0/0 next-hop 192.168.0.1;
}
}
This routing instance creates a new default route directing traffic to the Silver Peak appliance. Note
the route with the higher metric. If the first route is unreachable, traffic will be directed via the
second route.
PN 200059-001 Rev M 65
Silver Peak NX Series Appliances Network Deployment Guide Configuring the Router to Redirect Traffic
3 Create firewall filters that dictate which traffic uses the created routing instance:
This simply creates a filter that says traffic from Site A should use the created routing instance. That
is, traffic from 172.60.10.0/24 should use 172.70.10.101 as its default route.
4 Apply the filter to an interface. Note that similar to PBR, the filter should not be applied to the
interface directly connected to the Silver Peak appliance.
Note This configuration is valid for a Silver Peak appliance that is directly connected to the
Juniper device.
66 PN 200059-001 Rev M
Verifying Traffic Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
Verifying Traffic
Here, we want to verify that the tunnels are carrying traffic and that flows are being optimized.
Status column indicates whether a Click the icon for more information on which Silver
flow is being optimized or not. Peak technologies are being applied to the flow.
PN 200059-001 Rev M 67
Silver Peak NX Series Appliances Network Deployment Guide Verifying Traffic
• when appliance is up
If the bytes and packets are not incrementing, then the route policy and access list are not matching
the traffic that you want to redirect. If that’s the case, check the IP addresses you entered, as well as
the route policy.
68 PN 200059-001 Rev M
Verifying Traffic Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
PN 200059-001 Rev M 69
Silver Peak NX Series Appliances Network Deployment Guide Verifying Traffic
70 PN 200059-001 Rev M
Overview Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
Overview
This scenario deploys Site B in-line and the Site A network out-of-path using an available spare router
port. Policy-Based Routing (PBR) is configured on interfaces of Site A’s router to redirect traffic
destined for the WAN to the Silver Peak appliance.
Network Diagram
Figure 3-2 Out-of-Path Deployment with Policy-Based Routing (PBR): Router Mode [requires spare router port]
In this example, the Silver Peak appliance optimizes traffic to/from 172.60.10.0/24 and 172.80.10.0/24.
Summary
IP Addresses This deployment model requires two IP addresses (on the same or separate subnets):
• Silver Peak Appliance data path IP address (to originate and terminate tunnel)
• Silver Peak Management IP Address (for appliance configuration and
management)
PN 200059-001 Rev M 71
Silver Peak NX Series Appliances Network Deployment Guide Overview
Fail-Safe Behavior
Fail-safe behavior should always be tested before production deployment by ensuring that traffic
continues to flow in each of the following cases:
1 With the appliance in bypass state
2 With the appliance powered off
3 With the tunnels administratively down.
72 PN 200059-001 Rev M
Overview Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
1 Gather all the IP addresses Saves time and avoids mistakes. “Collecting the Necessary
needed for setup Information” on page 43.
2 Install the appliance into the Physical appliance: Connect the Site A appliance to Silver Peak Appliance Manager
network the Site A router, and insert the Site B appliance Operator’s Guide
between its WAN edge router and the Ethernet switch.
Quick Start Guides
Verify connectivity, connect power, and verify LEDs.
Virtual appliance: Configure the hypervisor, with the
required interfaces.
3 Configure Site A’s From a web browser, access and use the Initial “Using the Initial Config Wizard” on
appliancea Configuration Wizard to configure the appliance in page 44.
Router mode.
Reboot the appliance after finishing the configuration.
5 Configure the router Access the router’s command line interface, and “Configuring the Router to Redirect
configure the router for policy-based routing. Traffic” on page 64.
6 Site A Appliance: Create Use the Appliance Manager to configure Site A’s Silver “Verifying Traffic” on page 67.
tunnel and Route Policy Peak appliance.
entry
7 Configure Site B’s appliance Use the Initial Configuration Wizard to configure “Configuring Site B’s Appliance”
for in-line deploymenta Site B’s appliance in Bridge mode. on page 85.
Reboot the appliance.
a. IMPORTANT: The Appliance Next-hop IP Address must be the IP address of the WAN edge router. This may or may not be the same
as the LAN Next-hop IP Address for hosts on the LAN side of your network. If in doubt, check with your network administrator.
PN 200059-001 Rev M 73
Silver Peak NX Series Appliances Network Deployment Guide Overview
Table 3-2 Out-of-Path Deployment with Policy-Based Routing (PBR): Router Mode [Spare Router Port Available]
Hostname A B
Mode Out-of-Path (Router) In-line (Bridge)
Admin Password: Old admin admin
Admin Password: New / Confirm
Time Zone
NTP Server IP Address
License (for virtual appliance only)
mgmt0 IP Address / Maska 172.60.10.100 / 24 172.80.10.100 / 24
mgmt0 Next-hop IP Address 172.60.10.1 172.80.10.1
Appliance data path IP Address / Mask 172.70.10.101 / 24 172.80.10.101 / 24
Appliance data path Next-hop IP 172.70.10.1 172.80.10.1
b
LAN Next-hop IP Address (optional) not applicable ---
a. In this example, all mgmt0 IP addresses are in the same subnet. In your actual network, it’s likely that mgmt0 IP
addresses are in different subnets.
b. LAN next-hop IP is only required when there are subnets for which the Silver Peak appliance does not have a
configured IP address.
74 PN 200059-001 Rev M
Using the Initial Config Wizard with Site A’s Appliance Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
2 For the username and for the password, enter admin. The initial configuration page appears.
Read it, and click Next.
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
PN 200059-001 Rev M 75
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard with Site A’s Appliance
3 Select the MAC address for wan0. Make sure that the addresses match the MAC addresses
associated with the virtual interfaces of the Silver Peak virtual machine (VM).
4 Click Apply & Next. The License & Registration page appears. Enter the license details.
76 PN 200059-001 Rev M
Using the Initial Config Wizard with Site A’s Appliance Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
5 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
6 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Router and
then click +Add to add a WAN interface.
PN 200059-001 Rev M 77
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard with Site A’s Appliance
Configure the appliance data path IP next-hop router address and max WAN bandwidth.
78 PN 200059-001 Rev M
Using the Initial Config Wizard with Site A’s Appliance Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
8 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
9 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
PN 200059-001 Rev M 79
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard with Site A’s Appliance
11 Click Done. The appliance saves the settings and reboots automatically.
80 PN 200059-001 Rev M
Configuring the Router for Policy-Based Routing (PBR) Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
ip sla 1
icmp-echo 172.70.10.101
frequency 5
ip sla schedule 1 life forever start-time now
!
track 123 ip sla 1 reachability
Note Do not apply the policy route-map to the interface connected to the Silver Peak (in
this example, GigabitEthernet0/0), or you will create a routing loop.
interface GigabitEthernet0/0
description Connected to Silver Peak WAN0
ip address 172.70.10.1 255.255.255.0
interface GigabitEthernet0/1
description Connected to LAN
ip address 172.60.10.1 255.255.255.0
ip policy route-map silverpeak-lan-to-wan
interface GigabitEthernet0/2
description Connected to WAN
ip address 1.1.1.1 255.255.255.252
PN 200059-001 Rev M 81
Silver Peak NX Series Appliances Network Deployment Guide Configuring the Router for Policy-Based Routing (PBR)
2 Configure the WAN interface with the policy route-map named silverpeak-wan-to-lan.
interface GigabitEthernet0/2
description Connected to WAN
ip address 1.1.1.1 255.255.255.252
ip policy route-map silverpeak-wan-to-lan
82 PN 200059-001 Rev M
Configuring a Tunnel to the Remote Site Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
a In the Name field, assign a locally significant name. Silver Peak recommends using the naming
convention of SiteA-to-SiteB.
b In the Admin field, accept the default value, up, from the drop-down menu.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager prefills the IP address for the local appliance.
e In the Remote IP field, enter the Appliance data path IP address that belongs to the remote
appliance.
f In the Max BW field, enter the maximum bandwidth for this tunnel. This must be less than or
equal to the upstream bandwidth of your WAN connection. Or, select Auto Max BW so the
appliances use the lower of the two system bandwidths.
g Leave the Min BW at its default, 32 [Kbps].
PN 200059-001 Rev M 83
Silver Peak NX Series Appliances Network Deployment Guide Configuring a Tunnel to the Remote Site
4 Click Apply. The data entry area disappears, and the table displays the new tunnel.
84 PN 200059-001 Rev M
Configuring Site B’s Appliance Chapter 3 Out-of-Path with Policy-Based-Routing Redirection
For more information, see “Verifying Connectivity After Configuring Deployment” on page 19.
PN 200059-001 Rev M 85
Silver Peak NX Series Appliances Network Deployment Guide Configuring Site B’s Appliance
86 PN 200059-001 Rev M
CHAPTER 4
This chapter provides a step-by-step example for setting up Web Cache Communications Protocol
(WCCP) service.
The example uses a Cisco router paired with a single Silver Peak appliance. The Silver Peak appliances
participating in the WCCP service group must be deployed out-of-path (Router mode).
The example also compares two of the auto-optimization methods—subnet sharing (which, when
enabled, is the method that takes precedence), and TCP-based and IP-based auto-optimization. Both
methods require outbound (LAN–side) redirection; TCP/IP-based auto-optimization also requires
inbound (WAN–side) redirection. For more explanation, see “Determining the Need for Traffic
Redirection” on page 13.
In This Chapter
Overview See page 88.
Configuring the Site A Router for WCCP See page 92.
Using the Initial Config Wizard with Site A’s Appliance See page 94.
Using the Initial Config Wizard with Site B’s Appliance See page 110.
Creating Tunnels and Updating the Subnet Table See page 120.
PN 200059-001 Rev M 87
Silver Peak NX Series Appliances Network Deployment Guide Overview
Overview
In this scenario, the Silver Peak appliances are not connected in the direct path of the network traffic. As
a result, a network traffic redirection technique is used to forward traffic to the appliance.
Web Cache Communications Protocol (WCCP) supports the redirection of any TCP or UDP connections
to appliances participating in WCCP Service Groups. The appliance intercepts only those packets that
have been redirected to it. The appliance accelerates traffic flows that the Route Policy directs to a tunnel;
all other traffic passes through the appliance unmodified.
In the unlikely event that the appliance fails, WCCP on the WAN router removes the appliance from the
WCCP Service Group and resumes forwarding traffic normally, according to its routing tables.
At Site A, both the router and the participating appliance require a separate WCCP service group for each
protocol used in the tunnel. So, if a tunnel uses both TCP and UDP, you must create a separate WCCP
Service Group for each protocol (TCP and UDP) used in the A-to-B tunnel.
Network Diagram
Figure 4-1 Out-of-Path Deployment: Silver Peak Appliance peered with an L3 router using WCCP
The Silver Peak appliances optimize traffic to/from 10.110.31.0/24 and 10.110.11.0/24.
Note You don’t need a spare router port for this configuration. The Silver Peak appliance can
be connected to an existing or newly configured subinterface on the router via a VLAN trunk
such that a spare port on the LAN switch can be used for the physical connection.
88 PN 200059-001 Rev M
Overview Chapter 4 Out-of-Path with WCCP
Summary
IP Addresses This deployment model requires two IP addresses (on the same or separate subnets)
• Silver Peak Appliance data path IP address (to originate and terminate tunnels)
• Silver Peak Management IP Address (for appliance configuration and
management)
Configure WCCP on the Silver Peak appliance and the WAN router. Service Group IDs
on the router and appliance must match.
• Configure two WCCP v2 Service Groups on the Silver Peak appliance
(one for TCP and one for UDP)
• Configure two WCCP v2 Service Groups on the WAN router
(one for TCP and one for UDP)
Fail-Safe Behavior
Fail-safe behavior should always be tested before production deployment by ensuring that traffic
continues to flow in each of the following cases:
1 With the appliance in bypass state
2 With the appliance powered off
3 With the tunnels administratively down.
PN 200059-001 Rev M 89
Silver Peak NX Series Appliances Network Deployment Guide Overview
1 Gather all the IP addresses Saves time and avoids mistakes. “Collecting the Necessary
needed for setup Information” on page 91
2 Install the appliance into the Physical appliance: Connect the Site A appliance Silver Peak Appliance Manager
network to the Site A router, and insert the Site B appliance Operator’s Guide
between its WAN edge router and the Ethernet
Quick Start Guides
switch. Verify connectivity, connect power, and
verify LEDs.
Virtual appliance: Configure the hypervisor, with
the required interfaces.
2 Configure the Site A router for Access the Site A router’s command line interface “Configuring the Site A Router for
WCCP (CLI) to: WCCP” on page 92
• Configure an Access Control List (ACL) that
redirects all traffic from the Site A subnet to the
Site B subnet
• Configure two WCCP Service Groups — one
for UDP, one for TCP
• Associate the ACL with the Service Group
• Enable WCCP on the appropriate router
interface
3 Configure Site A’s appliance for Access the Initial Config Wizard to assign “Using the Initial Config Wizard
out-of-path deploymenta Appliance IP and Management IP addresses for with Site A’s Appliance” on
Site A’s appliance. page 94
Reboot the appliance.
4 Configure the WCCP Service • Create a pair of Service Groups (TCP, UDP) for “Configuring WCCP on Appliance
Groups on Site A’s appliance outbound redirection. A” on page 100
• Create a pair of Service Groups (TCP, UDP) for
inbound redirection.
5 Configure Site B’s appliance for Run the Initial Config Wizard to set up Site B’s “Using the Initial Config Wizard
in-line deploymenta Silver Peak appliance in Bridge mode. with Site B’s Appliance” on
Reboot the appliance. page 110
6 Verify appliance connectivity Ensure that the cable connections are sound and “Verifying Appliance Connectivity”
you haven’t misconfigured any IP addresses. on page 116
Do NOT proceed until you have verified
connectivity.
7 Enable subnet sharing This prepares each appliance to share local “Enabling Subnet Sharing” on
subnets. page 118
8 Create a tunnel and Route Policy Use the Appliance Manager. “Creating Tunnels and Updating
on Site A’s appliance the Subnet Table” on page 120
9 Test the connectivity from both Verify that the tunnel is up and that flows are being “Verifying Traffic” on page 124
ends optimized.
a. IMPORTANT: The WAN Next Hop IP Address must be the IP address of the WAN edge router. This may or may not be the same as
the Management Interface Next Hop IP Address for hosts on the LAN side of your network. If in doubt, check with your network
administrator.
90 PN 200059-001 Rev M
Overview Chapter 4 Out-of-Path with WCCP
Table 4-1 Out-of-Path Deployment: Silver Peak Appliance peered with an L3 router using WCCP
Hostname A B
Mode Router / Out-of-Path Bridge / In-Line
Admin Password: Old admin admin
Admin Password: New / Confirm
Time Zone
NTP Server IP Address
License (for virtual appliance only)
mgmt1 IP Address / Mask 10.10.10.1/24 ---
mgmt0 IP Address / Maska 192.168.1.7/24 192.168.1.9/24
mgmt0 Next-hop IP Address 192.168.1.1 192.168.1.1
Appliance data path IP Address / Mask 10.110.31.100/24 10.110.11.100/24
Appliance data path Next-hop IP 10.110.31.1/24 10.110.11.1/24
b
LAN Next-hop IP Address (optional) not applicable ---
WCCP Service Groups for outbound redirection 53 (TCP) ---
54 (UDP)
WCCP Service Groups for inbound redirection 55 (TCP) ---
56 (UDP)
WCCP Weight (default) 100 not applicable
a. In this example, all mgmt0 IP addresses are in the same subnet. In your actual network, it’s likely that mgmt0 IP
addresses are in different subnets.
b. LAN next-hop IP is only required when there are subnets for which the Silver Peak appliance does not have a
configured IP address.
PN 200059-001 Rev M 91
Silver Peak NX Series Appliances Network Deployment Guide Configuring the Site A Router for WCCP
1 Create an Access Control List (ACL) to redirect all traffic from Site A’s 10.110.33.0/24 subnet to
Site B’s 10.110.11.0/24 subnet.
CSR-1>enable
CSR-1>#
CSR-1(config)# configure terminal
CSR-1(config)# access-list 101 permit ip 10.110.33.0 0.0.0.255 10.110.11.0
0.0.0.255
Note If there were additional local subnets from which traffic originated, we would need
to create additional rules to make sure the ACL matched that traffic also.
2 Since you’ll be using two protocols, you’ll need two service groups. Therefore, create two WCCP
service groups (as placeholders) and associate the ACL with them. Here, we’ll create 53 to use (later)
with TCP and 54 to use (later) with UDP. Service Groups can be numbers between 51 and 255,
inclusive.
CSR-1(config)# ip wccp 53 redirect-list 101
CSR-1(config)# ip wccp 54 redirect-list 101
Note that we can reuse the same ACL because it matches traffic based on IP addresses. It’s the
WCCP service group which redirects traffic based on protocol.
Note On a Cisco Catalyst 6500, WCCP redirection can be done in hardware by adding the
keyword, accelerated, at the end of the global command, ip wccp 53 redirect-list 101. The
accelerated keyword allows the 6500 to do WCCP redirection (forwarding) in L2.
You must also associate the WCCP service group with Site A’s LAN-side interface. The interface
number below would be for your LAN–side interface.
CSR-1(config)# interface gigabitEthernet <number>
CSR-1(config-if)# ip wccp 53 redirect in
CSR-1(config-if)# ip wccp 54 redirect in
CSR-1(config-if)# end
92 PN 200059-001 Rev M
Configuring the Site A Router for WCCP Chapter 4 Out-of-Path with WCCP
Inbound Redirection
How you plan to optimize traffic affects whether or not you also need inbound redirection from the
WAN router (also known as WAN-side redirection):
If you enable subnet sharing (which relies on advertising local subnets between Silver Peak
appliances) or route policies (which specify destination IP addresses), then you only need outbound
redirection.
Silver Peak recommends using auto subnet sharing as a best practice.
If, instead, you default to TCP-based or IP-based auto-optimization (which relies on initial
handshaking outside a tunnel), then you must set up inbound and outbound redirection on the WAN
router.
This simply means creating another access list with the source and destinations addresses reversed
from the one shown in the last section (since incoming packets on the WAN side are destined to the
local LAN), and creating two new WCCP service groups to the WAN interface that’s using the new
ACL.
Note The best practice recommendation is to use auto subnet sharing (covered elsewhere in this
chapter), which does not require WAN side redirects. If you’re going to use auto subnet
sharing, then you can skip this section.
1 Add an entry to the Access Control List (ACL) to redirect traffic from Site B’s 10.110.11.0/24
subnet to Site A’s 10.110.33.0/24 subnet.
This entry will redirect traffic inbound from the other side of the network to the local Silver Peak.
This is necessary in cases where subnet sharing is not being used.
Note that the source and destination subnets are reversed from the previous example.
CSR-1>enable
CSR-1#
CSR-1(config)# configure terminal
CSR-1(config)# access-list 102 permit ip 10.110.11.0 0.0.0.255 10.110.33.0
0.0.0.255
This last entry (access-list 102) redirects inbound WAN-to-LAN traffic from the other side
of the network to the local Silver Peak. This is necessary in cases where subnet sharing is not
being used.
2 Create two WCCP Service Groups (as placeholders) and associate the new ACL (102) with them.
Here, we’ll create 55 to use (later) with TCP and 56 to use (later) with UDP.
Do not use the same Service Group numbers that are used in the previous (outbound redirection)
section. The best practice is to use two, unique Service Groups for inbound redirection.
CSR-1(config)# ip wccp 55 redirect-list 102
CSR-1(config)# ip wccp 56 redirect-list 102
3 You must also associate the WCCP service group with Site A’s WAN-side interface. The interface
number would be the one for your WAN facing interface.
CSR-1(config)# interface gigabitEthernet <number>
CSR-1(config)# ip wccp 55 redirect in
CSR-1(config)# ip wccp 56 redirect in
CSR-1(config)# end
PN 200059-001 Rev M 93
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard with Site A’s Appliance
Appliance A
2 For the username and for the password, enter admin. The initial configuration page appears.
Read it, and click Next.
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
94 PN 200059-001 Rev M
Using the Initial Config Wizard with Site A’s Appliance Chapter 4 Out-of-Path with WCCP
3 Select the MAC address for wan0. Make sure that the addresses match the MAC addresses
associated with the virtual interfaces of the Silver Peak virtual machine (VM).
For example, in the VMware client, you would check on the Virtual Machine Properties page.
PN 200059-001 Rev M 95
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard with Site A’s Appliance
4 Click Apply & Next. The License & Registration page appears. Enter the license details.
5 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
96 PN 200059-001 Rev M
Using the Initial Config Wizard with Site A’s Appliance Chapter 4 Out-of-Path with WCCP
6 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Router and
then click +Add to add a WAN interface.
Configure the appliance data path IP next-hop router address and max WAN bandwidth.
PN 200059-001 Rev M 97
Silver Peak NX Series Appliances Network Deployment Guide Using the Initial Config Wizard with Site A’s Appliance
8 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
98 PN 200059-001 Rev M
Using the Initial Config Wizard with Site A’s Appliance Chapter 4 Out-of-Path with WCCP
9 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
11 Click Done. The appliance saves the settings and reboots automatically.
PN 200059-001 Rev M 99
Silver Peak NX Series Appliances Network Deployment Guide Configuring WCCP on Appliance A
Note ACTIVE - Designated will be the state for one Silver Peak appliance — this is the device
that owns the communication for WCCP with the routers.
1 Click Add Service Group. The new WCCP Service group appears with the default settings.
2 In the Service Group ID field, enter the WCCP Service Group number you entered on the router as
a placeholder for the TCP protocol. On the router, we entered 53.
3 In the Admin field, accept the default of up.
4 In the Protocol field, leave tcp selected.
5 To access the Service Groups Advanced Settings page, click Advance Settings.
a In the Forwarding Method field, select either. Either allows the appliance and the router to
negotiate the best method for assignment. That is, GRE (Generic Routing Encapsulation) or L2.
b In the Weight field, keep the max default value of 100.
c In the Assignment Method field, leave the default of either. Either allows the appliance and the
router to negotiate the best method for assignment. That is, hash or mask.
d Leave Force L2 Return deselected.
e In the Password field, optionally enter a password.
f In the Assignment Detail field, select lan-ingress.
• If you’re not configuring the tunnel traffic for auto-optimization, then accept the default of
lan-ingress. This is the assumption made for this example, since all redirection will be from
the LAN to the WAN.
• wan-ingress assignment detail is only required when redirection is needed from the WAN
to the LAN, when using TCP/IP auto-optimization.
• custom is used to provide granular control of flow distribution. Contact Silver Peak
Technical Support for assistance.
g To save the settings and close the dialog box, click OK.
6 From the Interface field, select wan0.
7 For Compatibility Mode, select the option appropriate for your router. If a WCCP group is peering
with a router running Nexus OS, then the appliance must adjust its WCCP protocol packets to be
compatible. By default, the appliance is IOS-compatible.
8 In the Router IP Address field, enter the IP address of the WCCP router, 10.110.31.1.
9 Click Apply. A new WCCP Service Group for TCP appears.
a In the Forwarding Method field, select either. Either allows the appliance and the router to
negotiate the best method for assignment. That is, GRE (Generic Routing Encapsulation) or L2.
b In the Weight field, keep the max default value of 100.
c In the Assignment Method field, leave the default of either. Either allows the appliance and the
router to negotiate the best method for assignment. That is, hash or mask.
d Leave Force L2 Return deselected.
e In the Password field, optionally enter a password.
f In the Assignment Detail field, select lan-ingress.
• If you’re not configuring the tunnel traffic for auto-optimization, then accept the default of
lan-ingress. This is the assumption made for this example, since all redirection will be from
the LAN to the WAN.
• wan-ingress assignment detail is only required when redirection is needed from the WAN
to the LAN, when using TCP/IP auto-optimization.
• custom is used to provide granular control of flow distribution. Contact Silver Peak
Technical Support for assistance.
g To save the settings and close the dialog box, click OK.
6 From the Interface field, select wan0.
a For Compatibility Mode, select the option appropriate for your router. If a WCCP group is
peering with a router running Nexus OS, then the appliance must adjust its WCCP protocol
packets to be compatible. By default, the appliance is IOS-compatible.
b In the Router IP Address field, enter the IP address of the WCCP router, 10.110.31.1.
7 Click Apply.
• The data entry area disappears, and the table displays the new WCCP Service Group for UDP.
• State changes from INIT to ACTIVE, DESIGNATED.
• This means that the WCCP protocol is working properly with the router, and that this appliance
is Primary and Active.
State Definition
Note ACTIVE - Designated will be the state for one Silver Peak appliance — this is the device
that owns the communication for WCCP with the routers.
1 Click Add Service Group. The new WCCP Service group appears with the default settings.
2 In the Service Group ID field, enter the WCCP Service Group number you entered on the router as
a placeholder for the TCP protocol. On the router, we entered 55.
3 In the Admin field, accept the default of up.
4 In the Protocol field, leave tcp selected.
5 To access the Service Groups Advanced Settings page, click Advance Settings.
a In the Forwarding Method field, select either. Either allows the appliance and the router to
negotiate the best method for assignment. That is, GRE (Generic Routing Encapsulation) or L2.
b In the Weight field, keep the max default value of 100.
c In the Assignment Method field, leave the default of either. Either allows the appliance and the
router to negotiate the best method for assignment. That is, hash or mask.
d Leave Force L2 Return deselected.
e In the Password field, optionally enter a password.
f In the Assignment Detail field, select wan-ingress.
wan-ingress assignment detail is required when redirection is needed from the WAN to the
LAN, when using TCP/IP auto-optimization.
g To save the settings and close the dialog box, click OK.
6 From the Interface field, select wan0.
7 For Compatibility Mode, select the option appropriate for your router. If a WCCP group is peering
with a router running Nexus OS, then the appliance must adjust its WCCP protocol packets to be
compatible. By default, the appliance is IOS-compatible.
8 In the Router IP Address field, enter the IP address of the WCCP router, 10.110.31.1.
9 Click Apply. A new WCCP Service Group for TCP appears.
a In the Forwarding Method field, select either. Either allows the appliance and the router to
negotiate the best method for assignment. That is, GRE (Generic Routing Encapsulation) or L2.
b In the Weight field, keep the max default value of 100.
c In the Assignment Method field, leave the default of either. Either allows the appliance and the
router to negotiate the best method for assignment. That is, hash or mask.
d Leave Force L2 Return deselected.
e In the Password field, optionally enter a password.
f In the Assignment Detail field, select wan-ingress.
wan-ingress assignment detail is only required when redirection is needed from the WAN to the
LAN, when using TCP/IP auto-optimization.
g To save the settings and close the dialog box, click OK.
6 From the Interface field, select wan0.
a For Compatibility Mode, select the option appropriate for your router. If a WCCP group is
peering with a router running Nexus OS, then the appliance must adjust its WCCP protocol
packets to be compatible. By default, the appliance is IOS-compatible.
b In the Router IP Address field, enter the IP address of the WCCP router, 10.110.31.1.
7 Click Apply.
• The data entry area disappears, and the table displays the new WCCP Service Group for UDP.
• State changes from INIT to ACTIVE, DESIGNATED.
• This means that the WCCP protocol is working properly with the router, and that this appliance
is Primary and Active.
State Definition
Appliance B
2 For the username and for the password, enter admin. The initial configuration page appears.
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
3 Read it, and click Next. Select the MAC addresses for lan0 and wan0. Make sure that the addresses
match the MAC addresses associated with the virtual interfaces of the Silver Peak virtual machine.
For example, in the VMware client, you would check on the Virtual Machine Properties page.
4 Click Apply & Next. The License & Registration page appears. Enter the license details.
5 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
6 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Bridge and
configure the appliance data path IP next-hop router address and max WAN bandwidth.
a Leave Auto Tunnel and Auto Subnet Sharing deselected. We’ll take care of these features later.
Although it’s not technically necessary to deselect either one, we have chosen to do so for
tutorial purposes later in this chapter.
b Do not add tunnels. We’ll manually add remote appliances and create tunnels later.
8 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
9 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
Click Done. The appliance saves the settings and reboots automatically.
3 To ensure that local routing is working correctly, ping an address on the subnet from which WCCP
will be redirecting traffic.
To do that, use the same ping screen, specify either an address of a device or the router’s address in
that subnet, and ping with the -I option, as shown.
Tip Prior to putting a bridge mode appliance in production, it is always a good practice to test
connectivity with the appliance in bypass to make sure that the network will function in the event
the Silver Peak device fails to wire.
Subnet information is not shared between appliance until a tunnel comes up between them.
In the next few steps, we’ll enable subnet sharing on the appliances, but no subnet informations will
actually be shared until the tunnels are brought up in the next section.
Note You could have selected Auto Subnet Sharing in the Initial Config Wizard, instead of
doing this step. We do it here to highlight how the Subnet table changes after tunnels come up.
To enable subnets on B
We’ll repeat the same steps we performed for A.
1 On Appliance B, select Configuration > Subnets. The Subnets tab appears. Set the configuration.
a In the Name field, assign a locally significant name for the tunnel.
b In the Admin field, accept the default value, Up.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager prefills the IP address for the local appliance.
e In the Remote IP address field, enter the data path IP address of Appliance B.
f Leave Auto Max BW selected, so the appliance uses the lower of the two system bandwidths.
If you wanted to configure this manually, then you would deselect Auto Max BW and, in the
Max BW field, enter the maximum bandwidth for this tunnel. The value must be less than or
equal to the upstream bandwidth of your WAN connection.
g Leave the Min BW at its default, 32 [Kbps].
h Click Apply.
i Save the changes.
The tunnel status won’t change to Up until a tunnel is configured at both ends. That is, until after
we configure a tunnel from B to A.
a In the Name field, assign a locally significant name for the tunnel.
b In the Admin field, accept the default value, Up.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager prefills the IP address for the local appliance.
e In the Remote IP address field, enter the data path IP address of Appliance A.
f Leave Auto Max BW selected, so the appliance uses the lower of the two system bandwidths.
If you wanted to configure this manually, then you would deselect Auto Max BW and, in the
Max BW field, enter the maximum bandwidth for this tunnel. The value must be less than or
equal to the upstream bandwidth of your WAN connection.
g Leave the Min BW at its default, 32 [Kbps].
h Click Apply.
i Save the changes.
Within a few seconds, the Status of both tunnels should change to Up - active.
Click Refresh, if required.
Now that the tunnels are up, the appliances can begin advertising subnet information to each other.
After Appliance B learns the Appliance A’s subnets, it automatically send packets destined there
into the correct tunnel.
Notice that the subnet containing Site A’s end devices — the 10.110.33.0 subnet — does not appear
in the table.
This is because the Silver Peak at Site A doesn’t have an interface with an IP address in the
10.110.33.0 subnet. As a result, the local Silver Peak at Site A can’t advertise this subnet to
Appliance B.
So, we need to specifically configure Appliance A to advertise this subnet to Appliance B.
• Notice that Appliance B learned 10.110.31.0/24 and 10.110.33.0/24 from its peer, Appliance A.
• If Appliance A goes down, the subnets it advertises disappear from the table.
• The router knows that Appliance A is down and sends the traffic—unoptimized—to subnet
10.110.33.0/24.
Verifying Traffic
Here, we want to verify that the tunnels are carrying traffic and that flows are being optimized.
Status column indicates whether a Click the icon for more information on which Silver
flow is being optimized or not. Peak technologies are being applied to the flow.
Best Practices
Tips for Deployment
Inbound WCCP redirection is preferred over outbound [also known as ingress/egress] redirection
because inbound redirection is less CPU-intensive on the router. Inbound redirection is done in
hardware where as outbound is done in software.
• For Catalyst 6000/76xx deployments, use only inbound redirection to avoid using “redirection
exclude in”, which is not understood by the switch hardware and must be processed in software.
• For Catalyst 6000/76xx deployments, use L2 redirection for near line-rate redirection. Silver
Peak appliances automatically negotiate assignment and forwarding methods with all routers
and L3 switches from Cisco to the best possible combination that the router or L3 switch
supports.
WCCPv2 interception forwards all packets from the router or L3 switch to the appliance. Special
care should be taken when traffic redirected to the appliance has to be returned back to the router or
L3 switch. For many routers the return traffic is delivered via L2 so there is no CPU impact.
However, Catalyst 6000/76xx switches returns via GRE so the CPU can be negatively impacted
unless Force L2 return is enabled on the appliance.
• Force L2 Return should only be enabled when the interface/VLAN that the appliance is
connected to is not also an interface with the redirection applied to.
The appliance should always be connected to an interface/VLAN that does not have redirection
enabled – preferably a separate interface/VLAN would be provided for the appliance.
The appliance and Catalyst switch negotiate which redirect and return method to use when the
service group is formed. There can be many access VLANs on the aggregation switches. Redirection
is configured on all VLANs that need optimization. Layer 2 switching ports, including trunk ports,
are not eligible for redirection.
If Auto Optimization is used for matching traffic to be optimized via the appliance, WCCP
redirection must also be applied on the uplinks of the router or L3 switch to the core/WAN.
If WCCP redirection is needed on both the WAN and the LAN, the preferred configuration on the
appliance is to set the WCCP group configured on the WAN to wan-ingress and the group
configured on the LAN to lan-ingress.
• The configuration of wan-ingress and lan-ingress ensures that load balancing is symmetrical in
both directions of a flow.
• wan-ingress uses the destination address for distribution in the router/L3 switch table
• lan-ingress uses the source address for distribution.
If Route Policies are used for matching traffic to be optimized via the appliance, WCCP redirection
is not required on the core uplinks, only the access/LAN links. If Active/Active redistribution is
enabled with route policies, then flow redirection is required to handle asymmetrical flows caused
by load balancing. Flow redirection can handle millions of flows and ensures that the owner of a
given flow always receives the TCP flow for processing.
GRE
GRE is a protocol that carries other protocols as its payload:
In this case, the payload is a packet from the router to the appliance. GRE works on routing and switching
platforms. It allows the WCCP clients to be separate from the router via multiple hops. Because GRE is
processed in software, router CPU utilization increases with GRE redirection. Hardware-assisted GRE
redirection is available on the Catalyst 6500 with Sup720.
L2 Redirection
L2 redirection requires the appliance to be in the same subnet as the router or switch (L2 adjacency).
The switch rewrites the destination L2 MAC header with the appliance MAC address. The packet is
forwarded without additional lookup.
L2 redirection is done in hardware and is available on the Catalyst 6500/7600 platforms. CPU
utilization is not impacted because L2 redirection is hardware-assisted; only the first packet is
switched by the Multilayer Switch Feature Card (MSFC) with hashing.
After the MSFC populates the NetFlow table, subsequent packets are switched in hardware. L2
redirection is preferred over GRE because of lower CPU utilization.
There are two methods to load balance appliances with L2 redirection: hashing and masking.
This chapter provides a step-by-step example of a deployment where the Silver Peak appliance uses the
Virtual Router Redundancy Protocol (VRRP) to peer with the existing router, when no spare router port
is available.
In This Chapter
Overview See page 130.
Overview
In this deployment mode, the Silver Peak appliance uses the Virtual Router Redundancy Protocol
(VRRP) to peer with the existing router, when no spare router port is available.
• This requires changing the IP address of the router and adding the VRRP VIP (Virtual IP)
address to the router.
• The VIP address takes the existing router address; this way, you don’t need to modify the
client’s default gateway.
• The Silver Peak appliance becomes the primary default gateway for all users in that network.
• In the unlikely event that the Silver Peak appliance fails, the router automatically becomes the
default gateway.
• The remote location is configured In-Line.
Network Diagram
Figure 5-1 Out-of-Path Deployment: Silver Peak Appliance peered with an L3 router using Virtual Router Redundancy
Protocol (VRRP)
In this example, the Silver Peak appliance optimizes traffic to/from 10.110.31.0/24 and 10.110.11.0/24.
Summary
Appliance Placement Appliance shares LAN segment with existing equipment
• Appliance wan0 interface connects to Ethernet LAN switch
• Do not connect lan0 interface
The VIP must be the default gateway for the clients and servers on the LAN subnet.
NOTE: Typically, this would be the current default gateway, to avoid client
reconfigurations.
• The Silver Peak appliance must share the default gateway VIP with WAN router
using VRRP.
• The Silver Peak appliance must be configured with higher priority and preemption
to ensure VRRP reverts to the appliance.
Fail-Safe Behavior
Fail-safe behavior should always be tested before production deployment by ensuring that traffic
continues to flow in each of the following cases:
1 With the appliance in bypass state
2 With the appliance powered off
3 With the tunnels administratively down.
1 Gather all the IP addresses Saves time and avoids mistakes. “Collecting the Necessary
needed for setup Information” on page 133.
2 Install the appliance into the Physical appliance: Connect the Site A appliance Silver Peak Appliance Manager
network to the Site A router, and insert the Site B appliance Operator’s Guide
between its WAN edge router and the Ethernet
Quick Start Guides
switch. Verify connectivity, connect power, and
verify LEDs.
Virtual appliance: Configure the hypervisor, with
the required interfaces.
3 Configure the appliance In a browser, access and use the Initial “Using the Initial Config Wizard” on
Configuration Wizard to configure each appliance page 134.
— one in Bridge mode, the other in Router mode.
Reboot each appliance after finishing the
configuration.
4 Verify appliance connectivity Tests data path connectivity. “Verifying Appliance Connectivity”
on page 146.
Do NOT proceed until you verify connectivity.
5 Enable subnet sharing This prepares each appliance to share local “Enabling Subnet Sharing” on
subnets. page 147.
6 Create a tunnel on each Specify the local and remote endpoints for the “Creating Tunnels” on page 149.
appliance tunnel.
Afterwards, verify that the tunnels are up and the
subnet table has updated.
7 Configure Site A’s router Access the router’s command line interface, and “Configuring VRRP on a Cisco
configure the router for policy-based routing. Router” on page 151.
8 Configure VRRP on Site A’s Use two of the Configuration pages: Deployment “Configuring VRRP on Silver Peak
appliance and VRRP A1” on page 152
9 Test the connectivity from both Verify that the tunnel is up and that flows are being “Verifying Traffic” on page 154.
ends optimized.
Table 5-1 Out-of-Path Deployment: Silver Peak Appliance peered with an L3 router using Virtual Router Redundancy
Protocol (VRRP)
Hostname A1 B
Mode Out-of-Path (Router) In-line (Bridge)
Admin Password: Old admin admin
Admin Password: New / Confirm
Time Zone
NTP Server IP Address
License (for virtual appliance only)
mgmt0 IP Address / Maska 192.168.1.7/24 192.168.1.9/24
mgmt0 Next-hop IP Address 192.168.1.1 192.168.1.1
Appliance data path IP Address / Mask 10.110.31.100/24 10.110.11.100/24
Appliance data path Next-hop IP 10.110.31.2/24 10.110.11.1/24
b
LAN Next-hop IP Address (optional) not applicable ---
VRRP Group ID 1 ---
VRRP Virtual IP Address (VIP) 10.110.31.1 not applicable
VRRP Priority 130 not applicable
a. In this example, all mgmt0 IP addresses are in the same subnet. In your actual network, it’s likely that mgmt0 IP
addresses are in different subnets.
b. LAN next-hop IP is only required when there are subnets for which the Silver Peak appliance does not have a
configured IP address.
Appliance A1
2 For the username and for the password, enter admin. The initial configuration page appears.
Read it, and click Next.
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
3 Select the MAC address for wan0. Make sure that the addresses match the MAC addresses
associated with the virtual interfaces of the Silver Peak virtual machine (VM).
For example, in the VMware client, you would check on the Virtual Machine Properties page.
4 Click Apply & Next. The License & Registration page appears. Enter the license details.
5 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
6 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Router and
then click +Add to add a WAN interface.
Configure the appliance data path IP next-hop router address and max WAN bandwidth.
8 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
9 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
11 Click Done. The appliance saves the settings and reboots automatically.
Appliance B
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
13 Read it, and click Next. Select the MAC addresses for lan0 and wan0. Make sure that the addresses
match the MAC addresses associated with the virtual interfaces of the Silver Peak virtual machine.
For example, in the VMware client, you would check on the Virtual Machine Properties page.
14 Click Apply & Next. The License & Registration page appears. Enter the license details.
15 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
16 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Bridge and
configure the appliance data path IP next-hop router address and max WAN bandwidth.
a Leave Auto Tunnel and Auto Subnet Sharing deselected. We’ll take care of these features later.
Although it’s not technically necessary to deselect either one, we have chosen to do so for
tutorial purposes later in this chapter.
b Do not add tunnels. We’ll manually add remote appliances and create tunnels later.
18 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
19 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
21 Click Done. The appliance saves the settings and reboots automatically.
Tip Prior to putting a bridge mode appliance in production, it is always a good practice to test
connectivity with the appliance in bypass to make sure that the network will function in the event
the Silver Peak device fails to wire.
Note You could have selected Auto Subnet Sharing in the Initial Config Wizard, instead of
doing this step. We do it here to highlight how the Subnet table changes after tunnels come up.
To enable subnets on A1
1 On Appliance A1, select Configuration > Subnets. The Subnets tab appears. Notice that no subnets
are displayed.
2 Click Apply. The subnet table updates to include the local subnet.
If it doesn’t, try refreshing the page.
To enable subnets on B
1 On Appliance B, select Configuration > Subnets. The Subnets tab appears. Set the configuration.
Creating Tunnels
Create a tunnel between Appliances A1 and B. This involves accessing each appliance, in turn, and
creating a tunnel to the other (remote) appliance.
After the tunnels are up, we’ll verify that the subnet table has updated.
a In the Name field, assign a locally significant name for the tunnel.
b In the Admin field, accept the default value, Up.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager prefills the IP address for the local appliance.
e In the Remote IP address field, enter the data path IP address of the remote Silver Peak
appliance.
f Leave Auto Max BW selected, so the appliance uses the lower of the two system bandwidths.
If you wanted to configure this manually, then you would deselect Auto Max BW and, in the
Max BW field, enter the maximum bandwidth for this tunnel. The value must be less than or
equal to the upstream bandwidth of your WAN connection.
g Leave the Min BW at its default, 32 [Kbps].
h Click Apply.
i Save the changes.
The tunnel status doesn’t change to Up until a tunnel is configured at both ends. So, we’ll now
configure a tunnel from B to A1.
a In the Name field, assign a locally significant name for the tunnel.
b In the Admin field, accept the default value, Up.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager prefills the IP address for the local appliance.
e In the Remote IP address field, enter the data path IP address of the remote Silver Peak
appliance.
f Leave Auto Max BW selected, so the appliance uses the lower of the two system bandwidths.
g Leave the Min BW at its default, 32 [Kbps].
h Click Apply.
i Save the changes.
Within a few seconds, the tunnel Status changes to Up - active. Click Refresh, if required.
Now that the tunnels are up, the appliances can begin advertising subnet information to each
other.
Now that Appliance B has learned the remote appliance’s subnet(s), it automatically places packets with
destinations in learned subnets into the correct tunnels.
The Add VRRP area appears. Some fields display default values.
a Assign a Group ID number. You’ll use the same number for the primary and backup devices.
Here, we’re using 1.
b Leave Interface set to wan0.
c Leave Admin set to up.
d Leave the Advertisement Timer set to 1.
e In the Virtual Address field, enter the virtual IP that you’ll be using for both the primary and
backup appliances. In our case, it’s 10.110.31.1.
f Priority and Preemption work together. If two devices come up at the same time, the device with
the highest priority becomes the Master, and lower priority devices are backups.
If Preemption is enabled and a device with a higher priority comes online in the VRRP group,
it becomes the Master even if another device is already acting as Master. The lower priority
device then reverts to being a backup.
g We’ll set our Priority to 130, and enable Preemption.
h If you choose to use VRRP’s text authentication, then the Authentication String must be
specified in all members of the group. In this deployment, that would include Site A’s appliance
and the peered router. Here, we’ll leave it blank.
Verifying Traffic
Here, we want to verify that the tunnels are carrying traffic and that flows are being optimized.
c Click Apply. All traffic is then be handled by the backup device, which becomes the Master.
3 To verify the router’s status, access it and use the show vrrp command.
• With A1 up and acting as the Master (Cisco is backup)
Make sure to change the Silver Peak’s Admin state back to up when you’re done testing.
Status column indicates whether a Click the icon for more information on which Silver
flow is being optimized or not. Peak technologies are being applied to the flow.
In This Chapter
Overview See page 158.
Using the Initial Config Wizard for Site A See page 162.
Using the Initial Config Wizard with Site B See page 182.
Creating Tunnels and Updating the Subnet Table See page 193.
Configuring the Cisco Router for Policy-Based Routing (PBR) See page 198.
Overview
In this example, Site A deploys two primary appliances out-of-path (Router mode), and Site B deploys
a single appliance in-line (Bridge mode).
The peered appliances at Site A use the Virtual Router Redundancy Protocol (VRRP) to create and share
a common IP address, called the Virtual IP (VIP) address. Configuring for high availability assigns one
appliance a higher priority than the other appliance, thereby making it the Master, and the other, the
backup.
The appliance at Site B has separate tunnels going to each of the two appliances at Site A:
• If one of the appliances at Site A is down, then Site B only sends traffic to the appliance (tunnel)
that is up.
• If both appliances at Site A are up, then Site B sends traffic to the tunnel (appliance) that has
higher VRRP priority.
Network Diagram
Figure 6-1 Out-of-Path Deployment: Redundant Silver Peak Appliances using Policy-Based-Routing (PBR)
The Silver Peak appliances optimize traffic to/from 10.110.31.0/24 and 10.110.11.0/24.
• Each appliance needs a Silver Peak Appliance IP data path address (to originate
and terminate tunnels).
• The two appliances share one Silver Peak Appliance Virtual IP Address for VRRP.
• Each appliance needs a Silver Peak Management IP Address (for appliance
configuration and management).
Fail-Safe Behavior
Fail-safe behavior should always be tested before production deployment by ensuring that traffic
continues to flow in each of the following cases:
1 With the appliance in bypass state
2 With the appliance powered off
3 With the tunnels administratively down.
Table 6-1 Out-of-Path Deployment: Silver Peak Appliance peered with an L3 router using Virtual Router Redundancy
Protocol (VRRP)
Hostname A1 A2 B
Mode Router / Out-of-Path Router / Out-of-Path Bridge / In-line
Admin Password: Old admin admin admin
Admin Password: New / Confirm
Time Zone
NTP Server IP Address
License (for virtual appliance only)
mgmt1 IP Address / Mask 10.10.10.1/30 10.10.10.2/30 ---
mgmt0 IP Address / Maska 192.168.1.7/24 192.168.1.8/24 192.168.1.9/24
mgmt0 Next-hop IP Address 192.168.1.1 192.168.1.1 192.168.1.1
Appliance data path IP Address / Mask 10.110.31.100/24 10.110.31.101/24 10.110.11.100/24
Appliance data path Next-hop IP 10.110.31.1/24 10.110.31.1/24 10.110.11.1/24
LAN Next-hop IP Address (optional) b not applicable not applicable ---
VRRP Group ID 1 1 ---
VRRP Virtual IP Address (VIP) 10.110.31.254 10.110.31.254 not applicable
VRRP Priority 130 128 not applicable
a. In this example, all mgmt0 IP addresses are in the same subnet. In your actual network, it’s likely that mgmt0 IP addresses are in different
subnets.
b. LAN next-hop IP is only required when there are subnets for which the Silver Peak appliance does not have a configured IP address.
1 Gather all the IP addresses Saves time and avoids mistakes. “Collecting the Necessary
needed for setup Information” on page 160.
2 Install the appliance into the Physical appliance: Connect both appliances to Silver Peak Appliance Manager
network the same available subnet via an Ethernet LAN Operator’s Guide
switch. Verify connectivity, connect power, and
Quick Start Guides
verify LEDs.
Virtual appliance: Configure the hypervisor, with
the required interfaces.
3 Configure the peer appliances In a browser, access and use the Initial “Using the Initial Config Wizard for
at Site A Configuration Wizard to configure each appliance. Site A” on page 162
Reboot the appliances after finishing the
configuration.
4 Configure VRRP for the Site A You’ll configure one appliance to be the Master, and “Configuring VRRP on A1 and A2”
peers the other to be the Backup. on page 175
5 Configure flow redirection for When you create a cluster, the peers keep track of “Configuring Flow Redirection” on
the Site A peers which appliance owns each flow. If the path page 178
between client and server isn’t the same in both
directions, the flow is redirected to the appliance
that first saw it and “owns” it.
6 Configure Site B’s appliance In a browser, access and use the Initial “Using the Initial Config Wizard
Configuration Wizard to configure the appliance. with Site B” on page 182
Reboot the appliance after finishing the
configuration.
7 Verify appliance connectivity Tests data path connectivity. “Verifying Appliance Connectivity”
on page 188
Do NOT proceed until you verify connectivity.
8 Enable subnet sharing This prepares each appliance to share local “Enabling Subnet Sharing” on
subnets. page 191
9 Create a tunnel on each Specify the local and remote endpoints for the “Creating Tunnels and Updating
appliance tunnel. the Subnet Table” on page 193
10 Manually add Site A’s non-local Manually add subnets that aren’t directly connected “Configuring A1 and A2 to
subnets to an appliance interface so they can be advertised. Advertise Non-Local Subnets” on
page 196
11 Configure the router Access the router’s command line interface, and “Configuring the Cisco Router for
configure the router for policy-based routing. Policy-Based Routing (PBR)” on
page 198
12 Test the connectivity from both Verify that the tunnel is up and that flows are being “Verifying Traffic” on page 199
ends optimized.
Appliance A1
2 For the username and for the password, enter admin. The initial configuration page appears.
Read it, and click Next.
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
3 Select the MAC addresses for wan0 and mgmt1. Make sure that the addresses match the MAC
addresses associated with the vNICs in the hypervisor client.
Make sure that the addresses match the MAC addresses associated with the virtual interfaces of the
Silver Peak virtual machine (VM).
For example, in the VMware client, you would check on the Virtual Machine Properties page.
4 Click Apply & Next. The License & Registration page appears. Enter the license details.
5 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
6 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Router and
then click +Add to add a WAN interface.
7 Configure the appliance data path IP next-hop router address and max WAN bandwidth
Configure the Next-hop IP to be the physical address that the next-hop router will use — not the
VRRP virtual IP address. Otherwise, you’ll create a routing loop when the Silver Peak is the VRRP
Master.
9 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
10 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
Click Done. The appliance saves the settings and reboots automatically.
Appliance A2
12 For the username and for the password, enter admin. The initial configuration page appears.
Read it, and click Next.
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
13 Select the MAC addresses for wan0 and mgmt1. Make sure that the addresses match the MAC
addresses associated with the vNICs in the hypervisor client.
Make sure that the addresses match the MAC addresses associated with the virtual interfaces of the
Silver Peak virtual machine (VM).
For example, in the VMware client, you would check on the Virtual Machine Properties page.
14 Click Apply & Next. The License & Registration page appears. Enter the license details.
15 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
16 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Router and
then click +Add to add a WAN interface.
17 Configure the appliance data path IP next-hop router address and max WAN bandwidth.
Configure the Next-hop IP to be the physical address that the next-hop router will use — not the
VRRP virtual IP address. Otherwise, you’ll create a routing loop when the Silver Peak is the VRRP
Master.
19 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
20 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
22 Click Done. The appliance saves the settings and reboots automatically.
Using VRRP with two Silver Peaks acting as Master and Backup
If either Silver Peak acting as Master fails, the Backup assumes the role of Master and begins
optimizing traffic.
Because we want our network to behave deterministically — to minimize the amount of flow
redirection that is needed — we will configure A1 with a priority of 130, and A2 with a priority of
128. With A1 having the higher priority, it becomes the Master when both appliances are up.
Check the Preemption checkbox.
• This ensures that A1 becomes Master whenever it is up because it has the higher priority.
• A2, if it was acting as Master, reverts to backup when A1 assumes the role of Master.
• If preemption is not enabled, then whichever appliance is Master remains Master, even if a
device in that VRRP group has a higher priority. We always want A1 (which has a higher
priority) to be the Master, so we enable preemption.
.
The Add VRRP area appears. Some fields display default values.
a Assign a Group ID number. You’ll use the same number for the primary and backup devices.
Here, we’re using 1.
To refresh the page, reselect Configuration > VRRP from the menu.
3 To store your configuration, make sure to click Save Changes.
Note IMPORTANT — When configuring for flow redirection, the mgmt1 interfaces need to
be in a separate subnet from the mgmt0 interfaces.
An appliance that handles both directions of traffic for a flow can then optimize the flow properly.
Specifically, this sets the stage for TCP acceleration and CIFS acceleration.
This sequence of four diagrams illustrates how the need for flow redirection arises, and is resolved.
a Select Enable.
b In the Interface field, select mgmt1.
c Click Add Peer, and configure the IP address of mgmt1 on A2. In this example, it’s 10.10.10.2.
d Click Apply.
e Save the changes.
a Select Enable.
b In the Interface field, select mgmt1.
c Click Add Peer, and configure the IP address of mgmt1 on A1. In this example, it’s 10.10.10.1.
d Click Apply.
e Save the changes.
4 To verify that flow redirection is working, look to see that the State changes to OK, indicating that
the interfaces and flow redirection are configured properly on both sides.
Appliance B
2 For the username and for the password, enter admin. The initial configuration page appears.
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
3 Read it, and click Next. Select the MAC addresses for lan0 and wan0. Make sure that the addresses
match the MAC addresses associated with the virtual interfaces of the Silver Peak virtual machine.
For example, in the VMware client, you would check on the Virtual Machine Properties page.
4 Click Apply & Next. The License & Registration page appears. Enter the license details.
5 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
6 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Bridge and
configure the appliance data path IP next-hop router address and max WAN bandwidth.
a Leave Auto Tunnel and Auto Subnet Sharing deselected. We’ll take care of these features later.
Although it’s not technically necessary to deselect either one, we have chosen to do so for
tutorial purposes later in this chapter.
b Do not add tunnels. We’ll manually add remote appliances and create tunnels later.
8 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
9 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
11 Click Done. The appliance saves the settings and reboots automatically.
Tip Prior to putting a bridge mode appliance in production, it is always a good practice to test
connectivity with the appliance in bypass to make sure that the network will function in the event
the Silver Peak device fails to wire.
3 To ensure that local routing is working correctly, ping an address on the subnet from which PBR
(Policy-Based Routing) will be redirecting traffic. Here, that subnet is 10.110.33.0/24.
To do that, use the same ping screen, specify either an address of a device or the router’s address in
that subnet, and ping with the -I option, as shown.
Note You could have selected Auto Subnet Sharing in the Initial Config Wizard, instead of
doing this step. We do it here to highlight how the Subnet table changes after tunnels come up.
To enable subnets on A2
1 On Appliance A2, select Configuration > Subnets. The Subnets tab appears. Set the configuration.
To enable subnets on B
1 On Appliance B, select Configuration > Subnets. The Subnets tab appears. Set the configuration.
a In the Name field, assign a locally significant name for the tunnel.
b In the Admin field, accept the default value, Up.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager prefills the IP address for the local appliance.
e In the Remote IP address field, enter the data path IP address of Appliance B.
f Leave Auto Max BW selected, so the appliance uses the lower of the two system bandwidths.
If you wanted to configure this manually, then you would deselect Auto Max BW and, in the
Max BW field, enter the maximum bandwidth for this tunnel. The value must be less than or
equal to the upstream bandwidth of your WAN connection.
g Leave the Min BW at its default, 32 [Kbps].
h Click Apply.
i Save the changes.
The tunnel status won’t change to Up until a tunnel is configured at both ends. That is, until after
we configure a tunnel from B to A1.
a In the Name field, assign a locally significant name for the tunnel.
b In the Admin field, accept the default value, Up.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager prefills the IP address for the local appliance.
e In the Remote IP address field, enter the data path IP address of Appliance B.
f Leave Auto Max BW selected, so the appliance uses the lower of the two system bandwidths.
g Leave the Min BW at its default, 32 [Kbps].
h Click Apply.
i Save the changes.
The tunnel status won’t change to Up until a tunnel is configured at both ends. So, we’ll now
configure a tunnel from B to A1.
a In the Name field, assign a locally significant name for the tunnel.
b Enter the Remote IP address (that is, the data path IP address of Appliance A1).
c Click Apply.
a In the Name field, assign a locally significant name for the tunnel.
b Enter the Remote IP address (that is, the data path IP address of Appliance A2).
c Click Apply.
d Save the changes.
Within a few seconds, the Status of both tunnels should change to Up - active.
Click Refresh, if required.
Now that the tunnels are up, the appliances can begin advertising subnet information to each other.
Now that Appliance B has learned the remote appliances’ subnets, it automatically places packets
with destinations in the learned subnets into the correct tunnels.
Notice that the subnet where Site A’s end devices reside — the 10.110.33.0 subnet — does not
appear in the table.
This is because the Silver Peaks at Site A don’t have an interface with an IP address in that subnet.
As a result, the local Silver Peaks at Site A can’t advertise this subnet to Appliance B. We need to
configure A1 and A2 to advertise this subnet to other Silver Peaks.
• Notice that subnets learned from peer 10.110.31.100 ( A1) have a metric of 40, while others
were learned with a metric of 50. When Appliance B has a choice of two routes to a subnet, it
will prefer to send packets to the device having the lower metric. For subnet 10.110.33.0,
Appliance B will always route packets to A1 because it has the lower metric.
• If Appliance A1 goes down, the subnets it advertises disappear from the table, and Appliance B
will use the route advertised by peer A2 (10.110.31.100).
configure terminal
exit
interface gigabitEthernet 3
ip route-cache policy
ip policy route-map sp-vrrp
end
write mem
Verifying Traffic
Here, we want to verify that the tunnels are carrying traffic and that flows are being optimized.
You would expect the majority of the traffic to be in the tunnel to the VRRP Master, assuming it has been
the Master for an extended period of time. If there has been a recent change in Masters, this might not be
the case.
Appliance A1
Appliance A2
3 Click Apply.
All traffic is then be handled by the backup (A2), which becomes the Master.
Any flows that were going through the previous Master (A1) are redirected to that appliance by the
current Master (A2). This can be seen in the Flow Redirection statistics (see below).
If the previous Master (A1) had actually gone down (instead of having VRRP administratively
disabled), then those flows would have to be reestablished. As a result, they would flow through the
current Master (A2) and redirection would not take place.
Status column indicates whether a Click the icon for more information on which Silver
flow is being optimized or not. Peak technologies are being applied to the flow.
When the connection to the peer is functioning, the State column displays OK.
This chapter provides a step-by-step example for setting up HA (high availability) Silver Peak appliances
by using Web Cache Communications Protocol (WCCP) service with a Cisco router. If one appliance
goes down, the other then handles all the traffic. In an Active/Active deployment, the peered appliances
are also load balancing.
In this example, Site A deploys two active, redundant appliances (named A1 and A2) out-of-path (Router
mode) and, remotely, Site B deploys a single appliance (named B), in-line (Bridge mode). The focus of
this chapter is on the HA appliances; in practice, the remote appliance can be in either bridge or router
mode.
In This Chapter
Overview See page 204.
Overview
Web Cache Communications Protocol (WCCP) supports the redirection of any TCP or UDP connections
to appliances participating in WCCP Service Groups. The appliance intercepts only those packets that
have been redirected to it. The appliance accelerates traffic flows that match its Route Policy; all other
traffic passes through the appliance unmodified.
The two active Silver Peak appliances participating in the WCCP service group must be deployed
out-of-path (Router mode). In this example, those appliances are at Site A. For the purposes of this
specific example, Site B at the remote end deploys the appliance in-line (Bridge mode); there is no
inherent restriction on what mode it needs to be.
WCCP at Site A
Each of the peered appliances at headquarters uses WCCP to redirect traffic from the router to the
appliances.
WCCP redirects all traffic that is in a WCCP Service Group shared by the appliance and router.
A service group consists of a set of WCCP-enabled routers and appliances that exchange WCCP
messages. The routers send traffic to the appliances in the service group. The configuration of the
service group determines how traffic is distributed to appliances in the service group.
To use WCCP, you must create a separate WCCP Service Group for each protocol (TCP and UDP)
used in the SiteA-to-SiteB tunnel.
Network Diagram
Figure 7-1 Out-of-Path Deployment: Redundant Silver Peak Appliances peered with an L3 router using WCCP
The Silver Peak appliances optimize traffic to/from 10.110.31.0/24 and 10.110.11.0.0/24.
Summary
Appliance Placement Both appliances are attached in network, reachable by WAN router
• Each appliance’s wan0 interface connects to network
• Do not connect lan0 interface of either appliance
• Each appliance needs a Silver Peak Appliance IP data path address (to originate
and terminate tunnels)
• Each appliance needs a Silver Peak Management IP Address (for appliance
configuration and management)
Configure WCCP on Site A’s Silver Peak Appliances and the WAN router. Service
Group IDs on the router and appliance must match.
• Configure two WCCP Service Groups on each Silver Peak appliance
(one for TCP and one for UDP)
• Configure two WCCP Service Groups on the WAN router
(one for TCP and one for UDP)
Fail-Safe Behavior
Fail-safe behavior should always be tested before production deployment by ensuring that traffic
continues to flow in each of the following cases:
1 With the appliance in bypass state
2 With the appliance powered off
3 With the tunnels administratively down.
1 Gather all the IP addresses Saves time and avoids mistakes. “Collecting the Necessary
needed for setup Information” on page 209.
2 Install the Appliance A into Physical appliance: Connect both appliances to the Silver Peak Appliance Manager
the network same available subnet via an Ethernet LAN switch. Operator’s Guide
Verify connectivity, connect power, and verify LEDs.
Quick Start Guides
Virtual appliance: Configure the hypervisor, with the
required interfaces.
3 Configure the Site A router Access the Site A router’s command line interface “Configuring the Site A Router for
for WCCP (CLI) to: WCCP” on page 210
• Configure an Access Control List (ACL) that
redirects all traffic from the Site A subnet to the Site
B subnet
• Configure two WCCP Service Groups — one for
UDP, one for TCP
• Associate the ACL with the Service Group
• Enable WCCP on the appropriate router interface
4 Configure Appliance A1 In a browser, access and use the Initial Configuration “Using the Initial Config Wizard
Wizard to configure the appliance. with A1” on page 212.
Reboot the appliance after finishing the configuration.
5 Configure the WCCP Service • Create a pair of Service Groups (TCP, UDP) for “Configuring WCCP on A1” on
Groups on Appliance A1 outbound redirection. page 218
• Create a pair of Service Groups (TCP, UDP) for
inbound redirection.
6 Configure Appliance A2 In a browser, access and use the Initial Configuration “Using the Initial Config Wizard
Wizard to configure the appliance. with A2” on page 228
Reboot the appliance after finishing the configuration.
7 Configure the WCCP Service • Create a pair of Service Groups (TCP, UDP) for “Configuring WCCP on A2” on
Groups on Appliance A2 outbound redirection. page 234
• Create a pair of Service Groups (TCP, UDP) for
inbound redirection.
8 Configure flow redirection When you create a cluster, the peers keep track of “Configuring Flow Redirection” on
for the Site A peers which appliance owns each flow. If the path between page 235
client and server isn’t the same in both directions, the
flow is redirected to the appliance that first saw it and
“owns” it.
9 Configure Appliance B In a browser, access and use the Initial Configuration “Using the Initial Config Wizard
Wizard to configure the appliance. with B” on page 239
Reboot the appliance after finishing the configuration.
10 Verify appliance connectivity Tests data path connectivity. “Verifying Appliance Connectivity”
on page 245
Do NOT proceed until you verify connectivity.
11 Enable subnet sharing This prepares each appliance to share local subnets. “Enabling Subnet Sharing” on
page 248.
12 Create a tunnel on each Specify the local and remote endpoints for the tunnel. “Creating Tunnels” on page 250.
appliance
13 Manually add Site A’s Manually add subnets that aren’t directly connected to “Configuring A1 and A2 to
non-local subnets an appliance interface so they can be advertised. Advertise Non-Local Subnets” on
page 253
14 Test the connectivity from Verify that the tunnel is up and that flows are being “Verifying Traffic” on page 255
both ends optimized.
Table 7-1 Out-of-Path Deployment: Redundant Silver Peak Appliances peered with an L3 router using WCCP
Hostname A1 A2 B
Mode Router / Out-of-Path Router / Out-of-Path Bridge / In-Line
Admin Password: Old admin admin admin
Admin Password: New / Confirm
Time Zone
NTP Server IP Address
License (for virtual appliance only)
mgmt1 IP Address / Mask 10.10.10.1/24 10.10.10.2/24 ---
mgmt0 IP Address / Maska 192.168.1.7/24 192.168.1.8/24 192.168.1.9/24
mgmt0 Next-hop IP Address 192.168.1.1 192.168.1.1 192.168.1.1
Appliance data path IP Address / Mask 10.110.31.100/24 10.110.31.101/24 10.110.11.100/24
Appliance data path Next-hop IP 10.110.31.1/24 10.110.31.1/24 10.110.11.1/24
b
LAN Next-hop IP Address (optional) not applicable not applicable ---
WCCP Service Groups for outbound 53 (TCP) 53 (TCP) ---
redirection
54 (UDP) 54 (UDP ---
WCCP Service Groups for inbound 55 (TCP) 55 (TCP) ---
redirection
56 (UDP) 56 (UDP ---
WCCP Weight (default) 100 200 not applicable
a. In this example, all mgmt0 IP addresses are in the same subnet. In your actual network, it’s likely that mgmt0 IP addresses are in different
subnets.
b. LAN next-hop IP is only required when there are subnets for which the Silver Peak appliance does not have a configured IP address.
3 Since you’ll be using two protocols, you’ll need two service groups. Therefore, create two WCCP
service groups (as placeholders) and associate the ACL with them. Here, we’ll create 53 to use (later)
with TCP and 54 to use (later) with UDP. Service Groups can be numbers between 51 and 255,
inclusive.
CSR-1(config)# ip wccp 53 redirect-list 101
CSR-1(config)# ip wccp 54 redirect-list 101
Note that we can reuse the same ACL because it matches traffic based on IP addresses. It’s the
WCCP service group that redirects traffic based on protocol.
Note On a Cisco Catalyst 6500, WCCP redirection can be done in hardware by adding the
keyword, accelerated, at the end of the global command, ip wccp 53 redirect-list 101. The
accelerated keyword allows the 6500 to do WCCP redirection (forwarding) in L2.
4 You must also associate the WCCP service group with Site A’s LAN-side interface. In this chapter’s
example, you’d need to replace gigabitEthernet <port_number> with 10.110.33.1.
CSR-1(config)# interface gigabitEthernet <port_number>
CSR-1(config-if)# ip wccp 53 redirect in
CSR-1(config-if)# ip wccp 54 redirect in
CSR-1(config-if)# end
Note You can choose not to use an ACL on the Cisco router, thereby allowing all traffic to be
redirected to the appliance. The appliance will send back any traffic that doesn’t match its
policies.
An Alternative Practice
It’s considered a best practice to use separate inbound and outbound ACLs to guarantee maximum
flexibility in configuring redirection. Since a Service Group can only point to one redirect list, and we
are using a pair of service groups (one for TCP and one for UDP), that would require the use of a total of
four service groups if you are also doing inbound (WAN-to-LAN) redirection.
Here is a sample configuration for that scenario:
! Example with separate ACLs for WAN and LAN side redirects
CSR-1(config)# configure terminal
Then, later, when you’re configuring WCCP on the redundant Silver Peaks, it’s useful to force the same
flow to the same Silver Peak—in both directions—to avoid asymmetry. So, for each protocol (TCP,
UDP), a given flow would have an outbound Service Group’s Assignment Detail configured for
lan-ingress, and an inbound Service Group’s configured for wan-ingress. Again, this brings the total
number of Service Groups to four.
Appliance A1
2 For the username and for the password, enter admin. The initial configuration page appears.
Read it, and click Next.
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
3 Select the MAC addresses for wan0 and mgmt1. Make sure that the addresses match the MAC
addresses associated with the vNICs in the hypervisor client.
Make sure that the addresses match the MAC addresses associated with the virtual interfaces of the
Silver Peak virtual machine (VM).
For example, in the VMware client, you would check on the Virtual Machine Properties page.
4 Click Apply & Next. The License & Registration page appears. Enter the license details.
5 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
6 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Router and
then click +Add to add a WAN interface.
Configure the appliance data path IP next-hop router address and max WAN bandwidth.
8 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
9 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
11 Click Done. The appliance saves the settings and reboots automatically.
Configuring WCCP on A1
As a best practice, and for easier ACL management, we’ll configure two Service Groups (TCP, UDP) for
outbound traffic redirection and two Service Groups (TCP, UDP) for inbound traffic redirection.
Note ACTIVE - Designated will be the state for one Silver Peak appliance — this is the device
that owns the communication for WCCP with the routers.
1 Click Add Service Group. The new WCCP Service group appears with the default settings.
2 In the Service Group ID field, enter the WCCP Service Group number you entered on the router as
a placeholder for the TCP protocol. On the router, we entered 53.
3 In the Admin field, accept the default of up.
4 In the Protocol field, leave tcp selected.
5 To access the Service Groups Advanced Settings page, click Advance Settings.
a In the Forwarding Method field, select either. Either allows the appliance and the router to
negotiate the best method for assignment. That is, GRE (Generic Routing Encapsulation) or L2.
b In the Weight field, keep the max default value of 100.
c In the Assignment Method field, leave the default of either. Either allows the appliance and the
router to negotiate the best method for assignment. That is, hash or mask.
d Leave Force L2 Return deselected.
e In the Password field, optionally enter a password.
f In the Assignment Detail field, select lan-ingress.
• If you’re not configuring the tunnel traffic for auto-optimization, then accept the default of
lan-ingress. This is the assumption made for this example, since all redirection will be from
the LAN to the WAN.
• wan-ingress assignment detail is only required when redirection is needed from the WAN
to the LAN, when using TCP/IP auto-optimization.
• custom is used to provide granular control of flow distribution. Contact Silver Peak
Technical Support for assistance.
g To save the settings and close the dialog box, click OK.
6 From the Interface field, select wan0.
7 For Compatibility Mode, select the option appropriate for your router. If a WCCP group is peering
with a router running Nexus OS, then the appliance must adjust its WCCP protocol packets to be
compatible. By default, the appliance is IOS-compatible.
8 In the Router IP Address field, enter the IP address of the WCCP router, 10.110.31.1.
9 Click Apply. A new WCCP Service Group for TCP appears.
a In the Forwarding Method field, select either. Either allows the appliance and the router to
negotiate the best method for assignment. That is, GRE (Generic Routing Encapsulation) or L2.
b In the Weight field, keep the max default value of 100.
c In the Assignment Method field, leave the default of either. Either allows the appliance and the
router to negotiate the best method for assignment. That is, hash or mask.
d Leave Force L2 Return deselected.
e In the Password field, optionally enter a password.
f In the Assignment Detail field, select lan-ingress.
• If you’re not configuring the tunnel traffic for auto-optimization, then accept the default of
lan-ingress. This is the assumption made for this example, since all redirection will be from
the LAN to the WAN.
• wan-ingress assignment detail is only required when redirection is needed from the WAN
to the LAN, when using TCP/IP auto-optimization.
• custom is used to provide granular control of flow distribution. Contact Silver Peak
Technical Support for assistance.
g To save the settings and close the dialog box, click OK.
6 From the Interface field, select wan0.
a For Compatibility Mode, select the option appropriate for your router. If a WCCP group is
peering with a router running Nexus OS, then the appliance must adjust its WCCP protocol
packets to be compatible. By default, the appliance is IOS-compatible.
b In the Router IP Address field, enter the IP address of the WCCP router, 10.110.31.1.
7 Click Apply.
• The data entry area disappears, and the table displays the new WCCP Service Group for UDP.
• State changes from INIT to ACTIVE, DESIGNATED.
• This means that the WCCP protocol is working properly with the router, and that this appliance
is Primary and Active.
State Definition
Note ACTIVE - Designated will be the state for one Silver Peak appliance — this is the device
that owns the communication for WCCP with the routers.
1 Click Add Service Group. The new WCCP Service group appears with the default settings.
2 In the Service Group ID field, enter the WCCP Service Group number you entered on the router as
a placeholder for the TCP protocol. On the router, we entered 55.
3 In the Admin field, accept the default of up.
4 In the Protocol field, leave tcp selected.
5 To access the Service Groups Advanced Settings page, click Advance Settings.
a In the Forwarding Method field, select either. Either allows the appliance and the router to
negotiate the best method for assignment. That is, GRE (Generic Routing Encapsulation) or L2.
b In the Weight field, keep the max default value of 100.
c In the Assignment Method field, leave the default of either. Either allows the appliance and the
router to negotiate the best method for assignment. That is, hash or mask.
d Leave Force L2 Return deselected.
e In the Password field, optionally enter a password.
f In the Assignment Detail field, select wan-ingress.
wan-ingress assignment detail is required when redirection is needed from the WAN to the
LAN, when using TCP/IP auto-optimization.
g To save the settings and close the dialog box, click OK.
6 From the Interface field, select wan0.
7 For Compatibility Mode, select the option appropriate for your router. If a WCCP group is peering
with a router running Nexus OS, then the appliance must adjust its WCCP protocol packets to be
compatible. By default, the appliance is IOS-compatible.
8 In the Router IP Address field, enter the IP address of the WCCP router, 10.110.31.1.
9 Click Apply. A new WCCP Service Group for TCP appears.
a In the Forwarding Method field, select either. Either allows the appliance and the router to
negotiate the best method for assignment. That is, GRE (Generic Routing Encapsulation) or L2.
b In the Weight field, keep the max default value of 100.
c In the Assignment Method field, leave the default of either. Either allows the appliance and the
router to negotiate the best method for assignment. That is, hash or mask.
d Leave Force L2 Return deselected.
e In the Password field, optionally enter a password.
f In the Assignment Detail field, select wan-ingress.
wan-ingress assignment detail is only required when redirection is needed from the WAN to the
LAN, when using TCP/IP auto-optimization.
g To save the settings and close the dialog box, click OK.
6 From the Interface field, select wan0.
a For Compatibility Mode, select the option appropriate for your router. If a WCCP group is
peering with a router running Nexus OS, then the appliance must adjust its WCCP protocol
packets to be compatible. By default, the appliance is IOS-compatible.
b In the Router IP Address field, enter the IP address of the WCCP router, 10.110.31.1.
7 Click Apply.
• The data entry area disappears, and the table displays the new WCCP Service Group for UDP.
• State changes from INIT to ACTIVE, DESIGNATED.
• This means that the WCCP protocol is working properly with the router, and that this appliance
is Primary and Active.
State Definition
Appliance A2
2 For the username and for the password, enter admin. The initial configuration page appears.
Read it, and click Next.
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
3 Select the MAC addresses for wan0 and mgmt1. Make sure that the addresses match the MAC
addresses associated with the vNICs in the hypervisor client.
Make sure that the addresses match the MAC addresses associated with the virtual interfaces of the
Silver Peak virtual machine (VM).
For example, in the VMware client, you would check on the Virtual Machine Properties page.
4 Click Apply & Next. The License & Registration page appears. Enter the license details.
5 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
6 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Router and
then click +Add to add a WAN interface.
7 Configure the appliance data path IP next-hop router address and max WAN bandwidth.
9 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
10 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
12 Click Done. The appliance saves the settings and reboots automatically.
Configuring WCCP on A2
For an Active-Active deployment, we’ll configure the four WCCP Service Groups on A2 with the same
options and values used for A1:
To ensure that this appliance (IP address 10.110.31.101) shares the traffic equally with A1, we’ll
also accept the default weight of 100.
Note that Service Groups with outbound redirection use lan-ingress, and Service Groups with
inbound redirection use wan-ingress.
Across all groups, the Router IP Address is 10.110.31.1.
To configure WCCP on A2
You’ll be completing the same steps as you did for configuring the A1 appliance. For a review, see
“Configuring WCCP on A1” on page 218.
1 Go to Configuration - WCCP and select Enable WCCP.
2 Create two Service Groups for outbound redirection.: 53 for TCP, and 54 for UDP
3 Create two Service Groups for inbound redirection.: 55 for TCP, and 56 for UDP
4 Verify that the State of each WCCP Service Group changes from INIT to ACTIVE.
Note IMPORTANT — When configuring for flow redirection, the mgmt1 interfaces need to
be in a separate subnet from the mgmt0 interfaces.
An appliance that handles both directions of traffic for a flow can then optimize the flow properly.
Specifically, this sets the stage for TCP acceleration and CIFS acceleration.
This sequence of four diagrams illustrates how the need for flow redirection arises, and is resolved.
a Select Enable.
b In the Interface field, select mgmt1.
c Click Add Peer, and enter the IP address of mgmt1 on A2. In this example, it’s 10.10.10.2.
d Click Apply.
e Save the changes.
3 Select Configuration > Flow Redirection. The Flow Redirection page appears.
a Select Enable.
b In the Interface field, select mgmt1.
c Click Add Peer, and configure the IP address of mgmt1 on A1. In this example, it’s 10.10.10.1.
d Click Apply.
e Save the changes.
4 To verify that flow redirection is working, look to see that the State changes to OK, indicating that
the interfaces and flow redirection are configured properly on both sides.
Appliance B
2 For the username and for the password, enter admin. The initial configuration page appears.
Note At any future time, you can always access the Initial Config Wizard by going to the
Configuration menu and selecting Initial Config Wizard from the drop-down menu.
3 Read it, and click Next. Select the MAC addresses for lan0 and wan0. Make sure that the addresses
match the MAC addresses associated with the virtual interfaces of the Silver Peak virtual machine.
For example, in the VMware client, you would check on the Virtual Machine Properties page.
4 Click Apply & Next. The License & Registration page appears. Enter the license details.
5 Click Apply & Next. The Management Interface (mgmt0) page appears. Enter the appliance name
and management IP details.
6 Click Apply & Next. The Deployment Mode page appears. Under Deployment, select Bridge and
configure the appliance data path IP next-hop router address and max WAN bandwidth.
a Leave Auto Tunnel and Auto Subnet Sharing deselected. We’ll take care of these features later.
Although it’s not technically necessary to deselect either one, we have chosen to do so for
tutorial purposes later in this chapter.
b Do not add tunnels. We’ll manually add remote appliances and create tunnels later.
8 Click Apply & Next. The Date/Time Setting page appears. Configure the time zone and NTP server.
9 Click Apply & Next. The Change Admin Password page appears. If necessary, create a new username
and password.
11 Click Done. The appliance saves the settings and reboots automatically.
Tip Prior to putting a bridge mode appliance in production, it is always a good practice to test
connectivity with the appliance in bypass to make sure that the network will function in the event
the Silver Peak device fails to wire.
3 To ensure that local routing is working correctly, ping an address on the subnet from which Site A’s
router will be redirecting traffic. Here, that subnet is 10.110.33.0/24.
To do that, use the same ping screen, specify either an address of a device or the router’s address in
that subnet, and ping with the -I option, as shown.
Note You could have selected Auto Subnet Sharing in the Initial Config Wizard, instead of
doing this step. We do it here to highlight how the Subnet table changes after tunnels come up.
To enable subnets on A2
1 On Appliance A2, select Configuration > Subnets. The Subnets tab appears. Set the configuration.
To enable subnets on B
1 On Appliance B, select Configuration > Subnets. The Subnets tab appears. Set the configuration.
Creating Tunnels
From each appliance, you must create a tunnel to each remote appliance to which it will be sending
traffic.
We’ll create tunnels from Appliances A1 and A2 to B. Then we’ll create tunnels from B to A1 and to A2.
a In the Name field, assign a locally significant name for the tunnel.
b In the Admin field, accept the default value, Up.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager prefills the IP address for the local appliance.
e In the Remote IP address field, enter the data path IP address of the remote Silver Peak
appliance.
f Leave Auto Max BW selected, so the appliance uses the lower of the two system bandwidths.
If you wanted to configure this manually, then you would deselect Auto Max BW and, in the
Max BW field, enter the maximum bandwidth for this tunnel. The value must be less than or
equal to the upstream bandwidth of your WAN connection.
g Leave the Min BW at its default, 32 [Kbps].
h Click Apply.
i Save the changes.
The tunnel status won’t change to Up until a tunnel is configured at both ends. That is, until after
we configure a tunnel from B to A1.
a In the Name field, assign a locally significant name for the tunnel.
b In the Admin field, accept the default value, Up.
c Leave Auto MTU selected. This allows the tunnel MTU to be discovered and negotiated
automatically. When selected, this overrides the MTU setting.
d In the Local IP field, the Appliance Manager prefills the IP address for the local appliance.
e In the Remote IP address field, enter the data path IP address of the remote Silver Peak
appliance.
f Leave Auto Max BW selected, so the appliance uses the lower of the two system bandwidths.
g Leave the Min BW at its default, 32 [Kbps].
h Click Apply.
i Save the changes.
The tunnel status won’t change to Up until a tunnel is configured at both ends. So, we’ll now
configure a tunnel from B to A1.
a In the Name field, assign a locally significant name for the tunnel.
b Enter the Remote IP address (that is, the data path IP address of Appliance A1).
c Click Apply.
4 To add a tunnel to Appliance A2, click Add Tunnel.
a In the Name field, assign a locally significant name for the tunnel.
b Enter the Remote IP address (that is, the data path IP address of Appliance A1).
c Click Apply.
d Save the changes.
Within a few seconds, the Status of both tunnels should change to Up - active.
Click Refresh, if required.
Now that the tunnels are up, the appliances can begin advertising subnet information to each other.
Now that Appliance B has learned the remote appliances’ subnets, it automatically places packets
with destinations in the learned subnets into the correct tunnels.
Notice that the subnet where Site A’s end devices reside — the 10.110.33.0 subnet — does not
appear in the table.
This is because the Silver Peaks at Site A don’t have an interface with an IP address in that subnet.
As a result, the local Silver Peaks at Site A can’t advertise this subnet to Appliance B. We need to
configure A1 and A2 to advertise this subnet to other Silver Peaks.
• Notice that subnets learned from peer 10.110.31.100 ( A1) have a metric of 40, while others
were learned with a metric of 50. When Appliance B has a choice of two routes to a subnet, it
will prefer to send packets to the device having the lower metric. For subnet 10.110.33.0,
Appliance B will always route packets to A1 because it has the lower metric.
• If Appliance A1 goes down, the subnets it advertises disappear from the table, and Appliance B
will use the route advertised by peer A2 (10.110.31.101).
Verifying Traffic
Here, we want to verify that the tunnels are carrying traffic and that flows are being optimized.
Appliance A1
Appliance A2
Status column indicates whether a Click the icon for more information on which Silver
flow is being optimized or not. Peak technologies are being applied to the flow.
When the connection to the peer is functioning, the State column displays OK.
Best Practices
Tips for Deployment
Inbound WCCP redirection is preferred over outbound [also known as ingress/egress] redirection
because inbound redirection is less CPU-intensive on the router. Inbound redirection is done in
hardware where as outbound is done in software.
• For Catalyst 6000/76xx deployments, use only inbound redirection to avoid using “redirection
exclude in”, which is not understood by the switch hardware and must be processed in software.
• For Catalyst 6000/76xx deployments, use L2 redirection for near line-rate redirection. Silver
Peak appliances automatically negotiate assignment and forwarding methods with all routers
and L3 switches from Cisco to the best possible combination that the router or L3 switch
supports.
WCCPv2 interception forwards all packets from the router or L3 switch to the appliance. Special
care should be taken when traffic redirected to the appliance has to be returned back to the router or
L3 switch. For many routers the return traffic is delivered via L2 so there is no CPU impact.
However, Catalyst 6000/76xx switches returns via GRE so the CPU can be negatively impacted
unless Force L2 return is enabled on the appliance.
• Force L2 Return should only be enabled when the interface/VLAN that the appliance is
connected to is not also an interface with the redirection applied to.
The appliance should always be connected to an interface/VLAN that does not have redirection
enabled – preferably a separate interface/VLAN would be provided for the appliance.
The appliance and Catalyst switch negotiate which redirect and return method to use when the
service group is formed. There can be many access VLANs on the aggregation switches. Redirection
is configured on all VLANs that need optimization. Layer 2 switching ports, including trunk ports,
are not eligible for redirection.
If Auto Optimization is used for matching traffic to be optimized via the appliance, WCCP
redirection must also be applied on the uplinks of the router or L3 switch to the core/WAN.
If WCCP redirection is needed on both the WAN and the LAN, the preferred configuration on the
appliance is to set the WCCP group configured on the WAN to wan-ingress and the group
configured on the LAN to lan-ingress.
• The configuration of wan-ingress and lan-ingress ensures that load balancing is symmetrical in
both directions of a flow.
• wan-ingress uses the destination address for distribution in the router/L3 switch table
• lan-ingress uses the source address for distribution.
If Route Policies are used for matching traffic to be optimized via the appliance, WCCP redirection
is not required on the core uplinks, only the access/LAN links. If Active/Active redistribution is
enabled with route policies, then flow redirection is required to handle asymmetrical flows caused
by load balancing. Flow redirection can handle millions of flows and ensures that the owner of a
given flow always receives the TCP flow for proxy.
GRE
GRE is a protocol that carries other protocols as its payload:
In this case, the payload is a packet from the router to the appliance. GRE works on routing and switching
platforms. It allows the WCCP clients to be separate from the router via multiple hops. Because GRE is
processed in software, router CPU utilization increases with GRE redirection. Hardware-assisted GRE
redirection is available on the Catalyst 6500 with Sup720.
L2 Redirection
L2 redirection requires the appliance to be in the same subnet as the router or switch (L2 adjacency).
The switch rewrites the destination L2 MAC header with the appliance MAC address. The packet is
forwarded without additional lookup.
L2 redirection is done in hardware and is available on the Catalyst 6500/7600 platforms. CPU
utilization is not impacted because L2 redirection is hardware-assisted; only the first packet is
switched by the Multilayer Switch Feature Card (MSFC) with hashing.
After the MSFC populates the NetFlow table, subsequent packets are switched in hardware. L2
redirection is preferred over GRE because of lower CPU utilization.
There are two methods to load balance appliances with L2 redirection: hashing and masking.
1.877.210.7325
+1.408.935.1850
www.silver-peak.com