Virtualization 101

Timothy Stephan, Senior Director

vSphere Product Marketing
Introduction

Goals for this session

• Gain a fundamental understanding

of VMware’s core virtualization
technology
• Understand the customer benefits of
virtualization
• Explore how vSphere components
work together to form a complete
cloud infrastructure platform.
• Teach you how to get started today!

2
Agenda

What Exactly is Virtualization?

Who is VMware? What are We Trying to Solve?

How Will Virtualization Benefit Me?

Getting Started: vSphere Hypervisor and VMware Go

3
Agenda

What Exactly is Virtualization?

Who is VMware? What are We Trying to Solve?

How Will Virtualization Benefit Me?

Getting Started: vSphere Hypervisor and VMware Go

4
Virtualization Defined

Virtualization ● virt·tual·iz-a-tion ● [vur-choo-uhl-iz-ey-shuhn]

–noun
1. Today’s x86 computer hardware was designed to run a
single operating system and a single application,
leaving most machines vastly underutilized.
Virtualization lets you run multiple virtual machines on
a single physical machine, with each virtual machine
sharing the resources of that one physical computer
across multiple environments. Different virtual
machines can run different operating systems and
multiple applications on the same physical computer.
2. Something only VMware can do well :)

5
Virtualization & Cloud = Top Priorities for CIOs

CIO Technology 2010 Priority 2009 Priority

Virtualization 1 3

Cloud Computing 2 14

Source: Gartner CIO study, Q4 2009

6
Virtualization Defined

For those more visually inclined…

Traditional Architecture Virtual Architecture

7
Key Properties of Virtual Machines

Partitioning
 Runmultiple operating systems on
one physical machine
 Divide system resources between
virtual machines

8
Key Properties of Virtual Machines : Continued

Partitioning
 Run multiple operating systems on one physical machine
 Divide system resources between virtual machines

Isolation
 Faultand security isolation at the
VMware hardware level
 Advanced resource controls preserve
performance

9
Key Properties of Virtual Machines : Continued

Partitioning
 Run multiple operating systems on one physical machine
 Divide system resources between virtual machines

Isolation
VMware  Fault and security isolation at the hardware level
 Advanced resource controls preserve performance

Encapsulation
 Entirestate of the virtual machine can
be saved to files
 Move and copy virtual machines as
easily as moving and copying files

10
Key Properties of Virtual Machines : Continued

Partitioning
 Run multiple operating systems on one physical machine
 Divide system resources between virtual machines

Isolation
VMware  Fault and security isolation at the hardware level
 Advanced resource controls preserve performance

Encapsulation
 Entire state of the virtual machine can be saved to files
 Move and copy virtual machines as easily as moving and copying files

Hardware Independence
 Provision or migrate any virtual machine
to any similar or different physical server

11
Abstraction + Pooling = Reduced Complexity

Traditional View Virtual

Infrastructure

Exchange File/Print
Operating System Operating System Virtual Infrastructure
OS
Virtualization
OS
Virtualization
CPU
Pool

Memory
Pool
SAP ERP Oracle CRM
Storage
Operating System Operating System Pool
OS
Virtualization
OS
Virtualization

Interconnect
Pool

12
Abstraction + Pooling = Reduced Complexity

Exchange File/Print
Operating System Operating System Virtual Infrastructure

CPU
Pool

Memory
Pool
SAP ERP Oracle CRM
Storage
Operating System Operating System Pool

Interconnect
Pool

13
Agenda

What Exactly is Virtualization?

Who is VMware? What are We Trying to Solve?

How Will Virtualization Benefit Me?

Getting Started: vSphere Hypervisor and VMware Go

14
VMware is the Customer Proven Market Leader

 Company Overview
• $3.77 billion in 2011 revenues
• >$4.5 billion in cash and equivalents
• 30%+ operating margins
• ~11,000 employees worldwide
• 5th largest infrastructure software
company in the world
 Proven in the Trenches
• 300,000+ VMware customers
• 100% of Fortune 100
• 100% of Fortune Global 100
• 99% of Fortune 1000
• 97% of Fortune Global 500

15
VMware is the Customer Proven Market Leader

 Company Overview
• $3.77 billion in 2011 revenues
• >$4.5 billion in cash and equivalents
• 30%+ operating margins
• ~11,000 employees worldwide
• 5th largest infrastructure software
As
company in theof mid-2011,
world about 81%
 Proven inof
theinstalled
Trenches VMs are VMware-based
• 300,000+ Gartner,
VMware customers
June 2011
• 100% of Fortune 100
• 100% of Fortune Global 100
• 99% of Fortune 1000
• 97% of Fortune Global 500

Source: From Virtualization to Cloud Computing, Gartner IT Infrastructure, Operations and Management Summit, June 2011, Thomas Bittman

16
Our Mission – Aligns with your Mission

Enable Our Customers By

Simplifying and Automating IT in the Cloud Era

Leading to a
Optimizing IT Enabling Greater More Engaged,
Efficiency… Business Agility… Productive
Workforce

Enabling Business Transformation thru IT Transformation

18
VMware History – the Road to Cloud Infrastructure
Gen 1: Gen 2: Gen 3: Gen 4:
Client Hypervisor Server Hypervisor Virtual Infrastructure vSphere

Focus is on the
application
App App App App App App App App App App

OS OS OS OS OS OS OS OS OS OS

App App App App App

VMware Server (GSX)
OS OS OS OS OS
VIP
HOST OS ESX Server
Foundation of a
Private Cloud VIP

Central management Central management

via vCenter via vCenter

• Higher utilization • Complete OS/App fault • Centralized • Policy-based control of

• Some OS/App fault isolation management compute, network, and
storage resources
isolation • Encapsulation of virtual • Move virtual machines
machines while running • Security and fault tolerance
• Hardware Independence • Automatic business • Focus is on the application
continuity

19
Our Focus

Empowered, Secure
End-User Computing
Mobile Workforce

Faster
Time-to-Market
for Modern
Cloud Application Platform
Applications

A More Flexible,
Scalable, Efficient
Infrastructure Cloud Infrastructure & Management
for All Apps

Existing Datacenters Public Cloud Services

20
 VMware in the Leaders Quadrant for x86 Server Virtualization

Gartner, Inc. Magic Quadrant for x86 Server Virtualization Infrastructure, Thomas J. Bittman, Philip Dawson, Mark Margevicius, George J. Weiss, 30 June 2011.
This Magic Quadrant graphic was published by Gartner, Inc. as part of a larger research note and should be evaluated in the context of the entire report. The Gartner report is available upon request from VMware.

The Magic Quadrant is copyrighted 2011 by Gartner, Inc. and is reused with permission. The Magic Quadrant is a graphical representation of a marketplace at and for a specific time period. It depicts Gartner's analysis of how certain vendors measure
against criteria for that marketplace, as defined by Gartner. Gartner does not endorse any vendor, product or service depicted in the Magic Quadrant, and does not advise technology users to select only those vendors placed in the "Leaders" quadrant.
The Magic Quadrant is intended solely as a research tool, and is not meant to be a specific guide to action. Gartner disclaims all warranties, express or implied, with respect to this research, including any warranties of merchantability or fitness for a
particular purpose.

21
VMware and Our Partner Network…

Deep Expertise… Economics of Choice…

2,200+ Technology and 70+ vCloud

Consulting Partners Service Providers

Collectively committed to achieving transformative business

outcomes through continuous innovation and superior technology.
Together we have the complementary skills and shared commitment
to bring the best expertise and business solutions to each unique
customer environment.

Leverage Existing
Available Resources…
Relationships…

25,000+ 59,000+ VMware

Channel Partners Certified Professionals

22
 Broadest Software and Hardware Support in the Industry

Over 1400 ISV Partners Supporting Broadest ISV and IHV Support
over 2500 Applications on VMware
Majority ISVs Support VMware
 SAP and IBM have clear support
10 out of Top 10 statements for running their software on
VMware
Healthcare ISVs

 Oracle’s official support policy (Document

10 out of Top 10 ID #249212.1) states they will support
Finance ISVs Oracle Databases, applications &
middleware virtualized on VMware

 Microsoft fully supports all leading

5 out of Top 5 applications on VMware (and recently
Telecom ISVs extended support for Exchange 2010
DAG)

4 out of Top 5
Broadest x86 Hardware Support
Retail ISVs
 The broadest choice of x86 hardware
ISV Independent Sodtware Vendors
23
So, What’s the Problem?

© 2010 VMware Inc. All rights reserved

IT Mandate

IT will always be pressured to …

enable greater business agility,

while becoming more efficient.

25
 IT is Traditionally Forced to Focus on Non-Value-Add Activity

IT Investment
Overwhelming complexity
5% +
Infrastructure
Investment Brittle infrastructure
=
< 30% of IT budgets
23%
Application 42% goes to innovation and
Investment Infrastructure competitive advantage
Maintenance

30%
Application
Maintenance
Business Agility
Depends on IT Agility

Source: VMware Fortune 100 Customers

26
Before VMware: The State of at IT Infrastructure

Server Sprawl Power & Cooling

 36M physical x86 servers by  $1 for every $1 spent on
2011— a ten-fold increase servers
over 15 years  $29 bn in power and
 $140 bn in excess server cooling industry wide
capacity - a 3-year supply

Space Crunch Costs Operating Costs

 $1,000 / sq ft  $8 in maintenance for
 $2,400 / server every $1 spent on new
infrastructure
 $40,000 / rack
 20-30 : 1 server-to-admin
ratio

1. IDC, U.S. and Worldwide Server Installed Base 2007–2011 Forecast, Doc #207044, May 2007
272. IDC, Virtualization And Multicore Innovations Disrupt The Worldwide Server Market, Doc #206035, March 2007
3. Source: VMware
Obviously something isn’t working very well…

…so what should we do?

28
Agenda

What Exactly is Virtualization?

Who is VMware? What are We Trying to Solve?

How Will Virtualization Benefit Me?

Getting Started: vSphere Hypervisor and VMware Go

29
Core Benefits

1. Reduce 2. Dramatically 3. Enable Flexible, Agile

the Complexity Lower Costs IT Service Delivery
to simplify operations to redirect investment into to meet and anticipate the
and maintenance value-add opportunities needs of the business

Drive IT Agility to Increase Business Value

30
 How Do I Get Those Benefits?

Consolidation - One-time event that moves existing applications

onto a fewer number of servers

Containment - An ongoing effort to virtualize new applications

and manage growth of existing ones

Availability – Introducing virtualization to increase application

availability and data recoverability

…there are many more benefits of virtualization

What is vSphere?

Industry’s Most Complete and Robust

Virtualization Platform

vSphere Provides:
 Infrastructure consolidation and cost savings
 Application high-availability and business continuity
 Simple, yet powerful IT management

What Does vSphere Solve for Customers?

 Provides a Powerful Disaster Recovery Strategy
 Supports Running Mission-Critical Applications
 Simplifies IT Operations
 Delivers the Most Flexible Building Block for Leveraging
the Cloud
VMware vSphere Platform

vCenter Server

• High Availability
• Fault Tolerance
• vShield Zones
• vMotion • Hot Add
• Security APIs
• Storage vMotion • DRS
• Data Recovery
Application
Services

Availability Security Scalability

VMware vSphere

Compute Storage Network

Infrastructure
Services • Thin Provisioning
• Network I/O Control
• ESXi Convergence • Storage DRS
• Distributed Switch
• Auto Deploy • Profile-Driven Storage
• DRS • Storage I/O Control
• Storage APIs

DRS Distributed Resource Scheduler

High Availability

Description:
Enables the high availability of
virtual machines by restarting

X
them on a different vSphere
host in the event of a failure

Benefits: Resource Pool

• Minimizes downtime and IT
service disruption
• Reduce cost and complexity
compared to traditional
clustering
Live Machine Migration with Zero Downtime

Description:
Enables the live migration of virtual
machines from one host to another
with continuous service availability.

Benefits:
• Revolutionary technology that is the
basis for automated virtual machine
movement
• Meets service level and performance
goals

35
Better Storage Utilization and Efficiency

Description:
Provisioning storage only based on what is
needed now and grow into requested size
over time

Benefits
Benefits:
• Eliminate over-allocating storage
• Reduce CapEx purchases
• More granular controls over
storage resources
• Easy to convert from virtual disks
that were previously thick
(Storage vMotion)
Save Time During Disaster Recovery

Physical
Configure Install Configure Install Start “Single-step
hardware OS OS backup automatic recovery”
agent
Virtual
Restore Power
40+ hrs
VM on VM
< 4 hrs

Eliminate recovery steps

• No operating system re-install or bare-metal recovery
• No time spent reconfiguring hardware

Standardize recovery process

• Consistent process independent of operating system and hardware
The Results are Transformational

Financial Resources Human Resources Earth’s Resources

“Strategically, virtualization leads

inexorably down a path toward agility,
flexible sourcing and cloud computing.”
Tom Bittman, Gartner

Capital costs
reduced by
50% - 60% Average of 33%
reduction in routine
Delayed data admin time
Up to 80%
center expansion
E.g. provision reduction in
Operational costs a server in minutes datacenter
reduced by 25%+ energy costs

Source: Gartner; 29 July 2010 Q&A: Six Misconceptions About Server Virtualization, Tom Bittman.

38
Agenda

What Exactly is Virtualization?

Who is VMware? What are We Trying to Solve?

How Will Virtualization Benefit Me?

Getting Started: vSphere Hypervisor and VMware Go

40
vSphere Hypervisor - Overview

A Free Version of VMware’s Leading ESXi Architecture

Partitions a server into virtual machines

Reduces hardware, power, and cooling with the
performance and features of ESX

Plug-and-Play
Minimal configuration. Run VMs in minutes

OS-Independent, thin architecture

Unparalleled security and reliability

Full-featured
Superior consolidation and scalability

Ease of Use
Easy to manage with remote tools
Free download from vmware.com
Simple license upgrade to vSphere editions
 Hardware Requirements
Server
CPU
Minimum: Single socket, dual core
CPU
Ideal: Dual-socket, 4 or more cores per CPU
Memory:
Minimum: 2GB
Ideal: 8+GB
Memory
Network
Minimum: one NIC, plus one for Management interface
Ideal: One for Management Interface plus multiple NICs for
VMs
Storage Network
Local Storage (SATA/SAS):
Minimum: one 80GB drive
Ideal:2 mirrored drives (only for ESXi Installable)
plus 4 RAID5 drives for VMs
Storage
ESXi 4.0 Installable may be installed on 1GB+ USB 2.0 flash
storage device
Shared Storage
NFS, iSCSI, Fibre Channel: for VM storage
ESXi Installable requires local disk or 1 GB+ USB 2.0 flash
storage for the hypervisor
Installing vSphere Hypervisor

vSphere Hypervisor vSphere Hypervisor VMware Go

Embedded Installable
• Installed via SD flash or • Load Installer via CD • Go to
USB key internal to the or ISO image “go.vmware.com”
server • Simple 2-step 1. Register for an account
• Distributed with a new procedure: 2. Accept EULA
server 1. Accept EULA 3. Walk through
• No Installation -- Just 2. Select local drive for installation wizard
Turn It On! installation
Installing vSphere Hypervisor: Using VMware Go
VMware Go UI

vSphere Installation wizard:

VMware Go - FREE

• Simplified
installation with
HCL verification
• Configuration &
management
• Installation, HW/SW
inventory tracking
Get up and running in Minutes


1. Power on server , which boots
into hypervisor
2. Configure Admin Password
 3. (optional) Modify network
configuration
 4. Connect via vSphere Client
Point your browser to the configured
IP Address
Download & Install Windows-based
 vSphere Client
Start vSphere Client and log into
host
Manage Your ESXi Host: vSphere Client
Interacting with Virtual Machines

Virtual Machine Console

Or just use standard remote

access, e.g. RDP, ssh, etc.
Creating Virtual Machines

Use VMware Converter

Transfer existing physical
servers into virtual
machines
Import existing VMware and
3rd party virtual images

Create from Scratch

Specify CPUs (1-4),
Memory (1-256 GB),
Disks, Network interfaces
Load OS from ISO image
(over 80 supported OSes)

Import a Virtual Appliance

Hundreds to choose from on the
Virtual Appliance Marketplace
Download directly via vSphere
Client and deploy on host
Manage Individual VMs

Control capabilities
• Power on/off
• Suspend
 Captures state of VM to disk
 Can resume back to same state
• Snapshot
 Makes point-in-time copy of
virtual disk(s)
 Can have multiple snapshots in
a tree
Monitor Individual VMs
Easily Upgrade vSphere Hypervisor

vSphere Hypervisor vSphere – Kits or Editions

• FREE • Pools of computing resources
• Single server partitioning • Centralized management
• Production-class hypervisor • Built-in automation, availability and
• Advanced server resource manageability
management • All Kits or Editions include ESXi, starting
at $83 per CPU socket

The hypervisor is to Virtual Infrastructure what the engine is to a car, or the

BIOS to a PC – an required component, but not the whole solution.
VMware vSphere Deployment Architecture

• Deploy ESXi on each host

• Add vCenter Server to
Centrally manage ESXi
hosts
• Upgrade license file to
vSphere
VMware vSphere Storage Appliance:
Delivering shared storage capability to all SMBs
Traditional IT Environment VMware vSphere Storage Appliance

VSA runs as a virtual machine, with

your application workloads

Shared storage VSA uses server

hardware provides internal disk – no VSA clusters storage
centralized arrays shared storage across server nodes
of storage over a hardware
high-speed network required

Traditionally shared storage hardware VSA delivers high availability and

is required to use high availability and advanced vSphere features to SMBs
advanced vSphere features without shared storage hardware

53
Simplify Management with vCenter Operations

Overview
 Comprehensive dashboard for
infrastructure operations with
health, risk and efficiency scores
 Single tool to manage
performance, capacity and
configuration changes across
entire environment

Benefits
 End-to-end visibility into
infrastructure health
 Proactively ensure quality of
service and incident avoidance
 Optimize for efficiency and cost

54
In Summary

Virtualization reduces cost, improves high availability,

and simplifies operations

ESXi is a high performance, enterprise-class hypervisor

architecture

vSphere Hypervisor can be up and running with VMs in

minutes and is available for free. VMware Go can help!

vSphere Hypervisor can be upgraded to VMware

vSphere to deliver high availability and reduce OpEx
and CapEx.
Questions?
APPENDIX
 Visibility into Immediate and Potential Future Problems

Operations Short and Long Term Capacity

• Immediate • Forward
problems Looking
• What is • Are there
happening areas that I
right now? should be
• What do I concerned
need to pay about from a
attention to? capacity
perspective?
• Have I
deployed my
VI in the
most efficient
manner?

58
Ducati
PROFILE
“ We chose VMware as they are the leader in the market. When
you choose VMware, you know that you have a solution which
is always going to support you.”
- Daniel Bellini, CIO, Ducati
Industry Automotive
Objectives
Headquarters Bologna, Italy
Employees 1,100
•To meet their requirements in terms of growth and
compete/win against must larger brands, Ducati needed to
Annual Revenue €403 million
find a partner with the capabilities to support them.
VMware Solution
VMWARE PRODUCTS & SERVICES
•Cloud infrastructure has enabled Ducati to link it’s racing
and street bike businesses.
• VMware vSphere
•Ducati’s Cloud is 100% integrated on VMware technology
• VMware vCloud Director
• VMware View Business Impact
•0%-to-98% virtualized in less than 2 years, all with no down
time and supporting a 3x increase in production.
•Ducati’s IT infrastructure supports the production of 40K
bikes per year, and helps them outrace competitors
significantly larger.

59
 Centaur Fund Services
Profile “As a start up business, choosing the right IT infrastructure for our
business was absolutely crucial. We needed a scalable, resilient and
affordable solution... We’re 100% confident we made the right choice in
selecting VMware to provide our data center and desktop technology.”
— Karen Malone, Managing Director

Customer Objectives
 Create a highly resilient and scalable IT infrastructure without
Industry Finance
significant initial investment
Headquarters Dublin, Ireland
 Provide a robust remote working solution to reduce downtime
Employees 100+

Solution
The Numbers
 Deployed vSphere 4 for high availability and enhanced business
• 2 data centers in the UK continuity as well as a secure environment for cost effective growth
• Manages $200 billion in assets  Used View 4 for remote working capability
• 100% of server environment virtualized

Business Impact
 Reduced downtime to zero
VMware Products & Services  Increased speed of responses to client requests
 Maximized utilization of servers to reduce capital and operating costs
• vSphere 4  Increased ability to access applications remotely
• View 4

60
 Threadneedle
Profile “We have made huge savings across our IT…VMware’s server and
desktop virtualization solutions have enhanced the performance and
availability of our key business applications and our business
continuity and disaster recovery plans have been transformed.”
— Phil Whittock, Head of Infrastructure Architecture

Customer Objectives
 Maintain high availability for operational and regulatory purposes
Industry Finance  Reduce costs and streamline operations by consolidating data centers
Headquarters London, UK  Improve security and application delivery for end-users
Employees 100+

Solution
The Numbers
 Deployed VMware virtualization to consolidate data centers and
provide high availability for applications
• 5 data centers in the UK
• Manages £59 billion ($90 billion) in assets
 Used View to enhance desktop application and remote IT delivery
• Achieved 10:1 server consolidation
Business Impact
 Saved 6-figures in cost avoidance
VMware Products & Services  Virtualized over 80% of applications to improve delivery
 Reduced cost of ensuring high availability by 90%
• VMware Infrastructure 3  Reduced server provisioning by 3X
• View  Enhanced overall business agility

61
Start Virtualization Today
 How to get started
• vSphere Eval Download link: https://www.vmware.com/tryvmware/?p=vmware-vsphere5-
ent&lp=default

• vSphere Eval Guide: https://www.vmware.com/files/pdf/products/vsphere/VMware-vSphere-

Evaluation-Guide-1.pdf

• VMware Hardware Compatibility Guide:

http://www.vmware.com/resources/compatibility/search.php?lp=default

• vSphere Quick Start Classes: http://www.vmware.com/a/webcasts/details/260?lp=default

• vSphere Product Documentation: https://www.vmware.com/support/pubs/vsphere-esxi-vcenter-
server-pubs.html

 Additional resources
• vSphere Product Demos: https://www.vmware.com/tryvmware/?p=vmware-vsphere5-ent&lp=default
• VMware ROI/TCO Calculator: http://roitco.vmware.com/vmw/
• VMware Cost per App Calculator: http://www.vmware.com/technology/whyvmware/calculator/
• Get Started with vSphere Kits: http://www.vmware.com/products/datacenter-
virtualization/vsphere/small-business/compare-kits.html

62
Introducción al NFV

Network Function Virtualization

Para qué virtualizar la Red?
• Baja de costos apalancado en las economías de escala de la industria
del IT. (Manteniendo un Nivel de Servicio Carrier Class).

• Flexibilidad y rapidez para provisionar e iniciar servicios en diferentes

lugares sin tener que instalar nuevos equipos.

• Mejorar el Time to Market

• Mejorar la eficiencia operacional basada en la ventaja de contar con

una plataforma de red homogenea (física)

• Reducción de costos por: Reducción de consumo energético, espacio

ocupado y una mejora en el network monitoring

• Innovar creando un ecosistema abierto que posibilite contar con

nuevos proveedores.

Basado en la Fuente: ETSI EXPLAINING NETWORK FUNCTION VIRTUALIZATION

Prodip Sen Chair of the ETSI Industry Specification Group (ISG) on NFV
18/Marzo/2014
En qué consiste la virtualización de funciones de la red?
Es la combinación de dos conceptos:

• NFV: Es la separación del hw del sw (soportando funciones de red en

software sobre hardware de propósito general)
• SDN: Es la separación del plano de control del plano de datos,
pudiendo cambiar el comportamiento de los conmutadores de
forma externa.

Conjuntamente se pretende conseguir separar:

las funciones de red la capacidad para albergarlas

pudiendo gestionarse y escalarse por separado

Virtualización de Redes es la combinación de dos elementos:
NV = NFV + SDN

NFV
CONTROL CG-NAT POOL
BNG
MGMT
SW-defined Network Functions
 Separation of HW and SW
 No vertical integration
Pool - HW vendor ≠ SW vendor ≠ Mgmt vendor
admin
 Once network elements are SW-based,
Session
mgmt
UPnP HW can be managed as a pool of
IPv4 / NAT resources
IPv6
NAT
TR-069 DHCP ctrl.

SDN

Interconnecting Virtual Network

Functions (a.k.a. backplane)
 Separation of control and data plane
 Easy orchestration from SW domain
NFV permite hacer más flexible la red aprovechando que la capa de
HW es común para las diferentes aplicaciones.

Las funciones de la red están totalmente definidas por SW,

minimizando la dependencia y los condicionamientos del HW.

DPI CG-NAT GGSN/ VIRTUAL NETWORK

BRAS SGSN
FUNCTIONS FUNCTION
Firewall PE Router

Decoupled

COMMON HW
(Servers & Switches) CAPACITY
NFV ayuda a reducir la complejidad de la Gestión de Red porque el HW
puede ser tratado como un pool de recursos.

APPLIANCE APPROACH VIRTUAL APPLIANCE APPROACH

• Node sizing is determined by the • HW becomes interchangeable and

bottleneck of its functionalities aggregatable (pool)
• Capacity growth often leads to • Resource assignation becomes
node growth or silo HW purchase fully flexible and dynamic

SESSION MGT LIMITATIONS PER NODE PROCESSING CAPACITY BECOMES

LEADING TO 2nd NODE PURCHASE Vs. COMMODITY & MANAGED AS A CONTINUUM

SPARE CAPACITY FOR

EXTRA GROWTH
SESSION MGT (in any functionality)
LOAD = 95% LOAD = 40%
RESOURCES
SESSION MGT

SWITCHING
RESOURCES
SWITCHING
LOAD = 40% LOAD = 15%
 NFV Management and Orchestration

Os-Ma
NFV
OSS/BSS
Orchestrator

Or-Vnfm

EM 1 EM 2 EM 3 Ve-Vnfm Service, VNF and

VNF
Infrastructure
Manager(s) Description
VNF 1 VNF 2 VNF 3

Vn-Nf Vi-Vnfm
NFVI
Virtual Virtual Virtual
Computing Storage Network
Nf-Vi Virtualised Or-Vi
Virtualisation Layer Infrastructure
Vl-Ha Manager(s)
Hardware resources
Computing Storage Network
Hardware Hardware Hardware

Execution reference points Other reference points Main NFV reference points
NFV Reference Lab:
Fomentar el ecosistema / Evitar soluciones cerradas en un sólo vendor
7

x% of new
“network
element” to be
virtualized by
year 20YY
 AT&T Touts Its First Virtualized Functions
LightReading, 6/9/2015; CHICAGO -- Big Telecom Event -- AT&T has
been public with its bold proclamation that it will virtualize and software-
control 75% of its network by 2020, transforming a total of7 200 network
functions. And, according to the man helping to lead the charge, it's well
on the way to virtualizing the first 5% this year alone.
Those functions include:
1) vEPC, (AT&T already has customers and traffic running on)
2) SDN-controlled Network On-Demand service (now serving 100 cities), which lets
businesses change Ethernet network services and speeds on the fly.
3) Mobile call recording as a fully virtualized service (It as good traction with banks
and brokerage houses).
There are all also on AT&T's short-term virtualization agenda:
4) vDNS
5) vNetwork analytics,
6) vData platforms
7) vEdge Routers
 AT&T’s Stephens: More Than 40% of Network Functions Are Virtualized

SDxCentral, 26/7/2017
AT&T is well on its way to reaching its network virtualization goals. AT&T
CFO John Stephens told investors during the company’s second 7 quarter
earnings call yesterday that AT&T has virtualized more than 40 percent of
its network functions and is making progress toward its goal of virtualizing
55 percent of its network by year-end.
Stephens also attributed the company’s commitment to software-defined
networking (SDN) and network functions virtualization (NFV) to helping it
keep a tight reign on its costs.
 Qué virtualizar? Sobre qué aplicar NFV…

Sentido Común!
(Que se pueda hacer no implica que deba realizarse…)
• CPEs
• GWTs de MBH
• GGSN/CGSN
• SBC
• BRAS
• CG-NAT
In the case of CPEs, virtualisation allows to accomplish service migration
smoothly, while home operation and service deployment are simplified

Network functionalities at home equipment are minimised, making them as simple and stable
along the time as possible, whereas quick evolution can be supported by SW inside
operator’s network.
FROM…
Home environment Network environment

STB
CPE

NAT TR-069
IPv4 UPnP FW

DHCP Access Point Switch Modem

…TO Network environment

Home environment
STB
CPE UPnP FW

IPv4/IPv6
Access Point Switch Módem TR-069
DHCP NAT

15
15 15
Several alternatives exist…

Scenarios

1
Evolutionary approach, in which vCPE leverages on currently deployed
infrastructure (BRAS) while integrates virtualization technologies for some
specific vCPE functions

2
vCPE virtualized platform: Every vCPE function is virtualized and runs on x86
infrastructures. In particular, IPFE is implemented as a NF with the required
subset of functionalities for vCPE.

3
Fully virtualized network: vCPE IPFE function is integrated in a full SW vBRAS
that exploits full benefits of virtualization, both for PPPoE and IPoE users

NF: Network Function 16

16 IPFE: IP Front End
16
Evolutionary approach Scenario 1

Scenario BSS/OSS vCPE Orchestrator

1 vCPE NFs over general purpose HW WEB

Users
DHCP NAT
vCPE integrates currently Portal
1
deployed technologies (e.g.
BRAS) as long as virtualizaion
technologies (DHCP, NAT)
Operator Core
• Already implemented in MX960
1
current pilot (limited
scalability) ME60
MX9601
NAT

IPv4/IPv6 IPv6/IPv4
IPFE Core

ALU8950
AAA

NF: Network Function17

17 17
vCPE virtualised platform Scenario 2

Scenario BSS/OSS vCPE Orchestrator

vCPE NFs over general purpose HW WEB

2 IPv4/IPv6 Users
DHCP NAT
vCPE virtualized: Every vCPE IPFE Portal
function is virtualized and 2 1 2
runs on x86 infrastructures,
in which IPFE is implemented IPv4 over “X”

as a NF with the required

subset of functionalities

Operator Core

IPv6/IPv4
Core

ALU8950
AAA

NF: Network Function18

18 18
Fully virtualised network Scenario 3

Scenario BSS/OSS vCPE Orchestrator

vCPE NFs over general purpose HW WEB

3 Users
DHCP NAT
Fully virtualized: vBRAS IPv4/IPv6 Portal
Full vBRAS
function integrating full BRAS 1 2 3
functionality into a NF,
exploiting full benefits of 3
virtualization, both for PPPoE
and IPoE users

• Proposals from traditional Operator Core

and new-entrant industry L2 CPE

NAT

L3 CPE

IPv6/IPv4
Core

ALU8950
AAA

NF: Network Function19

19 19
Benefits for the Telco

• Network Virtualization will change our networks in the future

• As first use case, vCPE Project is complex enough in technology and
organization to give us hints about virtualization technology benefits and
changes
• Centralization: Internal plant vs external plant
• Proactive vs reactive operation
• Better home network diagnostics
• Better TTM for service evolution, migrations, troubleshooting, …
• Transformation CAPEX is reduced (COTS HW servers and SW licenses)
• Operational savings will arise in short time, paving the way to STB
virtualization and new Services

DISCOVER, DISRUPT, DELIVER

Industry interest on use cases (non exhaustive)

Use case Operators Vendors

vEPC AT&T, KDDI, NTT Docomo, Optus, SK Telecom, Connectem, Affirmed Networks, NSN,
SoftBank, T-Mobile, Verizon Wireless, Vodafone Samsung, Huawei, ZTE, Ericsson

vIMS DT, China Mobile Alcatel Lucent, Huawei, Mavenir, ZTE,

Ericsson
vCDN BT, Orange Akamai
vBRAS BT 6WIND, Cisco, NEC, Calsoft Labs, Huawei
vCPE BT, COLT, Orange, Telstra Calsoft Labs, NEC, Ericsson
vSTB DT, KPN Huawei
Enterprises firewalls and NTT Com Virtela
WAN accelerators

Carrier SDN Comcast, NTT, Telenor, KVH Telecom ConteXtream, Alcatel Lucent, Ciena,
Brocade, Cyan
Network Orchestration AT&T, DT, KDDI, Verizon Wireless Tail-f, Ericsson, x-ion, Cyan, Alcatel
Lucent
Service Chaining NTT Docomo, Telstra, Verizon Wireless Cisco, Juniper, HP, Ericsson,
ConteXtream, Huawei
Cloud RAN China Mobile Alcatel Lucent,Wind River
vRAN SK Telecom
Video Traffic Optimization Verizon Wireless

Video Transcoding Comcast Vantrix, Kontron

17 PoCs have been accepted so far by ETSI:
CloudNFV Open NFV Framework
Service Chaining for NW Function Selection in Carrier Networks
Virtual Function State Migration and Interoperability
Multi-vendor Distributed NFV
E2E vEPC Orchestration in an multi-vendor open NFVI environment
Virtualised Mobile Network with Integrated DPI
C-RAN virtualization with dedicated hardware accelerator
Automated Network Orchestration
VNF Router Performance with DDoS Functionality
NFV Ecosystem
Multi-Vendor on-boarding of vIMS on a cloud management framework
Demonstration of multi-location, scalable, stateful Virtual Network Function
SteerFlow: Multi-Layered Traffic Steering for Gi-LAN
ForCES Applicability for NFV and integrated SDN
Subscriber Aware SGi/Gi-LAN Virtualization
NFVIaaS with Secure, SDN-controlled WAN Gateway
Operational Efficiency in NFV Capacity Planning, Provisioning and Billing
NFV ISG PoCs participating in the ETSI NFV Poc ZONE:

PoC#1 - CloudNFV Open NFV Framework Project

PoC#2 - Service Chaining for NW Function Selection in Carrier Networks

PoC#4 - Multi-vendor Distributed NFV

PoC#5 - E2E vEPC Orchestration in a multi-vendor open NFVI environment

PoC#7 - C-RAN virtualisation with dedicated hardware accelerator

PoC#9 - VNF Router Performance with DDoS Functionality

PoC#12 - Demonstration of multi-location, scalable, stateful Virtual Network Function

PoC#13 - SteerFlow: Multi-Layered Traffic Steering for Gi-LAN

PoC#15 - Subscriber Aware SGi/Gi-LAN Virtualization

PoC#18 - VNF Router Performance with Hierarchical Quality of Service Functionality

PoC#19 - Service Acceleration of NW Functions in Carrier Networks

PoC#21 - Network Intensive and Compute Intensive Hardware Acceleration

Virtualización del Gi-LAN
Virtualización del Gi-LAN
Virtualización del Gi-LAN
 vCPE: Virtualisation allows to accomplish service migration smoothly,
while home operation and service deployment are simplified
Network functionalities at home equipment are minimised, making them as simple and stable along
the time as possible, whereas quick evolution can be supported by SW inside operator’s network.

FROM…
Home environment Network environment

STB
CPE

NAT TR-069
IPv4 UPnP FW

DHCP Access Point Switch Modem

…TO Network environment

Home environment
STB
CPE UPnP FW

IPv4/IPv6
Access Point Switch Módem TR-069
DHCP NAT
RFQ UNICA Infra – Modules
Modules Telco Cloud Solution

28
RFQ UNICA Infra – Modules
Ericsson Solution for Telco Cloud Solution

29
Introduction to Brocade’s vRouter 5600

Brocade vRouter 5600 is a router VNF, intended to support line rate speeds in
a fully virtualised environment.

Commercial version to be available before the end of the year. Early Access
version available in Telefónica for data plane evaluation only.

Data plane is based on Intel’s DPDK, while control CONTROL PLANE BASED
plane is based on enhancements over Open Source ON OPEN SOURCE
projects (Linux and Zebra, mainly). PROJECTS
(Zebra + Linux)
This slide set shows the results of data plane
performance tests conducted at TEF’s NFV
Reference. OPTIMIZED
DATA PLANE
The vRouter was deployed in the lab following (DPDK-based)

the recommendations from ETSI NFV-PER001

for assurance of high and predictable performance
in an NFVI.
Technical details

HW and SW configuration
Item Description
HW CPU: 2x Intel ® Xeon ® CPU E5-2697 v2 (2x 12 cores)
Memory: 64 GB
I/O: 4 NIC Intel X520 (82599 controller), with 2 10GE ports/NIC, evenly distributed among sockets
Host OS & Hypervisor Host OS: CentOS 6.5
Kernel version: 3.12.15 (CentOS6.5 config + mod for IOTLB support)
Hypervisor info: Libvirt 0.10.2, QEMU 0.12.1
Hugepages configuration: 56 GB (28 1G hugepages per socket)
VNF Model: Brocade Vyatta vRouter 5600 3.2 R0S4
NFV-PER requirements: 22 HW threads (11 cores), 4GB RAM memory (backed by host 1G hugepages),
4x 10GE ports in passthrough, all in a single socket
NFV-O & VIM In-house TEF’s NFV-O+VIM

Test details
Item Description
VNF configuration Number of routes: 8, 32k and 512k, evenly distributed among 4 ports, with respective network masks
/4, /16 and /20
Routes statically injected (no functional tests of routing protocols)
Traffic generation Tool: tidgen (in-house traffic generator)
Traffic profile:
• Packet sizes: 64, 96, 128, 192, 256, 384, 512, 768, 1024, 1500
• 75% TCP, 25% UDP
• Number of flows per port: 100k, 1M, 10M, evenly distributed along the route space
• Duration: 10 minutes per trial
VNF evaluation summary & Next steps

VNF EVALUATION

• Performance tests show 100% line rate throughput (no losses) with packet
size ≥ 192B
 Up to 80Gbps using all available ports in a COTS server
• Insensitive to lookup table size under realistic working conditions
 40M flows, 512k routes
• Efficient integration in Lab
 From memory stick to running lab deployment in less than 2 hours
 On site support from Brocade was excellent.

NEXT STEPS

• For the current feature set:

 For data plane, no additional tests are foreseen
 Dynamic control plane capabilities (routing updates, etc.) could be tested using the
commercial version.
• New features, expected in coming releases:
 Firewall functionality could be tested once available
Sobre los Servers:

Para tener en cuenta:

• Presentación de 6WIND (proveedor de servers):

• “Network Bottlenecks Multiply with NFV: Don't
Forget Data Plane Peformance“ (Founder & CEO,
Eric Carmès en “SDN2014”, Set2014)
Ejemplo de virtualización

Orchestrator: Cyan,
Cloudband (De ALU),
ECM Ericsson, etc

vDNS vAAA vBRAS vMail Server

VM sobre VM sobre VM sobre VM sobre
Red Hat 6.0 Red Hat 6.1 Ubuntu (Linux Free) Windows
500G HD 1TB HD 200G HD 200G HD
5GB RAM 10GB RAM 5GB RAM 5GB RAM
1 CPU 2 CPU 1 CPU 1 CPU

vSWITCH

Hypervisor: Para poder dar multitenant (servidores compartidos para

múltiples clientes) vmWare ESXi, kvm (No lleva licencia, Red Hat es
uno de los que lo provee), etc
Server: hp, IBM, Cisco, Dell, Oracle, etc: 2TBytes HD, 128 GB RAM, 8 CPUs (8 cores Intel)
 Servers: qué significa 8 vCPU?

Un proveedor de VNF ha requerido disponer de la siguiente plataforma:

8 vCPU, 16Gb mem, 100 Gb disk

Qué significa 8 vCPU?

Rta: En una tarjeta del Server hp C7000 hay:
2 CPUs con 8 cores cada una (16 en total).

La capa de virtualización (Por ej. el hypervisor KVM) permite dividir esos

cores a vCPU, y asignar los mismos a las Máquinas Virtuales (VM) que
correrán en la tarjeta.
Differences between Cloud Computing and
NFV
Network Virtualisation is not Cloud Computing

The network differs from the computing environment in 2 key factors:

1
NEED OF HIGH AND
Data plane workloads
PREDICTABLE PERFORMANCE
(which are huge!) (as with current equipment)

2
Network requires shape GLOBAL NETWORK VIEW IS
(+ E2E interconnection) REQUIRED FOR MANAGEMENT

…which are big challenges for vanilla cloud computing

…and most of industry is offering to Telcos just IT based

cloud products as network virtualization environments
What are the main differences between NFV and Cloud
Computing?

Goal

Design requirements Orchestration

Workloads
Virtualized
Infrastructure usage infrastructure
management
Infrastructure selection

Physical deployment
Although NFV and Cloud Computing use virtualization
techniques, they have different goals and requirements

NFV Cloud Computing

Main goal Deployment of e2e network services by Offering commodity computing through
composing network functions over a versatile different service models (IaaS, PaaS, SaaS)
infrastructure

FUNCTION
VIRTUAL
DPI CG-NAT GGSN/ (semantics)
BRAS SGSN NETWORK
Firewall PE Router FUNCTIONS

Decoupled

COMMON HW CAPACITY
(Servers & Switches) (resource
mgmt)

Main design Effective decoupling while ensuring carrier Efficiency by maximizing usage of resources
requirement grade performance
Network workloads are different from IT/Cloud workloads and hence
impose specific requirements over the infrastructure

NFV workloads Cloud workloads

Purpose Mainly moving packets (packet processing and Complex applications
forwarding)

Bandwidth Very high (>10Gbps) Low (<1Gbps)

Time Almost instantly (ns) Low (ms)

response
Point of Distributed (>100) Highly centralized (<10)
presence
Storage I/O Low (except for some VNFs like CDN) High or very high

Infrastructure Deterministic allocation of resources for Consolidating workloads over shared

usage ensuring carrier grade performance (high infrastructure for maximizing infrastructure
throughput with low latency) utilization
These specific requirements over the infrastructure make
that computation nodes are assembled in a different way

NFV workloads Cloud workloads

Network I/O Very high Low or medium

Computing Medium or high High or very high

Memory Low or medium High or very high

Storage I/O Low (except for some VNFs like CDN) High or very high

NFV Server Cloud Server

• Less RAM • Lot of RAM
Can anyone
Obviously, these types of servers are imagine a router
• High number • High number
of cores not exchangeable, and once deployed, of cores
do not fit well for different purposes!!
with just two legs?
• Less HDD • Lot of HDD

many interfaces few interfaces of 1Gbps

of 10Gbps or even 10Gbps!!

Infrastructure Requires proper component selection: Component selection is just focused on the
selection Latest generation of processors (Ivy Bridge or aggregated computing power, memory and hard
newer) with HW assist for all I/O and memory disk
operations (IOTLB, DDIO, etc.) and DPDK-
compatible NICs
And the physical deployment is totally different!!

NFV Cloud Computing

Physical A transit point with huge traffic, limited An end point with high traffic, lot of resources,
deployment resources, very distributed and cost limited highly centralized and with a massive cost

WAN WAN

Perimeter
security

To be
Perimeter
developed security

RAW traffic
from customers

MAN
Data Plane deployment

Typical DC deployment
Typical management of cloud infrastructure is not enough
for ensuring a predictable and high performance
NFV Cloud Computing
Virtualized NUMA view is needed to get proper and Hardware infrastructure is managed just as an
infrastructure predictable performance for NFV. aggregated view of resources (CPU, memory,
Avoiding bottlenecks in the hypervisor and OS is etc.)
management
critical, which implies avoiding the usage of Virtual Switching is used to simplify
virtual switching and relying on physical connectivity management
switching

44
NFV adds new abstraction layers that need their own
orchestration
NFV Cloud Computing
Orchestration NFV-O is required to compose network Focus on VMs lifecycle management. No notion
scenarios (orchestrate coherent capacity of neither VNF nor network scenario
allocation - VNFs and their physical
connectivity needs) and to manage the
network scenarios lifecycle

45
NFV vs Cloud Computing in a nutshell

NFV Cloud Computing

Deployment of e2e network services by composing Offering commodity computing through different service
Main goal
network functions over a versatile infrastructure models (IaaS, PaaS, SaaS)

Main design Effective decoupling while ensuring carrier grade Efficiency by maximizing usage of resources
requirement performance

Distributed (>100 pops) Centralised (<10 pops)

Workloads high bandwidth (>10Gbps) per NIC medium bandwidth (<1Gbps) per NIC
Low delay (<us) Medium delay (>ms)

Infrastructure Deterministic allocation of resources for ensuring carrier Consolidating workloads over shared infrastructure for
usage grade performance (high throughput with low latency) maximizing infrastructure utilization

Infrastructure Requires proper component selection: Component selection is just focused on the aggregated
selection Latest generation of processors (Ivy Bridge or newer) with computing power, memory and hard disk
HW assist for all I/O and memory operations (IOTLB, DDIO,
etc.) and DPDK-compatible NICs

Physical A transit point with huge traffic, limited resources, very An end point with high traffic, lot of resources, highly
deployment distributed and cost limited centralized and with a massive cost

Virtualized NUMA view is needed to get proper and predictable Hardware infrastructure is managed just as an
infrastructure performance for NFV. aggregated view of resources (CPU, memory, etc.)
management Avoiding bottlenecks in the hypervisor and OS is critical, Virtual Switching is used to simplify connectivity
which implies avoiding the usage of virtual switching and management
relying on physical switching

Orchestration NFV-O is required to compose network scenarios Focus on VMs lifecycle management. No notion of neither
(orchestrate coherent capacity allocation - VNFs and their VNF nor network scenario
physical connectivity needs) and to manage the network
scenarios lifecycle
Telco vision is virtualizing most of the network functions,
not just cloud-like workloads like control functions…
Data Plane must Control Plane can
be Distributed be Centralised
LOCAL PoPs REGIONAL DATA CENTRES
CDN (*)
v Video (*) SDP (*)
v NGIN (*)
Service
FUNCTION P-CSCF (*) Security (*) IMS (*) SRVCC (*) Domain
(software defined, NFV)
S/PGW (*) BNG (*) PE (*) DHCP (*) MME (*) PCRF (*)

IPv6
Network
DPI (*) CGNAT (*) DNS (*) DRA (*) SPR (*)
Router (*) Domain
HW and SW HW and SW
decoupling decoupling

OS + Hypervisor OS + Hypervisor Infrastructure

COTS HW COTS HW Domain

SDN Switching SDN Switching

CAPACITY
(Homogeneous
infrastructure) Interconnection

Devices
5G++

Places FTTx

Regional Data
Users Access Aggregation Local Points of Presence Core
Centres
Capillarity, Multiplexing Manage users and sessions, Switching, Control functions,
Capacity, Managed local services Transport Managed regional services
Mobility support
(*) Illustrative, since final VNFs deployed will depend on a natural selection processes regarding traffic and
digital world evolution and requirements.
Software Defined Networking (SDN): OpenFlow come to the rescue!
 Public Clouds
• Amazon AWS

• Fibertel

• Azure Microsoft

• GCP Google Cloud Platform

• Claro Cloud

• CenturyLink

• Opencloud