Cisco Presentation

© 2008 Cisco Systems, Inc. 1

Cisco Data Center Solutions

Wayne Simms Business Development Manager, Data Center

Jared Case Consulting Systems Engineer, Data Center

Date: September 27th, 2009

2© 2008 Cisco

Cisco Data Center 3.0

Data Center Evolution

3© 2008 Cisco


The Typical Enterprise Data CenterEnterprise Applications and Services

CommunicationsApplications

ComputingInfrastructure

NetworkingInfrastructure

Facilities(Power, Cooling,Cabling, and Physical Security)

StorageInfrastructure

“Siloed”, Low Utilization, Independent Operational Processes, Consistent Security, Sustainable BCP, SOA Difficulties, Power, HVAC

4© 2008 Cisco


Data Centers are under Increasing Pressure

Collaboration SLA MetricsEmpowered User Global Availability Reg. Compliance

New Business Pressures

Power & Cooling ProvisioningAsset Utilization Threat Prevention Bus. Continuance

Operational Limitations

“70% of typical IT budgets are allocated to run existing IT

applications and infrastructure, leaving only 30% available for new

initiatives.”(IBM Global CEO study, 2008)

5© 2008 Cisco


Conventional Data Center Model

New applications trigger build out of dedicated server, network, and storage infrastructure

Separate teams build and provision; leads to a slow, linear process

Tight Coupling of Apps and Infrastructure makes it tough to have capacity ahead of time

Growth capacity purchased for each application; no way to leverage unused capacity to other applications

Apps

Servers

Network

Storage

Business Need

Dedicated-FunctionPhysical Silos

6© 2008 Cisco


Convention leads to challenges & costs

Numerous inter-dependencies between Computing, Storage, and Networking for the Data Center– Virtualized computing needs shared storage– Network provides bandwidth, mobility, and expansion to Virtualized

computing– Shared storage requires network infrastructure (switches, cables,

circuits, etc.)

A comprehensive Architecture not only reduces risk, but creates leverage

7© 2008 Cisco


Data Center and Network EvolutionIT

Rel

evan

ce a

nd C

ontr

ol

Application Architecture Evolution

Data Center 1.0Mainframe

CENTRALIZED

Data Center2.0Client-Server and Distributed Computing

DECENTRALIZED VIRTUALIZED

Data Center 3.0Service Oriented and Web 2.0 Based

Consolidate

Virtualize

Automate

8© 2008 Cisco


Defy Conventional Wisdom

A Virtualized, Dynamic Data Center is like a three-legged stool…– If any of the legs are too short, it falls– Each leg is equally important

Changes or optimizations to one leg can improve the value of one or two of the others

More than ever, your individual choices have enterprise impact

Computing

9© 2008 Cisco


Cisco DC 3.0

10© 2008 Cisco


Increased resource utilization

Decreased power and cooling

Faster provisioning

Higher availability

Business continuity

Policies

Management

Security

Processes

Data center islands

Virtualization brings Great Benefits..…but it is not a “Free Lunch

11© 2008 Cisco


VirtualizationPlatform

ComputePlatform Network

Platform

Site Cost HVAC Power Dwelling

Platform Cost Storage Network Software Server

Organization Cost Complexity

VM Administrator Coordination

Costs

CostsCosts

Virtualization Has Been Promised As the Answer. However, Virtualization Solutions to Date May Only Address Part of the Problem, but Has Done So by Increasing Operational Expenses, Infrastructure Complexity, and Risk.

High ComplexityHigh Touch

Data Center Virtualization in Today’s EnvironmentIT Organizations Must Weave Together Complex Network, Compute,

Virtualization and Management Software

12© 2008 Cisco


Cisco experience with virtualization?

Virtual load balancing: Cisco ACE Virtual firewalls: Cisco ASA Virtual security (FW/VPN/IPS): Cisco ASA Virtual routing with MPLS Virtual SANs Virtual LANs (VLANs) Office-in-a-box (router/switch/voice/security/wireless):

Integrated Services Routers VoIP

Cisco has been a pioneer of Network Virtualization

13© 2008 Cisco


Overview

Unified Fabric• Enterprise-class top-

of-rack switch• Designed for server

connectivity• Lossless Low latency• I/O Consolidation• Standards-based

Unified Computing

• Platform for stateless computing and virtualization

• Multi-rack architecture• Form factor

independent• Enterprise-class x86• Standards-based

VN-Link

• Virtualization aware access layer

• Compatible with switching platforms

• Combine VM and physical network operations

• Standards-based

14© 2008 Cisco


VN-Link

15© 2008 Cisco


VMW ESX

Server

Nexus 1000V - VEM

VM #1

VM #4

VM #3

VM #2

What Can A Profile Contain?

Policy definition supports: VLAN, PVLAN settings ACL, Port Security, ACL

Redirect Cisco TrustSec (SGT) NetFlow Collection Rate Limiting QoS Marking (COS/DSCP) Remote Port Mirror (ERSPAN)

Nexus 1000V

VSMVirtual Center

16© 2008 Cisco


Accelerate Server VirtualizationBenefits of the Nexus 1000v

Security and Policy Enforcement

Operation andManagement

OrganizationalStructure

Enable VM-level security and policy

Scale the use of VMotion and DRS

Simplify management and troubleshooting with VM-level visibility

Scale with automated server & network provisioning

Enable flexible collaboration with individual team autonomy

Simplify and maintain existing VM mgmt model

17© 2008 Cisco


Data Center Ethernet (DCE)

18© 2008 Cisco


Unified Fabric

A Unified fabric is the end state network where LAN, SAN, and IPC traffic are converged onto a single network infrastructure

FCoE is the enabling technology for delivering a unified fabric and I/O interfaces. It provides seamless integration with existing FC SAN environments

19© 2008 Cisco


Why Unified Fabric?

Consolidation of Infrastructure– Cabling– Switches– Adapters

TCO Reduction– Equipment costs– Operational Costs

Enables Virtualization– Unified ports– Wire once & Walk away– Optimized for virtual machine

environments

Server

Management

Production Active

Production Standby

Clustering

SAN A

SAN B

Backup

SAN Edge

LAN Access

20© 2008 Cisco


Key Benefits of Unified Fabric

Reduce overall DC power consumption Extend the lifecycle of current data center

Wire hosts once to connect to any network Faster rollout of new applications & services.

Every host will be able to mount any storage target Improve Data management & resilience

Ubiquitous, scalable connectivity enables Virtual Machine portability

21© 2008 Cisco


Unified Compute Services

22© 2008 Cisco


Unified Computing System

22© 2008 Cisco Systems, Inc. All rights reserved. Cisco ConfidentialPresentation_ID

A single system that unifiesCompute: Industry standard x86Network: Unified fabricVirtualization: Control, scale, performanceStorage Access: Wire once for SAN, NAS, iSCSI

Embedded management Increase scalability without added complexityDynamic resource provisioningAbility to integrate with broad partner ecosystem

Energy efficient Fewer servers, switches, adapters, cables Lower power and cooling requirements Increase compute efficiency by removing I/O and

memory bottlenecks

23© 2008 Cisco


Key Differentiators of Unified Computing System

Unified Fabric: Decrease spending on power, cabling and management

Memory Expansion: Leading the industry in memory capacity

Service Profiles: Associate global policies throough server state retention and provisioning

UCS Manager: Control the entire UCS infrastructure

24© 2008 Cisco


Mgmt Server

Our SolutionMgmt ServerMgmt Server Embed management

Unify fabricsOptimize virtualizationRemove unnecessary – switches,– adapters,– management modules

Less than 1/2 the support infrastructure for a given workload

25© 2008 Cisco


Mgmt Server

Our Solution: Unified Computing SystemA single system that encompasses:– Network: Unified fabric– Compute: Industry standard x86– Storage: Access options– Virtualization optimized with VMware vSphere 4.0

Unified management model– Dynamic resource provisioning

Efficient Scale– Same effort for 1 or 320 blades

Lower cost– Fewer servers, switches, adapters, cables– Lower power consumption– Fewer points of management

26© 2008 Cisco


Our Solution: Unified Computing System

27© 2008 Cisco


SAN B

Our Solution: Unified Computing System Ideal Platform for Cloud Infrastructure—Single, scalable, integrated

Infrastructure (Network + Compute) virtualization; Storage framework

Dynamic resource provisioning

Mgmt SAN ALAN

28© 2008 Cisco


From cabling to your Data Center organization – UCS simplifies

From ad hoc and inconsistent…

…to structured, but siloed, complicated

and costly……to simple, optimized and

automated

What does your Data Center organization look like?

29© 2008 Cisco


SAN BMgmt SAN ALANBoot Policy

Service Profiles

Server Requirement

VHBA’s Operational Policies VNIC’s

•Name•Identity•Local Storage Policy•Firmware Update•Stats Policies•Scrubbing Policy

•Boot Devices•Boot Order

•Specific Blade•Blade Pool•Qualification Criteria

•Name•Identity•Fabric Connectivity•Configuration

•Scrub Policy•External Mgmt

•Name•Identity•Fabric Connectivity•High Availability•QOS Policy•Configuration

Service Profiles can be associated to any available server in a Unified Computing System which automatically includes full migration of Identities, firmware, and connectivity to LAN and SAN, etc

30© 2008 Cisco


UCS Workload Scalability via VMware vSphere 4.0

More VMs per Server = Lower power per VM Lower cooling per VM Lower cost per VM

Cisco Value Add Hypervisor Bypass

Cisco Value Add Memory Expansion

CPU

Mem

ory

VM VMVM

VM

VMVM

VM

VMVM

VM

Cisco Value VN-Link: NIV

10Gb Unified Fabric: DCE/FCoE

31© 2008 Cisco


Efficiency Through Architectural Innovation Increase efficiency by reducing

number of components– Unified Fabric & Fabric extender

Fewer switches, Fewer adapters

– Expanded memory Fewer servers, Fewer CPUs

– Embedded management Fewer points of management Coordinated control

– Integrated VMware Virtualization

Simplified design– Fewer components– More reliable

Customer Benefits– Lower CapEx – Lower OpEx– Increased business agility

32© 2008 Cisco


33© 2008 Cisco


Intel Case Study - FCoE Cost Analysis (per rack)

Estimated savings per rack = $20,400* (14%)

Assumptions:1. Comm & storage traffic on both 10Gb port2. 2 x 1Gb ports on MB (one port will continue to be used for maintenance)

*Does not include power, maintenance, and support costs

Standard Top of Rack 10Gb Top of RackQty Cost Qty Cost

Servers in a Rack 20 20Quad cards per Server 1 $400.00 0Dual Port HBA 1 $1,500.00 0Cat 5/6/7 cables 6 $180.00 1 $30.00Fibre Cables 2 $120.00 0SFP + Copper 0 2 $250.00GigE Ethernet Switch Port 6 $2,640.00 1 $440.00FC Switch Port 2 $2,400.00 0Dual Port CNA's 0 1 $1,500.00N5K Ports 0 2 $4,000.00

Sub-total (per host) $7,240.00 $6,220.00 TOTAL (per rack) $144,800.00 $124,400.00

https://intel.wingateweb.com/US08/scheduler/controller/catalogTitle: Realizing Benefits of Unified Networking: Deploying Fibre Channel over Ethernet (FCoE)

https://intel.wingateweb.com/US08/scheduler/controller/catalog�

Cisco Presentation

Documents

data center evolution

data center virtualization

dynamic data center

network evolution data

data centers

data center2

data centerdate

cisco systems