Server and Storage Consolidation

8/2/2019 Server and Storage Consolidation

1/36

The Right Technology. Right Away.

CDWG.com/consolidation l 888.508.4239

Reference Guide

Consolidation&Server Storage


2/36


At CDWG, were committed to

getting you everything you need

to make the right purchasing

decisions rom products and

services to inormation about the

latest technology. Our Reerence

Guides are designed to provide

you with an in-depth look at

topics that relate directly to

the IT challenges you ace.

Consider them an extension

o your account managers

knowledge and expertise, to

help your organization succeed.

We hope you fnd this guide to

be a useul resource.

Table of Contents

What is aCDWG

ReferenceGuide?Reerence Guide

Consolidation Chapter 1: Introduction to Optimization and Consolidation ............................................ 3

Consolidation vs. Optimization

Strategic Drivers

Chapter 2: Capacity and Performance Management.........................................................5 What Is C&PM?

Planning

Determining Current C&PM

Forecasting Monitoring and Reporting

Chapter 3: Optimization ............................................................................................................... 9 Server Optimization

Storage Optimization and Management

Chapter 4: Consolidation Strategies .......................................................................................25 Four Types o Consolidation

Defning a Consolidation Strategy Implementing a Consolidation Strategy

Chapter 5: Virtualization............................................................................................................29 Three Virtualization Models

Calculating the Value o Virtualization

Virtualization Technologies

Making It All Work

Glossary: ................................................................................................................32Index: ................................................................................................................35

Keep an eye on your mailbox orthe CDWG IT Investment Guide.It has even more products andinormation to help with your

consolidation initiative.

[ [


3/36

Today, organizations are in a constant battle to increase serverperormance with smaller budgets. It is generally understoodthat streamlining and consolidating servers and inrastructure willmaximize data-center utilization, enable government agenciesand educational institutions to more quickly respond to changingoperational needs, and help simpliy IT management. However,reaching a state o consolidation is just hal o the battle.

Two o the biggest buzzwords associated with this issue,

optimization and consolidation, are quite a bit more than mereindustry buzz. They are actually inrastructure strategy componentsthat can be planned, executed and ully realized. The purposeo this guide is to cut through the hype and demonstrate howoptimization and consolidation can reap real benefts in yourinrastructure.

Consolidation vs. OptimizationConsolidation and optimization are complementary strategieswhose combined goal is to decrease the total cost o ownership(TCO) o the IT inrastructure while increasing the return oninvestment (ROI) or in laymans terms, to get the most bang orthe IT buck.

Consolidation This is the process o bringing together separate

parts into one unifed whole. Through consolidation, the goal isto reduce the total number o inrastructure components (servers,storage, network) and increase the utilization o those resources tothe maximum possible beneft.

Optimization This involves making something as eectiveas possible. Through optimization, the intent is to ensure theinrastructure components you maintain are reaching their ullestpotential.

Strategic DriversA strategic driver is the motivating orce behind a strategy decision.The purchase o a new e-mail sotware solution, or example, maybe driven by specifc interoperability issues that indicated a needto increase an organizations capability to eectively communicate.The strategic drivers behind consolidation and optimization allinto two major categories the fnancial perspective and theinternal perspective. Both o these strategies help to achieve goalscommonly associated with fnance and internal processes.

to OptimizationIntroduction

IN THIS CHAPTER: Consolidation vs. Optimization

Strategic Drivers

Chapter 1

CDWG.com/consolidation l 3

and Consolidation

Consolidation is theprocess o bringingtogether separateparts into one unifed

whole.


4/36

Financial PerspectiveConsolidating and optimizing servers, storage and inrastructurewill yield fnancial benefts by maximizing the ROI, which is basedon the uture value o an expense or investment in this case,an IT investment. ROI is a metric that attempts to determine whatthe organization will receive in return or the cost o the servers,storage and inrastructure. Through consolidation and optimization,you can reduce the total size o the inrastructure and thusmaximize ROI by minimizing the need or investment.

For example, suppose you are able to reduce the total numbero servers needed or a project by 20 through consolidation omultiple applications onto a single server. You have, through betterIT management alone, saved the cost o 20 servers. This decreasesthe TCO.

The TCO o an IT investment in servers, or example iscalculated by frst measuring how much capital expenditure mustoccur (purchase), then mapping out all other associated costs such

as network connectivity charges and support costs. Reducing thenumber o servers, storage or other inrastructure components willreduce not only the capital expenditures or new or replacementsystems, but also inrastructure and support costs.

I each server needs network connectivity, or example, and yournetwork vendor charges $35 per month per port, then througha reduction o 20 ports, your organization will decrease the TCOby $700 per month ($8,400 per year). When you begin to actorin money saved through reduction in support costs, hardwarereplacements and, in some cases, sotware licenses, a much biggerpicture o TCO reduction emerges.

Internal PerspectiveIn addition to the direct fnancial benefts o consolidatingand optimizing servers, storage and inrastructure, there aremany internal benefts, such as improving (or maintaining)operational productivity and lowering process costs. An optimized

inrastructure is a productive inrastructure and contributes directly

to operations productivity by making the most o the resources

on hand.

Consolidation and optimization will lower process costs bysimpliying the inrastructure management, all o which contributeto increased IT eectiveness and efciency. Having ewerinrastructure components to manage simplifes management,which benefts the organization in terms o employee time and apossible reduction in sotware licensing ees.

Throughout the course o this guide, we will take an in-depth look

at the technologies and strategies available today that will yieldoptimization and consolidation results or your organization. Theseinclude the management process o capacity and perormancemanagement in Chapter 2, optimization in Chapter 3, consolidationstrategies in Chapter 4 and a close look at virtualization inChapter 5. Each chapter builds upon the previous one, and theyhave all been organized to provide a least-commitment to most-commitment view o technology and strategy as they progress.

1CHAPTER

4 l 888.508.4239

The Right Technology.

Right Away.

Simplicity =ProductivityLet CDWG help withconsolidating your serversand storage, and youllbe able to do morewith less.


5/36

To begin consolidating and optimizing inrastructure components,a base level o understanding must be established a oundation

to be used as a starting point. First, establish a clear denitiono what it is that youre trying to do and what you expect to gainrom your consolidation and optimization eorts. Obtaining thesemeasurements alls into the arena o capacity and perormancemanagement (C&PM).

What Is C&PM?C&PM consists o management processes with the common goal o

ensuring that IT and related inrastructure continue to meet currentand uture operational requirements in the most ecient mannerpossible. Whereas capacity speaks specically to the amount oresources currently available (or orecast as needed or uture use),perormance speaks more specically toward the most ecientutilization o resources, such as processing power, memory andstorage. C&PM is primarily concerned with:

Planning The establishment o a planning process to reviewcapacity and perormance management o IT resources that willensure resources are available to meet agreed-upon workloads

Current C&PM The actual current state o C&PM, whichinvolves the review o current workloads to determine whetherresources are available to meet the current need as well asdocumenting and understanding the infuences on resourcedemands, such as business drivers

Future C&PM Conducting C&PM orecasting at regularintervals to ensure that adequate resources are available to meetuture needs

Resource availability Ensuring that resources are availableto meet current demands, taking into account such concerns as

normal vs. peak workloads and contingency demands Monitoring and reporting Monitoring o perormance

and throughput (or load) on a specic piece or area o ITinrastructure, such as an individual server, group o servers,network device or location

PlanningC&PM planning is the overarching plan to understand the needs o

the unit or operational demands as they relate to the demand orcapacity, then taking steps to ensure that adequate capacity existsto meet those needs. C&PM planning is based upon some ormo previously agreed-to understanding o resource needs betweenunit/operations sta and IT personnel.

This understanding may be achieved through a Service LevelAgreement (SLA) or other ocial documentation o goals andmetrics. During the planning process, the service delivery manager

(or IT manager responsible or the establishment o the plan orC&PM) will communicate with unit and operations managersto determine acceptable levels o service, which will encompassboth the planning and provisioning o inrastructure capacity andavailability, as well as documenting these needs.

O course, to gain a clear planning perspective, an organizationmust rst have an accurate inventory o all its IT inrastructureassets. Sotware, such as Altiris Service & Asset ManagementSuite, aids service level management by combining enterprise assetand service management disciplines into a single

andCapacityIN THIS CHAPTER: What Is C&PM?

Planning

Determining Current C&PM

Chapter 2


Forecasting

Monitoring and Reporting

PerormanceManagement


6/36


7/36

> Other Types might include network andnon-real-time storage media, such as tape

Mainframe impact

Current mainrame site Used i multiple mainrames existwithin your organization

Mainrame impact site i known Will a new mainrame sitebe required? Will an existing site be reduced?

Current processing load peak During what hours ooperation is mainrame usage considered peak? (Otenorganizations help control mainrame costs by chargingline-o-business customers slightly more or mainrameusage during peak times than during nonpeak times.)

Current processing load nonpeak

Impact on processing-load peak (orecast) Based uponwhat is understood rom line-o-business partners concerningtheir growth, how do the service delivery or applicationmanagers believe peak-time mainrame processing load willbe impacted (usually indicated in percentage o growth perquarter)?

Impact on nonpeak-time processing load (orecast) Sameas previous but or nonpeak times

Impact on batch processing (i applicable) Based upon

what is understood rom line-o-business partners concerningtheir growth, how do the service delivery or applicationmanagers believe the processing load on mainrame batchprocessing will be impacted (usually indicated in percentage ogrowth per quarter)?

Midrange This data is captured per device, which is importantin understanding the rest o the metrics in this section; eachmidrange device represents its own line item, with processingand memory averages as well as impacts refected individually oreach host

Current midrange server names

Number o new or reduced servers (i known)

Current midrange processing average Average o midrange

processing, usually derived rom processing measurementscaptured during C&PM eorts

Current midrange memory average Average omidrange memory utilization, usually derived rom memorymeasurements captured during C&PM eorts

Impact on processing requirements

Impact on memory requirements

Storage Current mainrame storage requirements

Current midrange storage requirements

Current NAS requirements

Current SAN requirements

Current tape/optical/miscellaneous storage requirements

Forecast mainrame storage requirements

Forecast midrange storage requirements

Forecast NAS requirements

Forecast SAN requirements

Forecast tape/optical/miscellaneous storage requirements

As is evident rom this list, which is ar rom all-inclusive, you stillhave network and perhaps utility computing environments toconsider.

The process o orecasting capacity can quickly becomeoverwhelming. To be successul and accurately orecastrequirements, it is clear that at least two basic questionsneed to be addressed:

How well do you understand how operations impact theenvironment? I your organization provides services to outsideclients or example, students or citizens it is importantto capture what kind o services you oer as well as to whom

and how those services impact inrastructure requirements. The

2CHAPTER



8/36

easiest way to illustrate this is in terms o storage. For example,suppose your organization essentially has three types o internal/external customers. These customers require varying levels odata to be processed and stored by your systems:

> Customer type 1 requires 10MB o disk space over ve years



Once you understand the storage needs per customer or user,you can make the translation between business and storagerequirements. I the line-o-business or unit manager o thisdepartment were to speak with a service delivery manager andidentiy a potential to add 500 new customers or users over thenext three months, you could more accurately orecast and respond

to the need.

How are changes communicated? C&PM should be proactiveand it should be understood by all involved that the sooner the ITservice delivery managers are made aware o a potential impact,the more able the manager will be to respond to the changingneed. Sadly, however, many unit and operations managers havelittle direct communication with those that make the IT decisions.Thus, service delivery managers need to be sure not only that

operations stapersons understand the value o IT and ITstapersons understand their operations colleagues needs, butalso how the two relate.

The sooner IT managers learn o a potential growth requirement,the better able they are to plan and ensure resources are inplace, which translates to decreased time-to-market or services.It is thereore critical to the success o C&PM orecasting that ITmanagers are kept inormed o strategic operations directions.

Monitoring and ReportingMonitoring and reporting are critical processes to C&PM or tworeasons:

Maintenance/forecasting To maintain current C&PMmeasurements that will aid in identiying C&PM shortcomingsand address projected workloads, storage plans and resourceacquisition concerns. Sotware, such as IBMs Tivoli Monitoring

solution, provides monitoring or essential system resources, todetect bottlenecks and potential problems, and to automaticallyrecover rom critical situations.

Monitoring/measurement To monitor and measure the ITorganizations success in meeting agreed-upon standards orC&PM and provide reporting to unit and operations managers.

To provide the maximum value, a C&PM monitoring and reportingplan should be structured to include goals and goal indicators. Forexample, goals or C&PM may include:

Monitor and measure peak-load and transaction response times

Meet response times to Service Level Agreements

These goals may be measured by:

Peak-load and overall utilization rates

Percentage o peaks where target utilization is exceeded

Percentage o response time SLAs not met

C&PM can be as much about managing perceptions as it is aboutactual C&PM data collection. By clearly dening goals, suchas monitor and measure peak load and transaction response

time, and clearly dened measurements such as peak load andoverall utilization rates, a well-dened capacity and perormancemanagement plan begins to emerge.

2CHAPTER

8 l 888.508.4239


Right Away.

Call your CDWGAccount Manageror more inormation

on eectiveconsolidation solutions.

Unsure about how to go about measuring C&PM within

your organization? Goals and metrics or C&PM are

provided in the Control Objectives or Inormation and

related Technology (COBIT), which is available at nocharge rom http://www.isaca.org.


9/36

IT application and service delivery managers are continuallychallenged to deliver the maximum benet rom their existingresources. As the strain on inrastructure components such as

servers and storage devices increases, inrastructure managersneed to determine new ways to deliver more results with the sameamount o resources. Success in meeting these challenges can bemeasured in two words: eectiveness and eciency.

The eectiveness o an IT inrastructure is measured by its abilityto deliver results in a timely, correct, consistent and usable manner.Eciency is ocused on the provisioning o inormation systemsresources through the optimal (most productive and economical)use o resources.

Through optimization, your organization can realize the benets o:

Boosting eciency across your IT environment

Reducing costs

Maximizing storage

Improving/optimizing network utilization

Simpliying management

This chapter takes a close-up look at server and storageoptimization to examine how each can be delivered and measuredin terms o eectiveness and eciency.

Server OptimizationMany organizations today have a large number o serversto consolidate. The combination o client/server computing,distributed/decentralized computing and relatively inexpensive

hardware inexpensive enough to dedicate one box to a given

OptimizationIN THIS CHAPTER: Server Optimization

Storage Optimization and Management

Chapter 3



10/36

task or application has resulted in whats oten reerred to asserver sprawl.

Optimization and all the benets that come with it are great intheory, but how do you actually optimize an inrastructure inpractice? As discussed in the previous chapter, it is importantto tie a servers role to a specic business unction so that thecontribution o the inrastructure is clearly understood and becomes

measurable rom a line-o-business point o view.Optimization o a le and print server, or example, may be quitedierent rom optimizing a Web server. Each kind o server willhave specic requirements. And the inrastructures ability to meetthose requirements is really what youre driving toward whenstriving or optimization.

The rst step toward optimization is to establish a clearunderstanding o what servers are being used or within the

inrastructure and to dene server roles. Generally, servers will allinto one o the ollowing roles:

File and print

Database

Web

E-mail

Management

General application

Organizing servers by roles will allow you to examine each serverrole as an independent entity and establish like baselines. Theidea is to make adequate comparisons that measure the currentperormance and determine the overall level o eciency by role.Determining eciency can be accomplished by monitoring andevaluating inrastructure components and extrapolating a set obaseline measurements that will act as the starting point or theoptimization eort.

Essentially, this is a process o gathering the key perormancepoints that allow servers to add value to your inrastructure orexample, CPU, memory, storage and network utilization. I yourorganization ollows the traditional model, you will typically have alarge number o servers, each purchased to meet its own specicapplication need and used or little else but what they wereoriginally purchased or.

These servers are oten barely scratching the surace o their

perormance potential. Establishing baselines will not only allowyou to know what youre up against in terms o optimization butwill also contribute to consolidation eorts. Server baselines usuallyconsist o at least the ollowing metrics:

CPU utilization

Memory utilization

Disk utilization

Network utilization

3CHAPTER

10 l 888.508.4239


11/36

Many actors go into these metrics and will need to be taken intoconsideration to ensure a air and accurate measurement thatcan be used or management purposes. The model o the server,number o CPUs onboard, CPU processing speed and the amounto memory installed all contribute to the overall perormance o anapplication environment, so you must be certain to compare likesystems with like systems.

Next, you can explore opportunities to ocus on one or two actorsthat contribute to eciency, such as process execution time,memory usage, disk space usage, bandwidth usage or some otherresource. It is important to note that optimization oten requirescompromises to be made in the orm o a tradeo.

In much the same way that a vehicle may be particularly suitedto o-road travel but not well suited or on-road use, servershave particular traits that oten require a tradeo, where one is

optimized at the expense o others. For example, increasing thesize o cache may improve a servers runtime perormance but alsoincreases the memory consumption.

Once measurements have been classied, the next step is toanalyze the results and look or opportunities or improvement.Be sure to involve outside vendors and internal subject matterexperts as you begin to analyze the results. Pay particular attentionto those areas o your server inrastructure that are perormingwell. Determine why they are perorming well, and leverage the

opportunities presented.

Storage Optimization andManagementLike server optimization, storage optimization will requireclassication o roles, some monitoring and an assessment omonitoring results, but is generally less complex rom a hardware

perspective because storage devices dont process data in the

traditional sense o the word. Storage optimization speaksspecically to the most ecient use o storage inrastructure, whichis addressed through a storage discipline known as inormationliecycle management (ILM) and is infuenced by hierarchicalstorage management (HSM) and data replication.

ILMILM is a cradle-to-grave storage strategy that can be deployedto balance the costs o storing and managing inormation in directrelation to the inormations operational value. Much like serverscan be classied into specic server roles, inormation is classiedinto operational values.

When properly exercised, ILM will provide a strategic methodologyor aligning storage costs with organizational priorities. This hasthe potential to lower the TCO o the storage inrastructure throughreductions in hardware and sotware costs. ILM also reduces thestorage complexity by providing only the services necessary toachieve operational objectives.

The success o ILM is heavily dependent on accurate dataclassication. Data classication is the process by which operationalinormation requirements are gathered, then classied rstas inormation and subsequently as data. Once inormation is

classied as data, storage architects can begin to assess therelationship o the data to the current storage media.

HSMHSM is a data storage methodology that automatically moves databetween high-cost and low-cost storage media as needed to aidin storage optimization and cost management. High-speed storagedevices, such as hard disk drives, are typically more expensive thanslower devices, such as optical discs and magnetic tape drives.Although it might be ideal to have all data available at all times,

this setup is oten cost-prohibitive.


3CHA

PTERAccording to various

industry sources, a typicalserver environment

operates at only about 25to 30 percent o capacity.

Elements of Information Lifecycle Management

Operational RequirementsPartners present

operational inormationrequirements

Inormation ClassicationInormation is classied as data

by records managers, dataowners and legal counsel

Data ClassicationData is classied by storage

architects and storageadministrators


12/36

12 l 888.508.4239

In general there are three basic phases o storage:

Online storage Storage that is available on a network orreal-time access.

Near-line storage Storage that can be accessed as neededbut might not be immediately available. Larger-scale documentmanagement systems that utilize an automated tape system areone example o near-line storage: the user may need to wait a

ew minutes or the data to be accessed rom near-line storage.

Ofine storage Storage that is completely ofine and notdirectly accessible, such as tape backups or optical archives.Data stored ofine can only be accessed once it is retrieved andreconnected to the network.

Where inormation resides in any o these storage phases isdictated by its business value, inormation classication and dataclassication.

For example, data that may at one time have had a requirementto be kept online or 90 days may now be discovered through theILM process to actually have a business need o 15 days o onlinestorage, 15 days o near-line storage and 60 days o ofine storage.Because storage prices decrease the urther the data access isremoved rom real time, properly classiying inormation is a centralconcern to storage optimization. Once properly classied, data can

be reassigned to the proper media and thus make optimal use othe storage inrastructure.

Data ReplicationData replication is the copying o data to and rom sites to improvelocal service response times and availability. On the surace, itmay seem that data replication in and o itsel does not directly

Instead o deletingdata, your organization

should considerinormation liecyclemanagement.

3CHA

PTER File Server

(Online Storage)

connected tothe network

readily available

Near-Line

Storage

can be accessedas needed

might not be

Offline

Storage

disconnected fromthe network

not directly accessible

examples: tape backupsimmediately availableor optical archives


13/36

contribute to storage optimization. Actually, it is usually seen as adetractor rom an optimal state because it involves the replicationo data to multiple locations, which is quite inecient use ostorage space.

Optimization, however, is largely driven by eciency as a unctiono cost, and that cost is impacted by more than just the price ostorage. Organizations need the most ecient operation or thebudget available, and when network bandwidth is at a premiumand a large number o users need to access data remotely, datareplication can contribute to storage optimization by reducing thecost o the network inrastructure needed to sustain the fowo data.

For example, i you have seven operations sites geographicallydisbursed around the world, each with 100 users that need toaccess a series o large les, it might be more cost-ecient toreplicate a copy o these les to each o the local sites than tohave each user access the les directly. This illustrates the pointthat sometimes what may seem inecient on the surace mayactually deliver optimal results and contribute to overall storageoptimization.


Right Away.

CDWGs StorageSpecialists canwork with yourAccount Manager

to implement asuccessul consolidationstrategy.


3

CHAPTE

R


14/36

The blade snaps in. Your inrastructure expands.

888.508.4239

CDWG.cm/cd

Easy as1,2,3. Step 1:Choose enClosuresDetermine the number o BladeSystemenclosures needed or your proposedsystem, noting that each enclosure holds

up to eight ull-height or 16 hal-heightserver blades and has two bays ornetwork connectivity. Enclosures nowinclude either single-phase or three-phasepower supplies.

Step 2:Choose blaDes

Choose the specic server blade model(ull-height or hal-height), congurationand quantity to meet your needs. I ahal-height server blade is chosen, one

blade sleeve is required or each pair oserver blades. Hal-height and ull-heightserver blades may be mixed within thesame blade sleeve.

Step 3:Choose an interConneCt kitBeore choosing your interconnectoption, consider the type o storagesystem to be used (DAS, NAS, SAN).This will determine the type ointerconnect necessary.

Anyone who has ever been given the task o running down hundreds ocables, strung through the racks, just to add or remove servers will recognizethe benets o the blade server approach.

Blade servers boost utilization levels rom the 25-30% range to 80-90%.

Instead o a single application using a dedicated server, each applicationpulls the resources it needs rom a pool o servers. Blade servers eciency,fexibility, and cost-eectiveness are signicantly improving the productivityo organizations the world over.

The three-step process to the right highlights the various componentsnecessary to build an HP BladeSystem solution. It is an overview o the basic

component-conguration procedure and is not all-inclusive. We believe youwill nd it useul in understanding the essential, required building blocks oan HP BladeSystem.

Your CDWG team will work with you through every step o the process,helping you build and implement a server-optimization strategy tailored toyour technical and nancial needs, as well as any industry requirements.

To learn more about the benets o blade servers call your dedicated

CDWG account manager today or visit us online at CDWG.cm/pv.

Call your CDWG accountmanager or complete

blade confguration

Enclosure shown wiithboth 460c and 480c bladeservers; all sold separately


15/36

Y CDWG cc mg c p y fd g cd . 15

HP BladeSystem c7000 EnclosureThe HP BladeSystem c7000 Enclosure is a modular, lexible and scalableplatorm that consolidates the essential elements o a data center intoone 10U chassis.

Up to 16 multi-core dual processor server blades

Up to our redundant I/O abrics

5TB backplane to support current and uture I/O connections N+N and N+1 power redundancy or maximum coniguration lexibility

$5027.39CDWG 1000389

Cisco Catalyst Blade Switch 3020

F hP c-C bdsymThe Cisco Catalyst Blade Switch 3020 or HP c-Class BladeSystem provides HPcustomers with an integrated switching platorm that delivers Cisco resiliency,advanced security and enhanced manageability to the server edge while dramaticallyreducing cabling requirements.

Wire speed switching on 16 internal 1Gb ports

CiscoWorks Sotware that includes multilayer eature conigurations such as routingprotocols, ACLs, and QoS parameters

Supports Ciscos Identity-Based Networking Services (IBNS) which preventsunauthorized network access and insures that users get only their designatedprivileges

$5007.44CDWG 1000407

HP ProLiant

BL460c Blade ServerCall your CDWG account manager or complete blade coniguration

The Dual-Core Intel Xeon Processors 5110 (1.60GHz)allows a greater workload per processor

Memory intensive applications beneit rom the 1GB std.,32GB max. (PC2-5300) memory buers that allow asterspeeds and more capacity

With up to 2 SATA/SAS hard drive bays you get quicker accessto data and increased storage bandwidth

On-chip cache (4MB Level 2 Cache) increases dual-coreperormance and provides aster processor hit rate

$2299.99CDWG 1007007hP C Pc Three-year, our-hour, 24 x 7, onsite

$408.97 CDWG 1005222

bd p mp p v cmpg d w cgy.[[


16/36

Protecting the past and preparing or the uture o your organization starts with a reliable content management and datastorage solution. Your storage method should not only provide security or your current capacity, but must also be able toaccommodate changing technologies as well as your organizations uture growth. Luckily, increasingly ast responsetimes to store and retrieve your documents along with dramatic cost reductions make it even easier to implement

a storage system to accommodate your data storage needs and requirements.Many organizations have turned to storage area networks (SAN) as their preerred solution. A SAN device deliversbest-in-class perormance while providing both primary and secondary storage or database, e-mail, print, le, andWeb-service applications. By adding a SAN device to your network, you will be able to add capacity, increase eciencyand improve your le- and print-serving management.

SAN devices provide organizations with fexible, low-cost, high-capacity storage solutions or their highly sensitiveinormation. Built to grow with your organizations mounting storage needs, SAN devices are excellent investments.They provide smart organizations with the storage they need and the security they require.

Call your CDWG account manager today to nd out what an eective data-storage solution can do or your organization.

Overcome Storage Problems.

888.508.4239

CDWG.com/consolidation

HP StorageWorks 1500cs G2 Modular Smart Array Starter KitWith MSA30 SCSI drive enclosureThe HP StorageWorks 1500cs Modular Smart Array (MSA1500cs) Starter Kit utilizes a Fibre Channel storage area network (SAN) 2U controller shel that connectsto HP StorageWorks Serial ATA (SATA) and/or SCSI disk enclosures. The MSA1500cs with up to eight attached SATA enclosures can provide you with up to 48TB(96 500GB SATA disk drives) o native capacity. It provides you with a lexible, low-cost, high-capacity storage solution. Designed as a hardware oundation oruture solutions (to ensure maximum investment protection), the HP StorageWorks 1500cs Modular Smart Array Starter Kit includes (1) MSA1500, (1) MSA30 SCSIdrive enclosure, (1) 4/8 Base SAN Switch, (4) 4Gb SFP transceivers and (2) FCA2214 Host Bus Adapters with cables.

$14,999.40CDWG 1092139

Look to yourCDWG IT Investment Guide ormore inormation on these products.


17/36

Your CDWG account manager can help you fnd the right consolidation solution. 17

Your storage methodshould not onlyprovide security oryour current capacity,but must also be ableto accommodate

changing technologies.

HP StorageWorks Modular Smart Array 30 (MSA30) Greater perormance with Ultra320 SCSI transer rates

Up to 14 disk drives providing one o the highest spindle counts per rackspace resulting in a very high input/output per second

Single- and dual-bus, rack-mountable options insure access between the serverand the MSA30

Redundant power supplies and ans or higher availability

Advanced Data Guarding (RAID ADG) can withstand two simultaneous driveailures without downtime or data loss

MSA30 Single-bus Drive enclosure

$3115.18CDWG 529728

HP StorageWorks ModularSmart Array FamilyThe HP StorageWorks Modular Smart Array (MSA)Family provides you with HP ProLiant consistencyand simplicity, allowing you to utilize your current HPProLiant universal hard drives and easily migrate them

to an MSA product while protecting your valuable data.The MSA product amily is managed by the same toolsas the Smart Array controllers youre used to.

HP Ultra320 SCSI Universal Hard DrivesInternal, hot-pluggable72.8GB

15,000 rpm

$440.89 CDWG 442238 146GB

15,000 rpm

$655.35 CDWG 730117

CDWG 730117

CALL FOR PRICING

HP StorageWorks 4/8 Fibre Channel Fabric Switch Eight 4Gbps ports

Pay as you grow scalability rom single-switch to ull-abric connectivity

CDWG 876311

CALL FOR PRICING

HP StorageWorks 1500 Modular Smart Array

Starter Kit HA G2 Kit Includes one redundant controller, one FC I/O module, one 4/8 Base SAN

Switch, our 4Gb SFP transceivers, two FCA2214 HBAs and cables

CDWG 1092140

MSA30 Dual-bus Drive enclosure

$3376.53CDWG 529729


18/36

Five Strategies for Optimizing ServersHeres how you can get the most from your IT investment:

Use analysis tools. Many tools are bundled with servers and operating systems; robust third-party

applications are also available to make it easier to manage a server environment.

Put virtualization to work. By combining multiple servers into a virtual pool and allocating processing load

and tasks across a network, an enterprise can boost utilization rates from below 30 percent to 80 percentor higher.

Embrace a high-availability environment. Whether you equip a single server to handle failover needs or

map out a sophisticated disaster-recovery plan, ensure that servers are equipped to handle the demands of

enterprise systems, Web sites, e-commerce and other tasks. Even a brief failure can result in lost revenue

and unhappy staff, citizens and students.

Periodically reassess the IT environment. Over time, servers accumulate and conditions change. All too

often, this leads to an imbalance and underutilization of resources. Examine the overall situation at least

annually to ensure that youre allocating resources as needed.

Include the finance department in the analysis and decision-making process. Although server

optimization encompasses many technical issues, its essential to plug in business data and understandhow ROI (return on investment) and TCO (total cost of ownership) enter into the picture.

Telltale Consolidation SignsAccording to Steve Fink, director of IT Consolidation Solutions, HP, its time to consolidate servers when:

The average age of your server is three or more years

You are still utilizing direct attached storage

You have too many management tools

Your IT staff is minimal and seems always focused on maintenance

Youre not poised to consider open source applications and an open environment where you can move from

operating system to operating system

Theres a significant need to simplify and modernize your applications

You are completing a software refresh

Your Web servers have multiplied to the point where you have a management problem

Blade Servers Scaling Out and Scaling UpConsolidating servers often means purchasing new hardware. Going to new servers usually means scaling

out or scaling up (or some of each). Scaling out means more servers, typically with blade servers, such asan IBM eServer BladeCenter platform or HP BL ProLiant Series servers.

For example, rather than have 42 1U servers in a rack, you can have 84 servers in that same rack, says Bob

Lenard, director of Solutions and Channel Enablement, IBM X Series, IBM. The blade servers may be loosely

or tightly coupled, he notes, for example in a cluster so they can share workloads. And that rack can also

hold your I/O channels and switches, all consolidated into this environment, so your management becomes

better.

In terms of traditional U real estate, you may need up to 50 percent less space when you go to blades,

says Lenard. You may also need 25 percent less power and cooling, and you may have about one-third less

weight than a comparable 1U configuration which could be important, as older data center environments

may have a weight limit.The other direction to go in buying new hardware is scaling up to more powerful processing

environments to 4-, 8- and 16-way servers, such as IBM xSeries or eSeries, or HP ProLiant DL

servers. The raw performance improvements in processors let them handle more workload.

Hints and TipsConsolidation:

18 l 888.508.4239


19/36


Putting Virtualization to WorkThe raw power of todays storage technology is frequently overshadowed by the realities of making it work effectively.

Data and files end up scattered across a network, while data replication, performance monitoring and backups become

complex and unwieldy. The result for many organizations is costly system downtime and missed opportunities for

efficiency and growth.

Virtualization software addresses this challenge by better managing servers and storage solutions across the organization

and beyond. It enables more productive, flexible and scalable data centers while boosting data availability. Virtualization

achieves these gains by managing an operating system, application, process or storage solution across multiple servers or

systems thus maximizing processing power and reducing the need for additional hardware.

Whats more, because virtualization does not distinguish among physical media contained within servers, a system

administrator can add or swap devices at any time without reconfiguring the entire infrastructure. If one server or system

fails, another takes over instantly and performs the same tasks thus eliminating downtime and boosting availability.

The benefits of virtualization are well documented. It can create a utility computing environment that serves up data

on demand. It can handle data replication and performance monitoring, and manage backups, archiving and data

recovery far more deftly. Finally, it reduces costs by consolidating and eliminating servers, thereby cutting the real estate

needed for data centers, maintenance costs, administration time and cooling costs. As a result, a growing number

of businesses are finding that virtualization is a giant leap forward in the evolution of computing and information

technology.

An effective virtualization strategy offers a number of potential benefits:

Simplified administration. An effective server virtualization strategy allows an organization to make changes to the

computing environment quickly and seamlessly. IT staff can install operating systems and applications in a consistent

manner, and handle upgrades in a unified manner. In many instances, installations and upgrades take place in seconds

rather than hours or days.

Reduced server and power consumption costs. Virtualization typically complements a server consolidation strategy

and allows an organization to reduce the number of physical servers required to manage an IT environment. Not onlyis the organization able to reduce the direct cash outlay for servers, it is able to slash its energy bill by running fewer

servers and reducing the need to cool data centers.

Improved agility and flexibility. Moving away from the one server/one application model opens the door to a far more

robust computing environment and one that runs different operating systems and applications on the same server.

A virtualization strategy often boosts server utilization rates from below 20 percent to upwards of 80 percent.

Reduced licensing costs. Recently, a number of application providers have introduced virtual licensing agreements.

These licenses allow an organization to run virtual instances of an operating system on multiple servers at no extra

cost. For example, Microsoft allows customers to use up to four virtual instances of Microsoft Server 2003 R2 on a

single server.

Improve availability and stability. Virtualization boosts availability for key systems and applications. If a physical servercrashes, other virtual servers can take over and applications running on them remain unaffected. Some virtualization

software products automatically set up a new virtual server on a different machine if a hardware malfunction takes

place.

Less data-center space required. Not surprisingly, fewer servers translate directly into a reduced need for data-center

space. In todays compute-intensive environment, gaining control of server sprawl often leads to a substantial savings

in real estate and construction costs.

The Consolidation ProcessAs with any IT undertaking, there are a variety of factors to consider. Make sure youve thought about the IT economicsof the consolidation project, stresses Steve Fink, director, IT Consolidation Solutions at HP. Think of it as a technology

and business discussion.

To ease the decision-making process, CDWG can act as a trusted adviser throughout the consolidation process, says

Mike Beltrano, server product manager at CDWG. Beltrano describes CDWGs role as a four-step process:

1. Plan: Conduct assessments, take an inventory of current servers and applications, and recommend

a solution

2. Implement: Configure servers, ship and install new systems and any upgrades to existing systems

3. Manage: Provide any necessary integration and ongoing support services

4. Evolve: Help scale solutions as the business grows and fine-tune the IT environment as needs change

To help customers understand the benefits of server consolidation, and determine whether it makes sense, CDWG

offers an online Server Consolidation Profiler tool [at cdw.com/assessment], notes Beltrano. This easy-to-use tool asks

a series of simple questions and generates a customized report that details the benefits your organization may gain from

server consolidation.


20/36

Make the Switch to

a ProCurve NetworkProCurve Networking delivers enterprise networking solutions comprisingwired and wireless LAN and WAN networking products, services andsolutions. Whatever your needs, ProCurve has the solution thats exactlyright or you. With a variety o unmanaged 10/100Mbps, Web-managedand managed 10/100/1000Mbps switches along with chassis oering

integrated PoE on all 10/100/1000BASE-T ports, and 10GbE capability,ProCurve switches oer excellent investment protection, exibility and

scalability, as well as ease o deployment, operation and maintenance.

ProCurves Adaptive EDGE Architecture strategy is the industrys onlycomprehensive and inclusive network design strategy that is adaptable,scalable and completely interoperable or achieving command rom thecenter with control to the network edge.

ProCurve Networking Protection or Lie!

All ProCurve Networking switches come with a lietime warranty, makingProCurve an ideal long-term solution. A rock-solid lietime guarantee:

not just a lietime warranty, but also telephone support, Web support,ProCurve Manager (PCM) network-management sotware, downloadablesotware updates and next-business-day advance replacement. Fans andpower supplies are also covered or as long as you own your ProCurveNetworking product. ProCurve's ability to oer this level o warranty isa testament to the unparalleled reliability o ProCurve products and thelegendary HP quality rom which they arise.

ProCurve Networking by HP simply makes better sense i youre thinking

about the best ways to invest your budget. Now you can acquire theadvanced technology your company needs without worrying aboutdraining your resources over time on expensive maintenance agreements.ProCurve products set the standard or quality and reliability, and ourproducts and solutions are backed by the best warranty in the industry.

Call your CDWG account manager today to fnd out how we can put aProCurve networking solution to work or your organization.

[

888.508.4239

CDWG.com/consolidation

Microsoft Systems ManagementServer (SMS) 2003 R2Enterprise Edition Improve security o the Windows operating

system through vulnerability assessments o

common misconigurations Reliably target delivery o updates to Microsot

and third-party applications

Deliver critical applications reliably and easily tousers in the right place at the right time

Select License$419.991 CDWG 10439671Relects Select Level D pricing; U.S. Government Agencies willcollectively acquire at least 50,000 points rom each productpool elected during the term o the agreement; no media orlicense will be shipped; your ONLY proo o licensing is your

CDWG invoice; call your ederal CDWG account manager ordetails; CDWG holds GSA S chedule # GS-35F-0195J

APC Smart-UPS 1500VA USBand Serial RM 2U 120V Uninterrupted power protection with a

hot-swappable battery

1440VA, 980W, six outlets

Estimated runtime: up to 26 minutes athal load2

$663.33 CDWG 4684892Runtime may vary due to load and battery condition

For display only

Now you can acquire the advancedtechnology your company needs. [


21/36

Your CDWG account manager can help you fnd the right consolidation solution. 21

ProCurve Switch 3500yl series Ideal or ull-eatured Gigabit PoE Edge networks High-perormance, stackable orm actor

Planned uture IPv6 host and routing

Available in 24- or 48-port 10/100/1000BASE-T models, with our 10GbE ports

Lietime warranty on all parts or as long as you own the product

ProCurve Switch 3500yl-24G-PWR (J8692A) 24-port model $3006.46CDWG 953149ProCurve Switch 3500yl-48G-PWR (J8693A) 48-port model $5672.77CDWG 935403

CDW 953149

CDW 935401

ProCurve Switch 5400zl series Ideal or ull-eatured Gigabit PoE Edge network

Chassis orm actor

Planned uture IPv6 hosting and routing

Port density up to 288 10/100/1000BASE-T ports and 48 10GbE ports

Lietime warranty on all parts or as long as you own the product

ProCurve Switch 5406zl (J8699A) 48-port model $5546.09 CDWG 935401ProCurve Switch 5412zl-96G (J8700A) 96-port model $10,942.57CDWG 998992

To meet the evolving needs o security, mobility and converged applications, ProCurve Intelligent Edge switches the ProCurve Switch 3500 stackableand the ProCurve Switch 5400 chassis amily set a new standard. These next-generation switches oer wire-speed perormance and integratedGigabit Power over Ethernet (PoE), with scalability rom 24 to 288 ports per switch across the amily to provide your network with the lexibility oruture growth supporting other PoE systems such as VoIP.

Both are based on the ProVision ASIC, the ourth generation o custom-designed ASICs rom ProCurve that delivers intelligence and control to the edgeo a network. The ProVision ASIC provides all the perormance and unctionality needed today and also provides or uture application needs.


22/36

Server Virtualization:Lead the charge to increased productivity

no extra space requiredWhen your organization is looking to obtain higher utilization of hardwarewhile implementing a server consolidation initiative, virtualization is the keycomponent that will bring it all together. The virtualization will allow highersystems utilization when consolidating many under-used servers into a singlephysical server.

Using virtualization software, each application has its own virtual serverenvironment with its own operating system. The application assumes itsrunning on a dedicated server, yet all virtual environments share underlyingphysical server resources. By utilizing server virtualization your organization can:

Boost x86 server utilization to 6080%

Provision new systems faster with reduced hardware

Decouple application workloads from underlying physical hardware for

increased flexibility Dramatically lower the cost of continuity

What is the virtualization ROI?

Want to know the exact return on investment (ROI) you can expect fromadopting CDWGs recommended consolidation strategy and virtualizationsoftware? We know your upper management will. During the planning process,your CDWG account manager and pre-sales systems engineer will create acustomized ROI Analysis, using our Server Consolidation Calculator.

In a comprehensive, three-page report, the Analysis looks at key financialmetrics such as:

Payback period

Net present value (NPV)

Discounted return on investment (dROI)

Modified internal rate of return (mIRR)

The Analysis also includes three-year cash flows and detailed cost-benefitinformation such as server-hardware cost savings, server management anddeployment productivity, space cost-savings and more.

After the ROI discovery phase, CDWG will work with you to identify thevirtualization solution thats right for your server/storage environment. Callyour CDWG account manager today to get started on the server virtualizationsolution thats right for your organization and watch your productivityincrease.

888.508.4239

CDWG.cm/csdt

VMware TechnologyNetwork SubscriptionVMware Technology Network (VMTN)Subscription is a service for developers and testersthat provides a suite of VMware products, support

and upgrades in a convenient, low-cost annualsubscription. VMTN Subscription gives developersand testers the benefits of VMware virtualizationcapabilities for software development in a simple,flexible and affordable pricing model.

Use virtualization in development for less. VMware'spowerful suite of virtualization tools has beenvalidated by millions of existing users, and it isnow available in a package designed and pricedspecifically for development (non-production use).

$283.99 CDWG 1002062

VMware Workstation 5.5VMware Workstation 5.5 is powerful desktopvirtualization software for software developers/testers and enterprise IT professionals that allowsusers to run multiple operating systems andtheir applications simultaneously on a single PCin fully networked, portable virtual machines.VMware Workstation 5.5 accelerates the softwaredevelopment/testing process, increases userproductivity and enables hardware consolidation.

$179.99 CDWG 773141


23/36

Yu CDWG ccut mge c he yu fd the ght csdt sut. 23

Call or priCinG

VMware

Infrastructure 3Dt Cete Mgemet d otmzt SuteVMware Infrastructure is the most widely deployed software suite for optimizing and managing industry standard IT environments through virtualization from the desktop to the data center. The only production-ready virtualization software suite, VMware Infrastructure is proven to deliver results for over 20,000customers of all sizes and is used in a variety of environments and applications. The suite is fully optimized, rigorously tested and certified for the widestrange of hardware, operating systems and software applications. VMware Infrastructure provides built-in management, resource optimization, applicationavailability and operational-automation capabilities, delivering transformative cost savings and increased operational efficiency, flexibility and service levels.

Increase hardware utilization and reduce hardware requirements with server-consolidationratios commonly exceeding ten virtual machines per physical processor.

Reduce required data-center square footage, rack space, power, cooling, cabling, storageand network components by reducing the sheer number of physical machines.

Decrease labor costs by simplifying and automating labor and resource-intensive IT operations.

Improve application availability and continuity independent of hardware and operating systems.

Enable continuous uptime and non-disruptive maintenance of IT environments with livemigration of entire running systems.

Eliminate cumbersome software installations and configurations with virtual appliances.

Accelerate application development and deployment lifecycles.

Improve responsiveness to business needs with instant provisioning and dynamic optimizationof application environments.

Allow legacy systems to co-exist with new environments.

The sute cudes:

VMwe ESX Seve Platform for virtualizing servers, storage and networking

VMwe VMS High-performance cluster file system for storage virtualization

VMwe Vtu SMp Multiprocessor support for virtual machines

VMwe VtuCete Centralized management, automation and optimization for IT infrastructure

VMwe Ha Cost-effective high availability for virtual machines

VMwe DrS Dynamic balancing and allocation of resources for virtual machines

VMwe VMt Live migration of virtual machines without service interruption

VMwe Csdted Bcku Centralized backup software for virtual machines

CDWG 994692

Look to yourCDWG IT Investment Guide formore information on these products.


24/36


CDW, CDWG, CDW@work, CDWG@work and The Right Technology. Right Away. are trademarks of CDW Corporation.

At CDWG, we understand that downtime can bring you down. It threatens

productivity and progress. Thats why we offer a selection of extended

notebook warranties in a variety of service levels to fit your needs. Whether you

need four-hour turnaround or next-day, onsite support, weve got you covered.

Ask for an extended warranty with your notebook purchase to reduce downtime.

Stay up and running. Call 800.808.4239

or visit CDWG.com

Downtime got you down?

Try an extended warranty.


25/36

As stated in Chapter 1, consolidation is the process obringing together separate parts into one unied whole. Thisis accomplished through the systematic elimination o excess

resources to reduce inrastructure to its most ecient state.There are our specic types o consolidation that IT managersmay consider, ranging rom the easily implemented consolidationapproach o unied management to complete operationalconsolidation. This chapter will dene each o the our typeso consolidation, illustrate a ew ways in which each can beimplemented and dene a consolidation strategy to help youachieve consolidation within your organization.

Four Types o ConsolidationAs covered in the rst chapter, the main goal o consolidation is todecrease the TCO o the IT inrastructure while increasing the ROI.This can be accomplished through our types o consolidation:

Logical consolidation

Centralized consolidation

Physical consolidation

Operational (or application) consolidation

1. Logical ConsolidationLogical consolidation is the centralization, or uniying, omanagement o IT resources. Although logical consolidationdoesnt actually relocate or physically consolidate any servers, itdoes simpliy management, thereby easing the IT workload. Thisrst stage o consolidation is by ar one o the most cost-ecientand easiest to implement when properly planned.

Logical consolidation can be addressed rom two points o practice.First is that o standardization. A lack o standardization within anIT inrastructure leads to a condition known as server or system

sprawl. Sprawl is the accumulation o nonstandard, incompatibleor dicult-to-support inrastructure components. Standardizationis the cure.

Standardization

Logical consolidation is the rst step toward standardization.In the realm o IT, it must be understood that standardization isyour riend, variation is your enemy and complexity should neverbe eared. As with any standardization, however, there are oten

tradeos to be addressed, most notably in the area o fexibility.

Standardization is oten seen, quite inappropriately, as the tie thatbinds an organizations fexibility. However, the identication ospecic standard product and service oerings will ease the overallmanagement o the IT inrastructure by simpliying the serviceoerings. This will have a direct, positive impact on the governmentoce or educational institution in terms o cost and time to market.

For example, your line-o-business partners may be large

consumers o database resources. Thereore, conguringstandard product and service oerings that meet the needs o theorganization in terms o size, reliability, scalability, condentiality,integrity, availability, data retention and compliance will allow yourIT organization to negotiate and procure lower-cost products andservices. However, it may also decrease time-to-market by havingprecongured standard service oerings available to your lineo end users. Beore taking a deep-dive into the realm o uniedmanagement, consider logical consolidation rom a standardization

point o view. In doing so, your organization will begin to take

ConsolidationStrategiesIN THIS CHAPTER:

Four Types o Consolidation Dening a Consolidation Strategy

Implementing a Consolidation Strategy

Chapter 4



26/36

accountability or its assets and realize the larger value o IT as astrategic partner to drive down costs and decrease time to market.

Unied Management

Achieving logical consolidation addresses the need to simpliy

management and thus reduce overhead spent in supporting adisparate and dicult-to-manage environment. Standardizationis one approach, but it will not be all-inclusive. This is why logicalconsolidation not only includes standardization but also uniedmanagement. While standardizing on a single production platormor architecture is important, uniying management o IT resources isequally as important in cost reduction o resources.

For example, i your organization already uses Microsot

Windows Active Directory as the standard domain model, uniedmanagement may include the collapsing o unneeded AD domainsinto a single domain, or collapsing redundant or superfuousorganizational units (OUs) into a more organized and easy-to-manage structure.

Unied management goes much urther than simply domainmodels and systems architectures. It is oten more about tools andresource management that can be measured more in man hoursthan system perormance. Unied management may also include

the deployment o standard tools to address common needs.Antivirus tools such as Symantec Antivirus Enterprise Edition, orexample, can be used to manage the antivirus needs o all systemsrom a central location.

Enterprise patch management through Symantec LiveState PatchManager centralizes patch management and enables IT managersto keep devices secure, available and compliant with corporatestandards and can install patches on hundreds o clients and

servers in minutes, including mobile and remote devices.

Further, application deployment tools such as IBM DirectorSotware Distribution Premium enable organizations to easilycreate and distribute sotware packages or Windows and Linuxenvironments, saving travel and labor costs to install new orupdated production applications. Once youve reached an optimal

state o logical consolidation, the next stage in consolidationocuses on the physical consolidation and the benets ocolocation.

2. Centralized ConsolidationCentralized consolidation, sometimes called colocation, is thephysical consolidation o servers to a central location usuallya data center. This will signicantly simpliy maintenance costs

by allowing or the colocation o IT resources to support acolocated inrastructure and server environment. For example, anorganization that has 20 locations spread across three states, eachwith their own servers onsite, that consolidates servers into onecentral space will reduce costs by no longer needing to maintainIT support sta at those locations or travel to those locations orserver support.

Also, the management o each server will be simplied. As any ITadministrator will attest, there are times when you just need tolay hands on a box to solve a problem. By colocating servers toa central location, the cost o laying hands signicantly decreases.This also makes issues such as implementing security measures,backing up data and instituting an extensive ailover plan mucheasier because IT managers will not need to account or datalocated across a wide range o decentralized servers.

3. Physical Consolidation

Many organizations stop with colocation and never go much

4CHAPTER

26 l 888.508.4239


27/36

urther in terms o consolidation, which does a great disserviceto all the hard work that has gone into intelligent server design.Physical consolidation is the compiling o workload rom multipleservers onto a single server platorm.

As addressed in the previous chapter, it is not uncommon or

servers to be purchased or a specic purpose and not utilized orany other reason. These kinds o servers are not being utilized totheir ullest potential. The challenge that exists in this area is onethat is oten more managerial than technical that is, relatingcharges or server resources back to the line o business or the useo a particular server resource.

Although this task can be a challenging institutional question, itshould not be allowed to prohibit a valid opportunity to reduce theTCO through consolidation. Identiying underutilized servers and

consolidating them into one eciently utilized server will reducecosts by simpliying management and reducing power consumptionneeds, foor space, cooling costs and network connectivity charges,and may even save money in terms o licensing.

4. Operational (or Application) ConsolidationOperational consolidation is the consolidation o multiple platormoperating systems (OSs) and application systems onto a single

server. Although the most complex o all consolidation approaches,operational consolidation oten results in the highest reduction o

TCO by making the most ecient use o system resources within aserver. The challenge exists in ensuring that disparate applicationsystems and platorms dont interere with each other or over-complicate the environment rom either a system or operationalpoint o view.

For example, i through operational consolidation two applicationsare considered or deployment on the same server, steps must betaken not only to ensure that the platorm remains stable but alsothat the operational outage windows or each application allowenough time or server maintenance to be perormed. Installingan application used between 8 a.m. and 5 p.m. along with anapplication that is used rom 5 p.m. to 8 a.m. would leave a zerowindow o opportunity or routine systems maintenance such asthe installation o hotxes, service packs and upgrades to the

underlying OS.

Dening a Consolidation StrategyA consolidation strategy requires both a high level o technicalacumen to allow or an understanding o your organizationscurrent technical inrastructure, architecture and requirements, anda high level o business acumen to be sure that the consolidationstrategy being dened is in alignment with operational goals.

You wouldnt, or example, want to spend a great deal o timeconsolidating an application inrastructure on an environment

scheduled or replacement or retirement in the next six months.Budget issues, organizational goals and expectations resulting roma consolidation eort must all be discussed in detail in advance.

A successul consolidation strategy depends upon a detailedcapacity and perormance analysis, application analysis andoperations analysis to ensure that the strategy can gain the mostpositive impact on the IT inrastructure without adversely aectingapplication management, support or operations. Once youre sureyou have a solid understanding o the current goals and directiono the technology and the organization, the rst step will be toconduct a system-level analysis o the inrastructure.

Capacity and Perormance AnalysisAs discussed in Chapter 3, many actors contribute to systemmetrics and need to be taken into consideration to ensure a airand accurate measurement that can be used or managementpurposes. CPU utilization, memory utilization, disk utilization and

network bandwidth all come into play.

4CHAPTER


Operational consolidation is enabled through a technology

known as virtualization, which will be covered in depth

in Chapter 5.

One o the most direct ways to physically consolidate a

server environment is to invest in blade servers. A blade

server diers rom a standard server in that multiple

blade servers can share the same hardware chassis with

common services being provided, such as power. Eachblade remains independent rom other blade servers

insoar as each has its own network connectivity and

unctions independently o the others.

Blade servers such as the IBM eServer BladeCenter HS20,

which eatures a 3.20GHz Intel Xeon processor and 1GB o

RAM (but is only 2.5 inches thick), oer impressive power

and perormance within a compact ootprint. Similar blade

systems are available rom most major manuacturers.

The greatest benet o blade server computing is achieved

through savings realized in reduced power consumption

and foor space. However, savings are also derived through

decreased cabling inrastructure and the use o a single

monitor, mouse and keyboard per chassis.


28/36

These metrics all under the IT discipline o C&PM. A ull anddetailed C&PM analysis is important to understanding what thepresent state o the inrastructure is in terms o optimization andeciency.

As explored in Chapter 2, the purpose o C&PM analysis in this

regard is to identiy opportunities to increase eciency. Theseresults will play a big part in ensuring that the benets to eciencyrealized through optimization eorts are translated to the sameeciency benets during consolidation.

Application AnalysisOnce a C&PM analysis is perormed and the varying needs oeach candidate or consolidation are understood, the next matterto address is the application. Some applications may confict or

present challenges at the OS level. For example, when consideringconsolidating two Web server applications to a single server, youmight run into conguration challenges given the nature o theapplications as they compete or resources. Although it is certainlypossible to run two Web servers on the same server, it may bemore ecient to run applications that dont directly compete orsimilar resources such as a Web server and a database server.

Operational AnalysisOperational analysis is the process o level-setting the C&PMand application analyses against the needs o operations.Consolidation, in any regard, will certainly yield benets to the lineo business, but proper timing in relation to operations needs is alsoa very important actor to consider.

Conducting an operational analysis will aide in not onlyunderstanding how consolidation will benet the operations butwhen it will best benet the business or maximum impact. Someactors to consider during an operational analysis might be growthand new initiatives, hardware and sotware reresh times (as theyoten dictate the next major upgrade schedule), and desire o theoperations to derive savings rom TCO reductions.

Implementing a ConsolidationStrategyAs your organization prepares to move rom planning toimplementation, it is important to ensure that your plan hascovered all the major obstacles that may impede success. Theollowing checklist will help your organization properly prepare orthe task ahead.

Completing each o these steps will ensure that youre on the rightpath prior to implementation and that your approach is valid.

4CHAPTER

28 l 888.508.4239


Right Away.

CDWGcertied technicianscan congure rack

hardware to yourspecications beoreit ships.

Server Consolidation Checklist

Gain understanding o operational directions or IT and needs driving

consolidation.

Gather server location and inventory data.

Complete initial analysis o server environment.

Identiy consolidation candidates; keep them as simple as possible.

Develop executive-level assessment report with potential return on

investment (ROI).

Analyze server and network perormance.

Determine server unctions and network topology policies, such as

those set by the Storage Networking Industry Association.

Document interdependencies; local area network/wide area network

(LAN/WAN) constraints; security.

Develop and document a consolidation plan.

Document costs and expected TCO (total cost o ownership) and ROI.

Develop application and data proles.

Review architecture and perormance needs with application users.

Analyze application unctions, workload and transaction patterns.

Determine plans or growth o unctions and data.

Develop recommended consolidation solution and a plan to get there,

including best practices.


29/36

When an organization is dedicated to achieving the maximumpossible ROI and is in ervent pursuit o the lowest possible TCO oits IT inrastructure, virtualization technology must be considered.Virtualization is the process o presenting a logical grouping or

subset o computing resources so they can be accessed in ways

that give benets over the original conguration. It is also atechnology oten used to consolidate many physical servers intoewer servers hosting virtual machines.

A virtual machine is a partition o a virtual server that consists oCPU, memory, network and disk resources supporting a single OSinstalled in a virtual environment. Virtualization is usually made upo a host server running a base OS, such as Microsot Windowsor Linux, along with virtualization sotware such as VMwareor Microsot Virtual Server, which in turn, partitions out virtualmachines or use by other OSs. The host OS controls the overallserver and the virtualization sotware divides those resourcesamong various virtual machines (also called virtual servers) all

o which is completely transparent to the OS o the host serverplatorm and to the users.

The end result is that a single hardware platorm server runninga single core OS with virtualization sotware is capable oacting as multiple servers running multiple OSs, all o which arecompartmentalized rom each other and the base OS.

Beore virtualization, servers were able to support only a singleOS, which limited opportunities or consolidation. Running

multiple applications on a single OS oten results in unexpected oradverse consequences as two or more applications that were notspecically designed to run with each other compete or systemresources.

Now that virtualization has been introduced, hardware has becomecompletely independent o OSs and each application is ree to runwithin its own virtually dedicated environment. Hardware resourcesare assigned at the virtualization sotware level to each virtualmachine so that, rom an OS/application point o view, resources

do not confict.

VirtualizationIN THIS CHAPTER: Three Virtualization Models

Calculating the Value o Virtualization

Virtualization Technologies

Making It All Work

Chapter 5


x86 Architecture

Operating System

Application

CPU

Memory

Network

Disk

x86 Architecture

Virtualization Software

Operating System

CPU

Memory

Network

Disk

CPU

Memory

Network

Disk

Operating System

CPU

Memory

Network

Disk

Virtualization sotware enables multiple virtualized servers to rundierent operating systems on a single physical server.


30/36

This setup enables applications that could not have previouslycoexisted, such as an Apache Web server running on Red Hat Linuxand a Microsot Exchange E-mail Server on Windows 2003 Server,to run rom the same hardware platorm.

Three Virtualization ModelsVirtualization falls into three basic virtualization models:

Element virtualization This enables the optimal use oindividual servers, storage and other resources to meet demandwithin a single application environment or business process.Server virtualization is one example o element virtualizationwithin a single element silo.

Integrated virtualization This type o virtualization enablesthe optimal use o multiple inrastructure components to meetthe highest possible yield rom an environment. Integratedvirtualization may include a hardware platorm specically suited

to virtualization, such as Intel VT (Virtualization Technology),which is a set o hardware enhancements to Intel server andclient platorms that can improve traditional sotware-basedvirtualization solutions, along with an OS specically suited tothrive in a virtual environment.

Utility computing Also known as on-demand computing,utility computing is an architectural model or computing inwhich computing resources are made available to the user on

an as-needed basis. The goal is to maximize the ecient useo computing resources and minimize costs. Utility computingis enabled by storage virtualization, server virtualization andautomated inrastructure/service provisioning.

Each virtualization model has its own value within an organizationand calculating that value can be quite complex.

Calculating the Value oVirtualizationBy now, it should be clear that optimization, consolidation andvirtualization yield nancial benets and calculating those

benets may become crucial in driving adoption o virtualizationwithin your organization. However, calculating the return can geta bit complex.

Determining the payback period o an investment in virtualizationtechnology involves calculating the ollowing: the net presentvalue (NPV), which is the present value o a uture implementationor, in laymans terms, what it would be worth i implementedtoday; the discounted ROI (dROI), which is the present value oan asset divided by its cost to implement; and modied internalrate o return (mIRR). Unlike the internal rate o return (IRR), whichassumes the cash fows rom a project are reinvested at the IRR,the mIRR assumes that all cash fows are reinvested at the rmscost o capital. Thereore, mIRR more accurately refects theprotability o a project.

While these calculations may seem a daunting challenge, theywill all directly contribute to success in translating the value ovirtualization back to specic nancial objectives.

Virtualization TechnologiesNot all virtualization technology is entirely server-based. Thissection will cover the major contributors to server virtualization aswell as touch on other ways virtualization technology may benetyour organization, specically in the area o network le and SAN(storage area network) virtualization technologies.

Server-Based VirtualizationThere are two major developers o virtualization technology todaythat emulate hardware in the virtual server environment: Microsotand VMware. Both developers produce a server-class product thatis capable o running multiple virtual machines on a single server.Both are easy to manage, rom a traditional IT administrationpoint o view, and eature a plethora o add-on utilities to makemanagement practical.

For example, the VMware product eatures a Physical to Virtual(P2V) Assistant that simplies the migration o physical servers tothe VMware platorm. Microsot Virtual Server 2005 oers a similartool and includes a Web interace that provides administrators witha Web-based view o the status o their virtual computers. Last, butnot least, is the capability o both virtualization platorms to runmultiple OSs, including, on VMware, the IBM OS/2 Warp OS, which

5CHAPTER

30 l 888.508.4239

Another excellent use o virtualization technology isa rapidly deployed development environment. I your

organization requently requires server resources or

application development or application testing, installing

a virtual server or use by application teams can oer

a cost-eective and easily recyclable development-on-

demand environment. Dev-On-Demand is a phrase that

has been used to reer to any readily accessible virtual

environment or the purpose o development and/ortesting and is quickly becoming required equipment or

any lab environment.


31/36

is still in use today in some environments.

Another virtualization technology on the horizon will beincorporated into the newest version o Red Hat Enterprise Linux.RHEL version 5 will include Xen Open Source Virtualizationsotware.

Nondestructive MobilityA eature o virtualization that is oten not ully realized out othe starting gate, but yields great benets in terms o time andsimplication o IT management, is the ability or virtual machinesto be completely contained within a single le on disk. This enablesa single virtual server to be copied as a single le to anotherlocation within the inrastructure (another virtual server platorm)and unction as seamlessly as the original did. This unctionality

enables complete virtual server relocation (mobility) withoutdisruption o the original virtual le system in the process.

Network File VirtualizationNetwork le virtualization (NFV), sometimes called NAS (networkattached storage) virtualization, is an emerging virtualizationtechnology that enables users to continue to read/write data as thedata is being dynamically reallocated within networked storage.

This unctionality enables storage administrators to reallocatestorage without concern or data access.

Virtual Storage Area NetworkA virtual SAN (VSAN) is a specic section o a SAN that has beenbroken up into logical sections, also reerred to as partitions, toallow trac to be isolated within specic portions o a network. Ia problem occurs in one VSAN, that problem can be handled witha minimum o disruption to the rest o the SAN. VSAN technologyincreases storage fexibility and contributes to the optimizationo storage resources by making the best use o a storageinrastructure virtually autonomous.

Making It All WorkAs organizations continue to strive or increased serverperormance against strict IT budget constraints, IT managers needto be rugal about how they spend their IT budget. Optimizingservers and inrastructure can yield quick and easy ROI throughsimply understanding where the most potential or perormanceexists within the organization.

Consolidating, rst through unied management and later throughcolocation and virtualization, can enable organizations to morequickly respond to changing needs and to simpliy IT management.

Remember, optimization, consolidation and virtualization aremore than just industry buzzwords these days theyre survivaltechniques that can be planned, executed and ully realized.

As you proceed with your consolidation initiatives, CDWG can beo assistance to you not only as a dependable and convenient

source o products or your consolidation needs, but also as anadviser. We provide you with a single point o access to thousandso technology vendors.

We oer consolidation resources ar beyond hardware andsotware. In addition to our own team o specialists, we haveull-time access to dedicated vendor-assigned systems engineers,as well as service providers who oer a ull range o IT services. Weknow todays operational demands are constantly changing, so wecan help keep you aware o new capabilities and technologies.

Through planning, designing, implementing, managing andsupporting, well work with you to dene key objectives, determineyour technology goals and understand the purpose o yourconsolidation project. And we continue supporting you ater theproject has been completed with onsite assistance (whenrequested), 24 x 7 telephone tech support and your ongoingrelationship with your CDWG Account Manager. Because o ourocus on providing the right technology, right away, you can enjoy

the condence o working with the premier experts or serverconsolidation and other technology solutions.

At CDWG, weve designed our company with you in mind. Ourgoal is to ensure your satisaction over the long run. 5

CHAPTER



Right Away.

From MicrosotVirtual Server toVMware, CDWG

has your virtualizationneeds covered.


32/36

Application PortoliosA system o aligning and tracking multiple individual applications

that serve a larger purpose

Capacity and Perormance Management (C&PM)IT management processes ocused on ensuring that IT and related

inrastructure continue to meet current and uture organizational

requirements in the most ecient manner possible through

management o inrastructure capacity and perormance

Centralized ConsolidationThe physical consolidation o servers to a central location;

also called colocation

ConsolidationThe process o bringing together separate parts into one

unied whole

Data ClassicationThe process by which organizational inormation requirements are

gathered, then classied, rst as inormation and subsequently

as data

Data ReplicationThe copying o data to and rom sites to improve local serviceresponse times and availability

Dev-on-DemandShort or Development on Demand, a use o virtualization

technology to set up a rapidly deployable solution or application

development and application testing needs

Element VirtualizationThe process o presenting a logical grouping or subset o

computing resources so that they can be accessed in ways that

give benets over the original conguration within a single

element, such as a server or storage device

Hierarchical Storage Management (HSM)A data storage methodology that automatically moves data

between high-cost and low-cost storage media as needed toaid in storage optimization and cost management

ImagingA process wher

Server and Storage Consolidation

Documents

Transcript of Server and Storage Consolidation