capacity planning discipline for data centers

8/9/2019 Capacity Planning Discipline for Data Centers

1/34

Capacity PlanningDISC IPL INE FOR DATA CENTER DEC IS IONSTable of Contents

1 Introduction ...........................................................................1-12 The Value of Capacity Planning................................................2-1

Introduction ............................................................................. 2-1The Value Proposition for Capacity Planning..............................2-2

3 Capacity Planning Processes ...................................................3-14 How to Do Capacity Planning...................................................4-1

Three Steps for Capacity Planning .............................................4-1Determine Service Level Requirements......................................4-1Analyze Current Capacity ..........................................................4-5Plan for the Future ....................................................................4-9Capacity Planning Process ......................................................4-13

5 Using Capacity Planning for Server Consolidation ....................5-1Server Consolidation Defined....................................................5-1Finding the Answers Can Be Difficult .........................................5-2Predict the Future .....................................................................5-2Case Studies ............................................................................ 5-2Company A Finding Underutilized Resources for Useas Disaster Backup ...................................................................5-2Company B Avoiding Unnecessary Expenditures.....................5-4Conclusion ...............................................................................5-4

6 Capacity Planning and Service Level Management ...................6-1Service Level Agreements (SLAs) ............................................... 6-1Proactive Management of Resources to MaintainService Levels........................................................................... 6-2Service Level Management Scenarios........................................6-2

7 Product Focus: TeamQuest Model ............................................7-1Case Studies ............................................................................ 7-3TQ-EB01 Rev. A

Copyright 2004 TeamQuestCorporationAll Rights Reserved

Like what you see?Subscribe.

EMA Perspective....................................................................... 7-48 Bibliography...........................................................................8-1

See the final page for important legal notices.
http://www.teamquest.com/subscribe/http://www.teamquest.com/subscribe/


2/34

Introduction

1 IntroductionBoth server capacity planning and network capacity planning are important disciplines for managingthe efficiency of any data center. The focus of this eBook is on server capacity planning.

Traditionally, server capacity planning is defined as the process by which an IT departmentdetermines the amount of server hardware resources required to provide the desired levels of servicefor a given workload mix for the least cost. The capacity planning discipline grew up in the mainframeenvironment where resources were costly and it took a considerable amount of time to upgrade. Asthe data center transitioned to a distributed environment using UNIX, Linux, and Windows servers,over-provisioning and the introduction of cheap new boxes served as a replacement for capacityplanning. However, as the distributed data center matures we have found these practices to be lessthan optimal.

The average utilization on many servers in the distributed data center is way below acceptablemaximums, wasting costly resources. Further, as the economy evolves, IT is expected to do its shareto become more efficient, both in terms of hardware and human resources. And finally, IT is nowbecoming an accepted partner in the organizations drive to grow revenue while increasing profitsand remaining competitive for the long term. All of these factors require more mature IT processes,

including the use of capacity planning, a core discipline every company should adopt.

There are a number of types of capacity planning including the following:

1. Capacity benchmarking2. Capacity trending3. Capacity modeling

Benchmarking, or load testing, is perhaps the most common, but also the most expensive. The ideais, you set up a configuration and then throw traffic at it to see how it performs. To do this right, youneed access to a fully-configured version of the target system, which oftentimes makesbenchmarking or load testing impractical, to say the least.

Linear trend analysis and statistical approaches to trending can provide quick and dirty ways topredict when you will need to do something about performance, but they dont tell you what youshould do to optimally respond. Trending does not provide a way to evaluate alternative solutions toan impending problem, nor to understand problems unrelated to trending, e.g. server consolidationor adding an application.

That leaves modeling, which comes in a couple flavors: simulation and analytic modeling. Simulationmodeling can be very versatile and accurate, but requires a great deal of set up effort and time.Analytic modeling is fast and is potentially very accurate as well. The beauty of modeling is that youcan test various proposed solutions to a problem without actually implementing them. This cansave a lot of time and money.

This eBook explores the modeling approach to capacity planning. However, regardless of the toolsetyou use, if any, the key is to adopt mature IT processes, including the strategic use of capacityplanning.

This eBook is organized into a series of chapters, with Chapter 2 following this introduction:

2. The Value of Capacity Planning

The second chapter is excerpted from a paper1 written by Enterprise Management Associates,an analyst organization focused on the management software and services market. Itexplains the benefits of capacity planning.

Like what you see?Subcribe.

Capacity Planning: Discipline for Data Center Decisions Page 1-1


3/34

Introduction

3. Capacity Planning Processes

In this chapter we discuss capacity planning in context with other generally accepted ITprocesses, specifically using the ITIL model.

4. How to Do Capacity Planning

This section illustrates a methodology for capacity planning with examples using the

TeamQuest product line.

5. Using Capacity Planning for Server Consolidation

Server consolidation is a popular capacity planning technique for making IT operations moreefficient. This chapter explains server consolidation and provides two real-world exampleswhere capacity planning tools were used in consolidation projects.

6. Capacity Planning and Service Level Management

This chapter discusses key aspects of service level management and the relationshipbetween service level management and capacity planning.

7. Product Focus: TeamQuest Model

We conclude with another excerpt from EMAs paper1 that highlights TeamQuests capacityplanning solution, TeamQuest Model.

8. Bibliography

This eBook includes material from several resources listed here.

Whether you are learning about capacity planning for the first time, or you are an old hand looking forsome new points of view on the subject, we hope this eBook will prove useful to you.




4/34

The Value of Capacity Planning

2 The Value of Capacity Planning

Introduction

The boom-and-bust economy of the past five years has had an

unprecedented effect on IT organizations. In the late 1990s and2000, IT infrastructures grew at a frantic rate, creating a level ofcapacity and staffing that had never been seen before. Yet, with the economic tailspin of thesubsequent three years, most corporations are now seeking new ways to leverage IT to improvebusiness efficiency and automationwhile severely restricting ITs budgets for personnel, training,bandwidth, and equipment.

This roller coaster of supply and demand has left many IT managers looking for new, better ways tomanage their resources. While previous IT strategies focused primarily on managing growth, todaysIT organizations now are seeking methods to expand the number of applications and servicesavailable to support the business while cutting costs at the same time. In the current economy IT isincreasingly being asked to do more with less.

From a server perspective, the rise and fall of IT has instigated

several shifts in resource management. First, most IT organizationsare looking to make better use of existing assets by consolidatingunderutilized server resources and re-purposing capacity that waspurchased during the boom years, but has now been freed up bydown-sizing and/or reorganization of corporate business units. In arecent EMA survey, reclaiming and/or re-purposing hardware andsoftware that is underutilized was cited as a top priority by 57% ofIT executives responding.

Second, IT organizations are re-thinking their previous brute force approach to provisioning servercapacity. During the boom years, when growth was the top priority, many IT organizations respondedto performance problems simply by purchasing additional servers to handle the load. Today, whenevery IT dollar is critical, this over-provisioning approach is viewed as wasteful and inefficient,

particularly since downsizing, not growth, is the order of the day. In difficult economic times, ITorganizations must be able to measure server utilization in sharp detail, and make the best possibleuse of every processor cycle.

Lastly, many IT organizations have grown weary of traditional server-by-server resource managementefforts and are now evaluating next-generation technologies that would enable servers to pool anddynamically allocate their capacity to fit the specific needs of applications and services. This dynamicallocation of capacity, sometimes called virtualization, enables applications to draw capacity asneeded from a collective infrastructure, rather than relying on dedicated servers. This virtualizationtechnology could eventually pave the way for a new paradigm known as utility computing, in whichusers and/or applications draw only the capacity they need from a common infrastructure, just ashomes or businesses draw only the water or electricity they need each month from their local utility.

How can enterprises re-purpose underutilized capacity, implement structured methods for

provisioning servers, and prepare their IT environments for the coming wave of virtualization? Theanswer to all three questions is the same: capacity planning. By using capacity planning tools andprocesses, enterprises can measure server utilization trends, analyze future capacity needs, andpredict the exact server requirement for a given application, service, or general time period. Somecapacity planning toolsets also offer the ability to model the server environment, testing a variety ofwhat if scenarios against proposed server configurations to determine the most efficient use ofcapacity and server resources. In the following section, we examine the capabilities of capacityplanning and modeling tools, and evaluate the value that they can deliver to the enterprise.

This chapter is excerpted froma white paper1 by Enterprise

Management Associates (EMA),an analyst organization

specializing in themanagement software and

services market.

In a recent EMA survey,reclaiming and/or re-purposing hardware andsoftware that is underutilizedwas cited as a top priority by57% of IT executivesresponding.




5/34


The Value Proposition for Capacity Planning

In a difficult economy that has imposed severe restrictions on IT budgets, it may seemcounterintuitive to invest in new technology as a means of cost control. Yet, in the case of capacityplanning technology, a relatively small investment in software can lead to a very large savings inserver hardware and associated costs. While there may be some costs associated with deployingcapacity planning tools and processes, the value of that investment is typically extremely high.

How is that value achieved? One way is through the reductionor,in some cases, complete elimination of server over-provisioning.Capacity planning tools enable IT organizations to benchmark theexact workload that a server handles over time, then project trendsin traffic that the server might need to handle in the future. Thisdata, in turn, enables the IT organization to determine the exactserver requirements for a particular situation, ensuring that theserver environment will have enough capacity at all times. Thisprocess contrasts sharply with the estimated capacity allocationmethod in which the IT organization simply guesses at the capacityrequirement, then overbuys servers to ensure that there will alwaysbe sufficient processor power. In many cases, capacity planning

may enable the enterprise to slash its server acquisition budgetsignificantly.

The flip side of this server utilization data is that it helpsenterprises to identify underutilized servers and capacity so thatthey can be consolidated or re-purposed for other applications andservices. Capacity planning tools help organizations to identifyserver resources that might be idle, enabling IT departments toconsolidate several applications or services on a single server, orreallocate server processor power to other services that may need the capacity. This is an essentialcapability in todays IT organization, where capacity purchased during the boom years may besitting idly while budgets for new server capacity have often been slashed to the bone.

Capacity planning can also help preserve another crucial enterprise resource: IT staffing. Over thecourse of a servers life, most enterprises spend more money on maintenance, configuration, andupgrade of servers than they spent to purchase the server in the first place. For every underutilizedserver, there is a group of IT staff who have been tasked with installing, configuring, and maintainingthat server. If server resources can be used more efficientlyeither by eliminating servers or byenabling a single server to handle multiple applicationsthen the cost of IT staffing dropsproportionately.

A related benefit of capacity planning is the reduction of server downtime. Under the estimatedmethod of capacity allocation, enterprises typically purchase more servers than they need, thensimply re-provision their servers when there is a capacity overload. As a result, most IT operationscenters are tasked with maintaining and repairing more servers than they can use, yet they stillexperience outages or system performance problems when an overload forces the addition or hot-swapping of a new server in the environment. By contrast, enterprises that use capacity planning

tools generally can anticipate overloads before they occur, thus reducing or eliminating workload-related outages while enabling the IT operations center to reduce the number of servers it mustmaintain.

Capacity planning can also improve the overall performance and availability of the serverenvironment. While experienced IT staffers may be able to make an educated guess as to theircapacity needs at any given time, such guesses are fallible and may not take all contingencies intoaccount. As a result, many IT organizations that do not use capacity planning tools occasionally findtheir servers oversubscribed, delivering slow response times or timeouts that may slow business

Key Capacity PlanningValue Points

Reduce or eliminate serverover-provisioning

Identify and repurposeunderutilized servers

Reduce IT operationalexpenditures

Reduce server downtime

Improve serverperformance andavailability

Improve IT budget accuracy

Leverage futuretechnologies

Like what you see? Subcribe.



6/34


processes or cause customer dissatisfaction. With capacity planning tools, however, the enterprisecan track trendsor, in some cases, model new applications or servicesto anticipate potentialperformance and availability problems and eliminate them before they occur.

Capacity planning tools can also help deliver another key benefit in the current economybudgetpredictability. During the boom years, enterprises could afford to provision their servers on the fly,over-provisioning new services at the outset and then purchasing new capacity when existing servers

were oversubscribed. Under todays tight budgets, however, the ability to predict future server needsis crucial to making intelligent buying decisions. Capacity planning enables the enterprise to developan educated server purchasing strategy, and helps to reduce the need for panic purchasingsituations that a server vendor might exploit to charge premium prices for its equipment.

Finally, capacity planning technology is an essential element in the emerging paradigms of systemvirtualization and utility computing. In order to create an effective pool of capacity to support allnecessary applications and services, enterprises must have a way to track utilization and predictfuture capacity needs. This process of analyzing and projecting capacity requirements will becomeeven more complex when enterprises begin to employ computing grids that are designed tosupport the broad diversity of applications and services used in the enterprise. In the utilitycomputing model, capacity planning will become as important to the enterprise IT organizations as itis to power, telephone, or water utilities today.

The next chapter places capacity planning in the overall context of IT processes.




7/34

Capacity Planning Processes

3 Capacity Planning ProcessesThe IT Infrastructure Library (ITIL)2 describes a standard set of established processes by which an ITorganization can achieve operational efficiencies and thereby deliver enhanced value to anorganization. ITIL is most commonly used in Europe, originally having been created for the U.K.

government. Because it provides a useful best-practice framework that can be put to use relativelyinexpensively, ITIL is growing in popularity as a means to manage IT services within both commercialand governmental organizations.

As shown in the following diagram, Capacity Management is part of the Service Delivery area of ITIL.Capacity Planning is part of the ITIL Capacity Management process.

Figure 3-1ITIL Service Delivery Processes

Further, ITIL defines three sub-processes in the Capacity Management process:

Business Capacity Management: This sub-process is responsible for ensuring that thefuture business requirements for IT Services are considered, planned and implementedin a timely fashion.

Service Capacity Management: The focus of this sub-process is management of theperformance of live, operational IT Services.

Resource Capacity Management: The focus in this sub-process is the management of theindividual components of the IT infrastructure.

The types of capacity planning that we will cover in this eBook focus on the last two sub-processes,service and resource capacity management. Specifically, when we build a single system model, we

are usually performing aspects of resource capacity management rather than service capacitymanagement. This is because services typically span multiple systems. When we combine modelsfrom multiple systems, then we are performing aspects of service capacity management.




8/34

Capacity Planning Processes

The capacity planning function within capacity management interfaces with many other processes asindicated in the following diagram:

Figure 3-2Workloads and Service

As you can see, service level management is where all good IT processes begin, since this is the cruxof the problem. At a minimum, service level management provides service definitions, service leveltargets, business priorities, and growth plans. Within given architectural constraints, capacityplanning determines the cheapest optimal hardware and software configuration that will achieverequired service levels, both now and in the future. Capacity planning then feeds configuration andasset management to make this all a reality so that service delivery can begin. On another processbranch, performance management verifies capacity plans and service levels and provides revisedinput into capacity planning when necessary.

The next chapter provides a high-level approach capacity planning best practices.




9/34

How to Do Capacity Planning

4 How to Do Capacity Planning3

It is very common for an IT organization to manage system performance in a reactionary fashion,analyzing and correcting performance problems as users report them. When problems occur,hopefully system administrators have tools necessary to quickly analyze and remedy the situation. In

a perfect world, administrators prepare in advance in order to avoid performance bottlenecksaltogether, using capacity planning tools to predict in advance how servers should be configured toadequately handle future workloads.

The goal of capacity planning is to provide satisfactory service levels to users in a cost-effectivemanner. This chapter describes the fundamental steps for performing capacity planning. Real lifeexamples are provided using TeamQuest Performance Software.

Three Steps for Capacity Planning

In this chapter we will illustrate three basic steps for capacity planning:

1. Determine Service Level Requirements

The first step in the capacity planning process is to categorize the work done by systems and to

quantify users expectations for how that work gets done.

2. Analyze Current Capacity

Next, the current capacity of systems must be analyzed to determine how they are meeting theneeds of the users.

3. Planning for the Future

Finally, using forecasts of future business activity, future system requirements are determined.Implementing the required changes in system configuration will ensure that sufficient capacitywill be available to maintain service levels, even as circumstances change in the future.

Determine Service Level Requirements

We have organized this section as follows:

a. The overall process of establishing service level requirements first demands an understanding ofworkloads. We will explain how you can view system performance in business terms rather thantechnical ones, using workloads.

b. Next, we begin an example, showing workloads on a system running a back-end Oracle database.

c. Before setting service levels, you need to determine what unit you will use to measure theincoming work.

d. Finally, you establish service level requirements, the promised level of performance that will beprovided by the IT organization.

Workloads Explained

From a capacity planning perspective, a computer system processes workloads (which supply thedemand) and delivers service to users. During the first step in the capacity planning process, theseworkloads must be identified and a definition of satisfactory service must be created.




10/34


Figure 4-1Workloads By Department

Sales

Marketing

Engineering

A workload is a logical classification of work performed on a computer system. If you consider all thework performed on your systems as a pie, a workload can be thought of as some piece of that pie.Workloads can be classified by a wide variety of criteria:

who is doing the work(particular user or department)

what type of work is being done(order entry, financial reporting)

how the work is being done(online inquiries, batch databasebackups)

It is useful to analyze the work done on systems interms that make sense from a business perspective, using business-relevant workload definitions.For example, if you analyze performance based on workloads corresponding to businessdepartments, then you can establish service level requirements for each of those departments.

Business-relevant workloads are also useful when it comes time to plan for the future. It is mucheasier to project future work when it is expressed in terms that make business sense. For example, itis much easier to separately predict the future demands of the human resources department and theaccounts payable department on a consolidated server than it is to predict the overall increase intransactions for that server.

An Example Using Workloads

The TeamQuest View chart below shows 24 hours (cropped to fit the page) of CPU utilization on anIBM F50 PowerPC system. The chart is useful, but it provides a birds eye view of performance, atbest.

Figure 4-2CPU Utilization

In Figure 4-3, below, the Process Table chart of TeamQuest View reveals that during the same 24hour period, 10,642 individual processes ran on this system. All of the utilization information for allof those processes was displayed together in our CPU utilization chart. Wouldnt it be nice if we couldshow a similar chart, but display utilization based on the major functions being performed on this




11/34


system? Using TeamQuest Performance Software, we can do just that, by going through a processcalled workload characterization.

Figure 4-3Process Table

We will leave the detailed instructions for performing workload characterization to another eBook, oryou can refer to the TeamQuest Performance Software documentation. In a nutshell, workloadcharacterization requires you to tell TeamQuest Performance Framework how to determine whatresource utilization goes with which workload. This is done on a per process level, using selectioncriteria to tell the TeamQuest Performance Framework how to determine which processes belong towhich workloads.

Figure 4-4, below, shows a list of workloads that have been characterized so that the work of each of

the 10,642 processes is attributed to one of seven workloads. These workloads are defined accordingto the type of work being done on the system.

Figure 4-4Workload Definitions


http://www.teamquest.com/http://www.teamquest.com/


12/34


If you look carefully you will see six explicitly defined workloads in Figure 4-4, but we said there wereseven. The reason is that there is always an OTHER workload in addition to the explicitly definedworkloads. Any resource utilization that does not match the characterization for any of the explicitlydefined workloads becomes associated with OTHER. This ensures that no performance data fallsthrough the cracks simply because it didnt match any of the defined workloads.

In Figure 4-5, below, pink bars show utilization that did not match the characterization criteria for

any of our define workloads. There is little or no pink in the chart, demonstrating that we have done agood job of explicitly characterizing most of the work done on this server.

Figure 4-5CPU Utilization by Workload

All we did was define some workloads based on the type of work being performed on this server, butnotice how much more useful the information is that is provided in our new chart. Workloads can bevery powerful.

Determine the Unit of Work

For capacity planning purposes it is useful to associate a unit of work with a workload. This is ameasurable quantity of work done, as opposed to the amount of system resources required toaccomplish that work.

To understand the difference, consider measuring the work done at a fast food restaurant. Whendeciding on the unit of work, you might consider counting the number of customers served, theweight of the food served, the number of sandwiches served, or the money taken in for the foodserved. This is as opposed to the resources used to accomplish the work, i.e. the amount of Frenchfries, raw hamburgers or pickle slices used to produce the food served to customers.

When talking about IT performance, instead of French fries, raw hamburger or pickle slices, we

accomplish work using resources such as disk, I/O channels, CPUs and network connections.Measuring the utilization of these resources is important for capacity planning, but not relevant fordetermining the amount of work done or the unit of work. Instead, for an online workload, the unit ofwork may be a transaction. For an interactive or batch workload, the unit of work may be a process.

The examples given in this chapter use a server running an appointment scheduling applicationprocess, so it seems logical to use a calendar request as the unit of work. A calendar requestresults in an instance of an appointment process being executed.




13/34


Establish Service Levels

The next step now is to establish a service level agreement. A service level agreement is anagreement between the service provider and service consumer that defines acceptable service. Theservice level agreement is often defined from the user's perspective, typically in terms of responsetime or throughput. Using workloads often aids in the process of developing service levelagreements, because workloads can be used to measure system performance in ways that makes

sense to clients/users.

In the case of our appointment scheduling application, we might establish service level requirementsregarding the number of requests that should be processed within a given period of time, or we mightrequire that each request be processed within a certain time limit. These possibilities are analogousto a fast food restaurant requiring that a certain number of customers should be serviced per hourduring the lunch rush, or that each customer should have to wait no longer than three minutes tohave his or her order filled.

Ideally, service level requirements are ultimately determined by business requirements. Frequently,however, they are based on past experience. Its better to set service level requirements to ensurethat you will accomplish your business objectives, but not surprisingly people frequently resort tosetting service level requirements like, provide a response time at least as good as is currentlyexperienced, even after we ramp up our business. As long as you know how much the business willramp up, this sort of service level requirement can work.

If you want to base your service level requirements on present actual service levels, then you maywant to analyze your current capacity before setting your service levels.

Analyze Current Capacity

There are several steps that should be performed during the analysis of capacity measurement data.

a. First, compare the measurements of any items referenced in service level agreements with theirobjectives. This provides the basic indication of whether the system has adequate capacity.

b. Next, check the usage of the various resources of the system (CPU, memory, and I/O devices).This analysis identifies highly used resources that may prove problematic now or in the future.

c. Look at the resource utilization for each workload. Ascertain which workloads are the major usersof each resource. This helps narrow your attention to only the workloads that are making thegreatest demands on system resources.

d. Determine where each workload is spending its time by analyzing the components of responsetime, allowing you to determine which system resources are responsible for the greatest portionof the response time for each workload.

Measure Service Levels and Compare to Objectives

TeamQuest Model is a tool that can help us check measured service levels against objectives. Forexample, after building a model of our example system for a three-hour window, 7:00 AM 10:00 AM,the display below (Figure 4-6) shows the response time and throughput of the seven workloads that

were active during this time.By looking at the top line of the table, you can tell that the model has been successfully calibrated forour example system, because the total Measured AR% and Modeled AR% are equal. AR stands forActive Resource. An active resource is a resource that is made 100% available once it has beenallocated to a waiting process. In this case, the active resource is CPU.




14/34


Figure 4-6Response Time and Throughput

Because no changes have been made in the system configuration, modeled response time and

throughput for each workload should closely match reality. In our example, the response time meansthe amount of time required to process a unit of work, which in the case of our application, is anappointment request process. So this report provides us with an appointment request averageresponse time for each of our workloads that we could compare with desired service levels.

Measure Overall Resource Usage

It is also important to take a look at each resource within your systems to see if any of them aresaturated. If you find a resource that is running at 100% utilization, then any workloads using thatresource are likely to have poor response time. If your goal is throughput rather than response time,utilization is still very important. If you have two disk controllers, for example, and one is 50% utilizedand the other is swamped, then you have an opportunity to improve throughput by spreading thework more evenly between the controllers.

Figure 4-7Overall Resource Usage

The table above (Figure 4-7) shows the various resources comprising our example server. The tableshows the overall utilization for each resource. Utilization for the four CPUs are shown togethertreated as one resource, otherwise each resource is shown separately.




15/34


Notice that CPU utilization is about 64% over this period of time (7:00 AM 10:00 AM on January 02).This corresponds with the burst of CPU utilization that was shown earlier in Figures 4-2 and 4-5.

No resource in the report seems to be saturated at this point, though hdisk2 is getting a lot more ofthe I/O than either hdisk0 or hdisk1. This might be worthy of attention; future increases in workloadsmight make evening out the disparity in disk usage worthwhile.

Measure Resource Usage by Workload

Figure 4-8, shows the same period again, only now resource utilization is displayed for theAPPOINTMENTS workload. Note that this particular workload is using 59% of the CPU resource, nearlyall of the 64% utilization that the previous table showed as the total utilization by all workloads.Clearly, the APPOINTMENTS workload is where a capacity planner would want to focus his or herattention, unless it is known that future business needs will increase the amount of work to be doneby other workloads on this system. In our example, that is not the case. Ramp-ups in work areexpected mainly for the APPOINTMENTS workload.

Figure 4-8Resource Utilization by Workload

The previous charts and tables have been useful for determining that CPU Utilization is likely to be adetermining factor if the amount of work that our system is expected to perform increases in the

future. Furthermore, we were able to tell that the APPOINTMENTS workload is the primary user of theCPU resources on this system.

This same sort of analysis can work no matter how you choose to set up your workloads. In ourexample, we chose to treat appointment processes as a workload. Your needs may cause you to setup your workloads to correspond to different business activities, such as a Wholesale Lumber unit vs.Real Estate Development, thus allowing you to analyze performance based on the differentrequirements of your various business units.




16/34


Identify Components of Response Time

Next we will show how to determine what system resources are responsible for the amount of timethat is required to process a unit of work. The resources that are responsible for the greatest share ofthe response time are indicators for where you should concentrate your efforts to optimizeperformance. Using TeamQuest Model we can determine the components of response time on aworkload by workload basis, and you can predict what the components will be after a ramp-up in

business or a change in system configuration.

A components of response time analysis shows the average resource or component usage time for aunit of work. It shows the contribution of each component to the total time required to complete a unitof work.

0

0.05

0.1

0.15

0.2

0.25

0.3

Step: 1

Seconds

All Other AR Queues Queue Delay

All Other AR Queues Service

20-58-00(aixdemo) Queue Delay

20-58-00(aixdemo) Servicehdisk0(aixdemo) Queue Delay

hdisk0(aixdemo) Service

hdisk1(aixdemo) Queue Delay




CPU(aixdemo) Queue Delay

CPU(aixdemo) Service

Figure 4-9APPOINTMENTS Components of Response Time

Figure 4-9, above, shows the components of response time for the APPOINTMENTS workload. Notethat CPU service time comprises the vast majority of the time required to process an appointment.Queuing delay, time spent waiting for a CPU, is responsible for the rest. I/O resources made only anegligible contribution to the total amount of time needed to process each user call.

The ASMAIN workload, shown in Figure 4-10, is more balanced. There is no single resource that is theobvious winner in the contest for the capacity planners attention (however, make note of the queuedelay for hdisk2.)




17/34


0

1

2

3

4

5

6

7

8

9

Step: 1

Seconds



20-58-00(aixdemo) Queue Delay20-58-00(aixdemo) Service









Figure 4-10ASMAIN Components of Response Time

Plan for the Future

How do you make sure that a year from now your systems wont be overwhelmed and your IT budgetover-extended? Your best weapon is a capacity plan based on forecasted processing requirements.

You need to know the expected amount of incoming work, by workload, then you can calculate theoptimal system configuration for satisfying service levels.

Follow these steps:a. First, you need to forecast what your organization will require of your IT systems in the future.

b. Once you know what to expect in terms of incoming work, you can use TeamQuest Model todetermine the optimal system configuration for meeting service levels on into the future.

Determine Future Processing Requirements

Systems may be satisfying service levels now, but will they be able to do that while at the same timemeeting future organizational needs?

Besides service level requirements, the other key input into the capacity planning process is aforecast or plan for the organizations future. Capacity planning is really just a process fordetermining the optimal way to satisfy business requirements such as forecasted increases in the

amount of work to be done, while at the same time meeting service level requirements.Future processing requirements can come from a variety of sources. Input from management mayinclude:

Expected growth in the business Requirements for implementing new applications Planned acquisitions or divestitures IT budget limitations Requests for consolidation of IT resources




18/34


Additionally, future processing requirements may be identified from trends in historicalmeasurements of incoming work such as orders or transactions.

Plan Future System Configuration

After system capacity requirements for the future are identified, a capacity plan should be developedto prepare for it. The first step in doing this is to create a model of the current configuration. From this

starting point, the model can be modified to reflect the future capacity requirements. If the results ofthe model indicate that the current configuration does not provide sufficient capacity for the futurerequirements, then the model can be used to evaluate configuration alternatives to find the optimalway to provide sufficient capacity.

Figure 4-11Setting Up TeamQuest Model

Our base model in Figure 4-11 was representative of 300 users generating appointment/calendarrequests on our system. From there we have added steps in increments of 30 users until theincoming work has increased by one-half again as much. This ramp-up in work has been added onlyfor the APPOINTMENTS workload, because no such increase was indicated for other workloads when

we did our analysis of future processing requirements in the previous section.

TeamQuest Model will predict performance of the current system configuration for each of the stepswe have set up. Figure 4-12 is a chart generated using TeamQuest Model showing the predictedresponse time for each workload.




19/34


300

390

APPOINT

MENTS

A

SMAIN

BOURNE

GPMS

OTHER

STAFF

SYSTEM

0

2

4

6

8

10

12

1416

18

20

A PPOINTMENTS

ASMAIN

BOURNE

GPMS

OTHER

STAFF

SY STEM

Figure 4-12Predicted Response Time

As we can see, the response time starts to elongate after 390 users.

Figure 4-13 is a chart showing predicted response time using a stack bar chart that also shows thecomponents of response time. Notice the substantial increase in CPU wait time after the number ofuses reaches 360. It seems that the performance bottleneck is CPU resource.

Seconds




20/34


0

0.2

0.4

0.6

0.8

1

1.2

1.4

300 330 360 390 420 450

Seconds




20-58-00(aixdemo) Service









Figure 4-13

Predicted Response Time

Figure 4-14 shows the same stack bar chart, but this time TeamQuest Model was told to predictperformance if the system involved was changed to a p670 1100Mhz 4-CPU system. Clearly, thenewer, faster architecture not only allows us substantial growth, but reduces our overall responsetime to a more realistic level and still allows us the headroom to experience additional growth ifneeded.




21/34


0

0.2

0.4

0.6

0.8

1

1.2

1.4

300 320 360 390 420 450

Seconds













Figure 4-14

Predicted Response TimeAfter Upgrade

Capacity Planning Process

In summary, we have shown these basic steps toward developing a capacity plan:

1. Determine service level requirementsa. Define workloadsb. Determine the unit of work

c. Identify service levels for each workload

2. Analyze current system capacitya. Measure service levels and compare to objectivesb. Measure overall resource usagec. Measure resource usage by workloadd. Identify components of response time

3. Plan for the futurea. Determine future processing requirementsb. Plan future system configuration

By following these steps, you can help to ensure that your organization will be prepared for thefuture, ensuring that service level requirements will be met using an optimal configuration. You will

have the information necessary to purchase only what you need, avoiding over-provisioning while atthe same time assuring adequate service.

The next chapter provides example applications of capacity planning for todays data center issues.




22/34

Using Capacity Planning for Server Consolidation

5 Using Capacity Planning for Server Consolidation4

With the right tools, you can locate underutilized capacity and move applications and subsystems totake advantage of that capacity, oftentimes delaying the purchase of additional servers. This is justone of many IT optimization strategies that falls under the definition of server consolidation.

This chapter explains server consolidation and provides two real-world examples where performancemanagement and capacity planning tools were used in consolidation projects. These projects savedthe organizations involved thousands of dollars in one case, and millions in another, mainly byhelping them to use resources they already had on hand.

Server Consolidation Defined

Consolidation by definition is the act that brings together separate parts into asingle whole. Typically, server consolidation refers to moving work or applicationsfrom multiple servers to a single server. Server consolidation is just one aspect ofcapacity planning.

Why, Oh Why?

There are many reasons to consolidate servers, but the underlying reason isultimately, To save money. The goal is to lower the Total Cost of Ownership (TCO)

information technology and increase Return on Investment (ROI). By increasingncy, IT departments can do more with less and make optimal use of what they already have.

In a typical

forefficie

server consolidation effort, you can increase efficiency by:

ments

l space, hardware systems and software

This a server consolidation was used to increase efficiency by:

ing servers

Whe ran on one or a few big boxes. This simplified

ol means a hodge-podge of servers and techniques used to

introduce management

Reducing complexity Reducing staff require Increasing manageability Reducing costs for physica Increasing stability/availability

ch pter will provide examples where Delaying new purchases Extending the life of exist Optimally utilizing existing servers

n mainframes ruled the earth, everythingmanagement. Now even where there are mainframes in an organization there are frequently a greatmany distributed servers as well. Running many distributed boxes offers a great deal of flexibility,and when business departments own those boxes rather than IT departments, they get to havetheir own playground. Applications can be rolled out without going through a bureaucratic quagmire,and expensive mainframe expertise is not required. This flexibility allows for rapid deployment, butthe cost is a management nightmare.

Decentralized management and contrmanage them. Management is primarily reactive, oftentimes with data center managers beingconsulted only after disaster strikes, when problems are far out of hand. Because of this, it is notunusual for the IT department to provide the impetus to consolidate servers.

Today there is a general trend toward server consolidation. The idea is towhere chaos previously ruled, reducing complexity and optimizing existing IT resources.




23/34


Finding the Answers Can Be Difficult

Server consolidation sounds simple enough, but it is not. How much capacity isavailable on each server? What resources are underutilized? Is the CPU? Whatabout memory?

What happens if work is moved in from another server? Will there be enough

resources for both applications? Will I/O be a bottleneck after the move? Whatwill response time be like?

Can more than one application be moved to that server? What are the overall effects of increasing thework on that server? Is there any space for future growth? How long will it be before businessworkloads increase to the point that the CPU or other resources are over-utilized?

These are all true concerns and legitimate questions when talking about server consolidation, andwithout the proper tools, finding answers can be difficult.

Predict the Future

The tool of choice for many server consolidation experts is TeamQuest Model. TeamQuest Model usesanalytic modeling based on queuing theory to rapidly evaluate various hypothetical system

configurations. It also predicts performance based on projected business growth. Most importantlyfor this chapter, it can predict how applications will perform after servers are consolidated.

With TeamQuest Model you can:

1. Measure work being done on each individual server.2. Model effects of consolidating work from multiple servers onto a single

server.3. Quickly identify any performance bottlenecks that are likely to be

caused by the additional work being done on the server.4. Predict effects of future growth in your business using what if

scenarios. What if the work grows by X percent over X amount oftime?

5. Predict effect on performance of changing hardware configurations.

What if another CPU is added? What if disk drives are added, or anothercontroller? What if the make or model of server is changed?

The most important benefit of using TeamQuest Model is that you can see the effects of changesbefore any work or data are actually moved. You can then be more confident regarding the effects ofthe server consolidation before you make any purchases and before data or applications are moved.

Case Studies

Below are two examples of how server consolidation was used in different situations, saving bothcompanies money. The examples are real, but we have changed the names to protect our customersinterests. Company A used capacity planning and server consolidation to accomplish big costsavings in their disaster recovery/management plans. Company B discovered they could make do

with existing capacity, and canceled an order for a large server.

Company A Finding Underutilized Resources for Use as Disaster Backup

A disaster management consultant was hired for Company A. Company A had three locations, one inCalifornia, the second in New York, and the third in Kansas City. All three sites used a large SANenvironment so that each site had access to the same data. There were a total of some 900 servers.The consultant was to come up with a disaster management plan that would be able to recover 200 ofthe most critical servers in case of a disaster. 100 of these mission critical servers were located in the




24/34


California office, 50 in New York and 50 in Kansas City. In case of a disaster, the mission criticalservers had to be recovered and operational within a 24-hour time period.

This consultant started looking at the workloads and utilizations of each server, and found that therewas one business application per server. Upon interviewing system administrators (SA), theconsultant learned that Company A had at one point tried running multiple applications per server.This resulted in unacceptably poor performance and the people responsible for the applications

pointing fingers at each other.

The company thought they could resolve the problem quickly by putting each individual businessapplication on its own server. This turned out to be a very difficult, time consuming and expensivetask. To avoid having to separate applications again in the future, a company-wide policy wasinstituted requiring each business application be hosted on its own dedicated server.

H ndsight being 20-20, it would have been prudent to have TeamQuesti View installed onthe servers, helping to analyze performance bottlenecks. It might have been possible toeliminate problems with just a few carefully chosen changes or upgrades. Even better,

TeamQuest Model could have been used to explore alternative configurations to meetcapacity requirements.

Instead, the company chose what appeared to be the most

expedient solution to their performance finger-pointingproblem; they hosted just one application per server. This isnot an uncommon solution, but it usually leads to inefficientuse of IT resources.

Using TeamQuest View the disaster management consultantanalyzed a few of the larger critical servers and found they wereless than 3% to 6% utilized at any given time of the day ornight. There were vast amounts of underutilized capacity thatcould potentially be used as backup in the event of a disaster.

The consultant then used TeamQuest Model to evaluate apotential 10-to-1 server consolidation. He modeled 10 of themission critical servers from California, and moved the work to a mission critical server in Kansas

City. The model showed that even with the consolidation of these servers, the server in Kansas Citywould be less than 71% utilized.

The consultant then did an analysis of the historicalgrowth of the workloads on servers that resided inCalifornia for the preceding year. He used what-ifanalysis capabilities of TeamQuest Model to predict theutilization for the next 12 months based on the growthpatterns of the previous year. This revealed that after theconsolidation and the projected growth for the next 12months, the server would still be less than 80% utilized.

The consultant convinced management to let him use theserver in Kansas City as a backup site for the 10 missioncritical servers. He mirrored the individual server fromCalifornia to the one server in Kansas City, and foundperformance to be within a small percentage of what hadbeen projected with TeamQuest Model.

The consultant is currently modeling the entire enterprise, including all three sites in the analysis. Itis quite likely that no new computer resources will be required to provide disaster back-up, resultingin a very significant savings, a very pleasant surprise for upper management.

Underutilization MeansOpportunity

Consolidation Saves the Day




25/34


Company B Avoiding Unnecessary Expenditures

Company B was a small company with a System Administrator (SA) who was in charge of six serverswithin his department. One of those servers was leased, and the lease was expiring at the end of themonth. The hardware vendor convinced Company B to order the newest, latest, greatest (and mostexpensive) server on the market. If the new server arrived as scheduled, service levels could bemaintained, but the SA was concerned. What would he do if the new server did not arrive before theold leased equipment had to be returned?

The work being done on the leased server included an instance of Oracle and two customapplications. Could that processing be temporarily moved to any of the other five servers?

The SA used TeamQuest Model to analyze all six servers. He found that no single server would be ableto take the entire load of the leased server, but he did find a way to distribute the load over twoexisting machines. He could move the Oracle instance to a machine that had another small instanceof Oracle running on it. If he added two more CPUs to that server, there would be sufficient capacityfor the next 12 months with the current growth patterns. He foundanother server that had enough capacity to host the customapplications without any modifications for the next 12 months.

With sufficient capacity to cover growth over the next 12 months,

there was no need for the leased server or the replacement serverthat had been ordered. The SA quickly took these findings to hismanagement, and the changes were successfully implemented overthe next weekend. The order for the new server was cancelled, andthe SA received a healthy bonus for saving the company asubstantial amount of money.

Eliminated Need forExpensive Replacement

Conclusion

These are just two examples of how using a good capacity planning tool to perform serverconsolidation can save money. Company B saved thousands of dollars, and Company A literally savedmillions, dramatically demonstrating that performance management and capacity planning is bestdone with the proper tools. To do otherwise creates risk instead of minimizing it.

To make optimal use of your organizations IT resources, you need to know exactly the amount ofcapacity you have available and how that capacity can be utilized to its fullest.




26/34

Capacity Planning and Service Level Management

6 Capacity Planning and Service Level ManagementIn this chapter we discuss service level management and its relationship to capacity planning.

We hear a lot about service level management today, but what do we really mean when by it? In the

boxes below we show side-by-side definitions of service level management and capacity planning.The service level definition is from Sturm, et al, Foundations of Service Level Management.5

You can clearly see that these two critical IT processes really share common goalsprovide adequateservice for the least cost. So how do you really do this? You must employ an integrated approach toservice level management and capacity planning, as this section describes. Stated slightlydifferently:

Service Level Management (SLM)

is the disciplined, proactivemethodology and procedures usedto ensure that adequate levels ofservice are delivered to all IT users inaccordance with business prioritiesand at an acceptable cost.5

Capacity planning

is the process by which an ITdepartment determines the amount ofserver hardware resources required toprovide the desired levels of servicefor a given workload mix for the leastcost.

Capacity planning requires a firm understanding of required service levels

Service level management is successful when the necessary IT architecture and assetsare available

Operational processes should incorporate both capacity planning and SLM to maximizesuccess

We will now examine two key aspects of service level management, setting service level agreementsand maintaining service levels over time as things change.

Service Level Agreements (SLAs)

Many of the IT organizations we talk to do not have any type of SLA mechanism in place and this is aconcern. How can you adequately provision a service if you do not have goals to shoot for? And howcan you plan upgrades if you do not know what the business expects for this service over time? Werecommend at least recording the following information for each service:

The business processes supported by the service The priority to the business of these processes The expected demand for this service and the seasonality, if any, of that demand The expected growth in demand for this service over the next three years The worst response time or throughput that is acceptable for this service

You do not have to make this process formal with signed agreements, but it is a very good idea forboth IT and the affected business units to have a copy of the above information to resolve conflictsshould they arise.

With this minimal information, or a more complete and formal SLA if you have it, you are now ready touse capacity planning to ensure performance that is adequate at a minimum cost.




27/34


The key SLA questions to answer are:

What hardware configuration is required for the performance and workload level? Can the service be provisioned on existing systems and meet the requirements? What hardware upgrade plan must be in place for anticipated business growth?

There are many ways to determine how much hardware capacity you need to deliver a service level.

You can perform load testing, but this assumes you can afford exact duplicates in the test lab of theproduction floor, and load testing fails in a consolidated environment anyway. As an alternative, youcan use capacity modeling to determine how much hardware you will need.

About the only way to find out if a service can be provisioned on an existing system with the specifiedservice level is to use capacity modeling. And what about after it is up and running? How do youmaintain the service level as workloads grow? When do you need to upgrade, if at all? Capacitymodeling is often the best way to know for sure. We will present best practices for doing this in amoment.

Proactive Management of Resources to Maintain Service Levels

Essentially you must answer the following questions:

Do predicted service levels match real service levels? Can we consolidate applications or servers to save money and still provide the required

service? If growth plans change, what does this mean to our upgrade plan?

How do you manage resources to maintain service levels and minimize costs? First you need tomeasure service levels and compare to specified and predicted. This requires a performancemanagement system with workload measurement. Are you providing adequate service levels but atthe expense of over-provisioning? You can find this out by (a) looking at historical resource utilizationfor the service and the system and looking for headroom, or (b) by building a model and stressing it tosee where it breaks.

If you have a service over-provisioned for now, is it also for the foreseeable future? If yes, you shoulddefinitely consolidate or provision a new service here. It is always a good idea to check with business

units or other client organizations periodically to verify growth plans in the SLA. What if they recentlychanged? The best thing to do is get out your model of this service and plug in the revised growthprojections and see how your upgrade plan looks now.

Service Level Management Scenarios

The following scenarios illustrate capacity planning applications for service level management.

Provisioning a new service to meet an SLA

1. Perform load planning and testing.2. Baseline the test runs.3. Build a model of the test environment.

4. Change the model parameters to the production environment.5. Solve the model for the projected workloads.6. Compare predicted response time/throughput to service level

parameters.7. Make necessary changes to planned provisioning.




28/34


Adding an SLA to an existing service

1. Build a model of the production environment.2. Solve the model for the projected workloads.3. Compare predicted response time/throughput to service level

parameters.

4. Make necessary changes to planned provisioning.

Developing an upgrade plan under an SLA

1. Gather projected growth estimates from business units.2. Build a model of the existing service based on current

configuration and workload.3. Solve the model for one, two and three year growth projections.4. Compare predicted response times and throughputs to the SLA.5. Use modeling what-if scenarios to find least cost, just-in-time

upgrade path.

To summarize, the goals of service level management and capacity planning are tightly interwoven.Both strive to provide adequate levels of service at an acceptable cost. And both are much moresuccessful when implemented together as part of a disciplined and unified process.

The final chapter introduces the TeamQuest solution for capacity planning.




29/34

Product Focus: TeamQuest Model

7 Product Focus: TeamQuest ModelOne of the leading vendors in the capacity planning space isTeamQuest Corp., which makes tools both for performance analysisand capacity planning. Founded in 1991 as a spin-off of Unisys

Corp., TeamQuests software has become one of the industrysmost established capacity planning solutions, and is used by suchmajor corporations as 20th Century Fox, AT&T, Nike and 3M Corp. While TeamQuest offers a variety ofproducts designed to help enterprises optimize performance, its primary capacity planning tool isTeamQuest Model (Model), which offers full planning and modeling capabilities. The full TeamQuestproduct line is depicted in Figure 7-1.

One of the first criterion that distinguishesModel from its competitors is its supportfor multiple hardware platforms. Whilethere are many vendor- or platform-specifictools for capacity planning, the number ofmulti-platform solutions is much smaller.

Many vendors have found it difficult toinstrument tools that can collect andextrapolate data coming from platforms asdifferent as Microsoft Windows and Unix,but that capability is essential for mostlarge enterprises, which usually support avariety of different servers.

Another key differentiator is Modelsmultiple system modeling capability thatallows management of the capacity of amulti-tiered application from a holisticpoint of view. Models ability to specify andplay what-if with the horizontal scaling of a tier is unique. The impact of changing or adding CPUs maybe easily modeled, showing the impact of adding servers to the tier. Additionally, Model supportssplit workloads, showing application flows from one server to the next, recognizing that todaysapplications do not move in a single path through the tiers.

From a functional perspective, Models chief benefit is its ability to provide predictive analysisrequired for long-term capacity planning. It is a PC-based application that automatically retrieves datafrom Windows and Unix servers and then builds models of systems and applications that enable theuser to analyze what if scenarios and predict the effects of proposed changes on systems andapplications before they are made. This capability is essential to IT organizations not only forplanning system purchase and configuration, but to predict the impact of IT changes and escalatingbusiness workloads on the performance of business services.

Two example TeamQuest Model screens are shown below; the first depicts predicted performance as

the number of users increases over time, and the second depicts the projected performance if aproposed system upgrade is performed. Clearly the proposed upgrade will adequately prepare thesystem for future workloads. In fact, it may be overkill. TeamQuest Model would allow the easyexploration of additional, perhaps less expensive, alternatives.

Most of this chapter isexcerpted from a white paper1

by Enterprise ManagementAssociates (EMA), an analyst

organization specializing in themanagement software and

services market.

Figure 7-1Team uest Performance Framework




30/34


0

0.2

0.4

0.6

0.8

1

1.2

1.4

300 330 360 390 420 450

Seconds













Figure 7-2

TeamQuest Model Shows Predicted Response Time

0

0.2

0.4

0.6

0.8

1

1.2

1.4

300 320 360 390 420 450

Seconds




20-58-00(aixdemo) Servicehdisk0(aixdemo) Queue Delay








Figure 7-3

Predicted Response Time After Upgrade

Models most obvious benefits are in planning the purchase and deployment of new hardware. Thesoftware offers an array of pre-built capacity profiles, enabling IT organizations to use real-world datato perform what if queries on the infrastructure impact or performance of planned hardwareinstallations. Using Model, enterprises can get a real-life picture of how a new server will affectcapacity and performance, even before that server is purchased or installed.




31/34


Model also enables enterprises to test the effects of various server consolidation scenarios, making iteasier to find the most efficient configuration of hardware and expose potential consolidationconflicts before they occur. Model offers an advantage over other capacity planning products in thisarea, because it can integrate, analyze, and compare data from multiple servers to aid in serverconsolidation efforts.

Aside from server acquisition and consolidation, Model offers a number of other potential benefits for

IT planning. For example, Model can be used to measure the effectiveness of proposed workloadbalancing strategies, such as examining the impact of increased activity on a particular server or ofmoving some types of processing to a different part of the day. In addition, Model has been used tomodel capacity for newly introduced applications, or even the impact of infrastructure consolidationthat may take place as a result of mergers or acquisitions.

While Model can make a significant difference in efforts to evaluate and measure the impact of futureIT initiatives, it can be equally effective in facilitating day-to-day activities, such as the ongoingmonitoring and modeling of server capacity. For example, Model can be used to predict when aparticular server will run out of capacity, given the projected business workload growth rate. This datacan be used to help IT organizations precisely plan their strategies for system upgrade, as well asproviding data on exactly what the upgrade requirements might be.

Model can also be used to support the development, analysis and enforcement of established servicelevels, whether they are delivered by internal IT organizations or external service providers. Modelcan help organizations model the time, costs, and resources required to maintain different levels ofperformance, enabling service providers and their end-user customers to reach mutually-satisfactoryservice level agreements that are based on realistic expectations and scientific analysis, rather thanwishful thinking. With Model, IT organizations can benchmark a variety of service levels to determinetheir costs and then set realistic customer expectations for maintaining them.

In fact, the sophistication of Models ability to predict performance is one of the elements that sets itapart from other capacity planning tools on the market. IT organizations can use Model to predict avariety of performance metrics, including response time, throughput, and queue length as well asresource utilization. Model is the only capacity planning tool on the market that offers a choicebetween analytic modeling and discrete event simulation, which means that users can employ Modelto gain insights on overall performance or on the impact of specific IT events in the enterprise, such

as infrastructure upgrades.

Case Studies

Clearly, Model offers a range of advantages on paper, but how does it work in practice? To answer thisquestion, EMA interviewed two Model users: the U.S. Patent Office and a US GovernmentSubcontractor that maintains records and data about U.S. military personnel and their dependents.The following is a synopsis of those interviews.

United States Patent and Trademark Office

Three years ago, the U.S. Patent Office (www.uspto.gov) had a major performanceproblem in a critical application. A key search system was suffering major performance

problems, yet IT technicians needed a method to model and design specific scenariosto optimize the proposed solution. The technicians contacted Hewlett-Packard, makerof the Patent Offices HP MeasureWare performance management software, and an HPcontact suggested that the TeamQuest Model performance-modeling tool could savevaluable time in optimizing the proposed solution.

By employing analytic modeling and validating its search system application using Model, the PatentOffice found the I/O bottleneck immediately. After locating the problem, technicians used Model tohelp test a variety of channel striping sizes and find the appropriate Fibre Channel configuration.




32/34


Model enabled the Patent Office to aggregate statistics for each storage drive across the FibreChannel, making it possible to identify bottlenecks and then determine the best method for dividingI/O across the channel to eliminate those bottlenecks.

Today, the Patent Office is using Model regularly to import, analyze, and enhance data fromMeasureWare. Model captures workload data from existing systems, and applies that data to thePatent Offices new system model, projecting workload growth over the lifespan of the server. This

approach enables the Patent Office to avoid over-provisioning over the life of the system, because theIT organization can accurately predict the need for future capacity.

And what of the Patent Offices search system? In early 2000, the Patent Office was awarded aDepartment of Commerce Silver Medal Award for the Patent Search Performance Teams improvementof its multimillion-dollar search engine project. IT executives at the Patent Office say that TeamQuestModel has saved them millions of dollars in over-provisioning costs, and helped make the searchsystem project a roaring success.

US Government Contractor

A contractor to the US Government that provides services to many government agencies (and whomrequested anonymity) has a database engine that contains information on hundreds of thousands ofmilitary personnel and their dependents. Just a few years old, the system has become enormouslypopular, growing from two Unix servers to nearly 200 since it began operations.

Such rapid growth was good news for the company, but bad news for its IT organization, which wasscrambling to keep up with the skyrocketing demand for server capacity. To have any chance to catchup, the company needed automation and planning technology to facilitate growth. The company alsoneeded performance reporting and tracking capabilities to ensure good service and debugperformance problems.

In 2001, the companys sole server provider, Sun Microsystems, recommended Model to thecompanys IT department. TeamQuest technicians came to the company and installed Model on sixWeb servers to collect data and create some sample models for IT staff. Once company officials gavethe go-ahead, the full Model package was installed and operational within a few days.

Since its installation, TeamQuest Model has enabled the company to save millions of dollars in over-

provisioning costs, according to officials associated with the Model deployment. In one instance,Model modeling helped the company find a configuration that spread the workload across existingSun servers, reducing the need for new servers from 30 to 20 and saving the organization millions ofdollars.

Today, Model is helping the company maintain consistent levels of service across its governmentcustomer base, which is a crucial element to its business. Model enables the company to modelserver performance and establish service level agreements with realistic thresholds, making it easierfor the organization to meet or beat customer expectations, even as its server infrastructurecontinues to evolve.

EMA Perspective

In a difficult economy, every IT dollar matters. The days of solving performance problems byindiscriminately buying more servers are over, and there is an excellent window of opportunity fordeploying management technology that can optimize server utilization and performance. ITorganizations are looking for ways to make better use of existing server capacity while reducing theneed to buy more processor power.

At the same time, it is important to note that IT budgets are tight, and that most IT organizations arenot prepared to purchase new management or planning software unless they are guaranteed a clear




33/34


and fast return on their investment (ROI). Todays management tools must be affordable, easy todeploy, and show strong results in a very short period of time.

Capacity planning tools meet all of these criteria. Through capacity planning, IT organizations canrealize major cost savings immediately by reducing or eliminating the need to over-provision theirserver environments. Capacity planning and modeling technology also can guide IT through theserver consolidation process, and can help set realistic goals for service levels that can be

maintained over time.

EMA has established that TeamQuest is a superior partner for capacity planning. TeamQuest hasbeen in the performance assurance business since 1991, and has a legion of satisfied customers. Thecompany has an excellent reputation in the industry, and all of the customers interviewed for thisreport offered high praise for the TeamQuest organization, from pre-sales to post-sales. The Modelproduct has several features that separate it from its competitors, including multi-platform supportand the ability to do broad or event-driven capacity analysis and modeling.

EMA recommends TeamQuest Model to IT executives that are concerned with doing more with lessin todays difficult economy. Model clearly demonstrates the ability to recognize major cost savingsand increased efficiencies in a short amount of time, generating excellent value and a strong ROI forthe IT investment dollar.




34/34

Bibliography

8 Bibliography1. The Value Proposition for Capacity Planning, Enterprise Management Associates, 2003.

2. The Office of Government Commerce web site, http://www.ogc.gov.uk.

3. How to Do Capacity Planning, TeamQuest Corporation, TQ-WP23 Rev. B.4. Consolidating Applications onto Underutilized Servers, TeamQuest Corporation, TQWP20

Rev. A.

5. Sturm, Rick & Morris, Wayne, Foundations of Service Level Management, SAMS, 2000.

Like what you see?Subscribe.

TeamQuest Corporation

AmericasOne TeamQuest WayClear Lake, Iowa 50428USA+1 641 357-2700+1 800 551-8326

[email protected]

Europe, Middle East and AfricaBox 1125405 23 GteborgSweden+46 (0)31 80 95 00

United Kingdom38 The Old WoodyardHagley HallHagleyWorcestershire DY9 9LQ

+44 (0)1562 [email protected]

Asia PacificLevel 6, 170 Queen StreetMelbourne, VIC 3000Australia+61 3 9641 [email protected]

Legal Notices

TeamQuest and the TeamQuest logo are registered trademarks in theUS, EU, and elsewhere. All other trademarks and service marks are theproperty of their respective owners. No use of a third-party mark is tobe construed to mean such marks owner endorses TeamQuestproducts or services.

The names, places and/or events used in this publication are purelyfictitious and are not intended to correspond to any real individual,group, company or event. Any similarity or likeness to any realindividual, company or event is purely coincidental and unintentional.

NO WARRANTIES OF ANY NATURE ARE EXTENDED BY THE DOCUMENT.Any product and related material disclosed herein are only furnishedpursuant and subject to the terms and conditions of a licenseagreement. The only warranties made, remedies given, and liabilityaccepted by TeamQuest, if any, with respect to the products describedin this document are set forth in such license agreement. TeamQuestcannot accept any financial or other responsibility that may be theresult of your use of the information in this document or softwarematerial, including direct, indirect, special, or consequentialdamages.

You should be very careful to ensure that the use of this information

and/or software material complies with the laws, rules, andregulations of the jurisdictions with respect to which it is used.

The information contained herein is subject to change without notice.Revisions may be issued to advise of such changes and/or additions.

U.S. Government Rights. All documents, product and related materialprovided to the U.S. Government are provided and delivered subject tothe commercial license rights and restrictions described in thegoverning license agreement. All rights not expressly granted thereinare reserved.

capacity planning discipline for data centers

Documents