VMware Capacity Planner - Creating A Capacity Blueprint

 

W H I T E

Capacity Planning: The Blueprint for Server Consolidation

P A P E R

 

VMWARE   W H I T E

PAPER

Table of Contents

Executive Summary .......................................................... ...................................... 3 Growing Support for Consolidation..................................................... .................. 3 Standard Approaches Approaches to Capacity Planning ................................................. ......... 3  

Inventory .................................................. ......................................................... 3

 

Utilization of Current Capacity ...................................................... ................... 3

 

Utilization Variation over Time ...................................................... ................... 4

 

Analysis: Determining How to Consolidate ...................................................... 4

CapacityPlanner............................................................................................. ......... 5  

Analyze ..................................................... ......................................................... 6

 

Plan ........................................................... ......................................................... 6

 

Model ........................................................ ......................................................... 6

 

Monitor ..................................................... ......................................................... 6

How CapacityPlanner CapacityPlanner Work Workss ................................................................ .................. 6

2

 

Data Collector .................................................... ................................................ 6

 

Data Manager .................................................... ................................................ 7

 

Information War Warehouse ehouse ......................................................... ............................ 7

 

Data Analyzer .................................................... ................................................ 7

 

CapacityPlanner Dashboard.............................................................................. 7

 

Rigorous Information Collection Collection..................................................... ..................................................... .................. 7

 

Intelligent Capacity Analysis........................................................... .................. 7

 

VMWARE   W H I T E

PAPER

Capacity Planning: The Blueprint for Server Consolidation Executive Summary

Standard Approaches to Capacity Planning

Server consolidation has become a key element in IT planning in recent years because it offers a range of benefits — from lower hardware, software, and personnel costs to improved reliability and the management efficiencies of standardized infrastructure.

 To chart the most effective approach approach to consolidation in a particular IT environment, analysts need a variety of information on the existing infrastructure and data on how the hardware in that existing infrastructure is being used.

Virtualization is a key enabling technology encouraging the server consolidation trend and is now becoming a core part of strategic IT planning. Planning for server consolidation through virtualization can be a complicated undertaking. Large IT environments have dozens — sometimes hundreds or thousands — of servers ser vers running a multitude of applications and services for a wide range of departments, owners, and business domains. Deciding how to best combine these into fewer, more manageable physical resources, while at the same time planning for future expansion, unexpected demands, and organizational changes, can be a daunting task. VMware Capacity Planner provides insights into the server resources available in the IT infrastructure and the ways those resources are being used. It integrates and displays data on server inventory and performance so system analysts have the information they need to build the greatest possible efficiency into a server consolidation plan.

Growing Support for Consolidation Server consolidation has been a hot topic in recent years. Industry analysts from such organizations as IDC and Gartner along with writers for top trade publications have found that server consolidation offers such benefits as: • Reduced hardware and maintenance costs • More efficient use of datacenter datacenter space • Simplified and more consistent operating operating environment environment • More effective management of enterprise IT resources resources • Improved reliability and and flexibility Virtualization has shown itself to be one of the most important approaches for enabling consolidation. It is being adopted as a core part of strategic IT planning in organizations of all sizes.  This white paper explains the issues involved in capacity

Inventory   To make good decisions about capacity capacity planning, and about server consolidation as a step in the process, the project team must begin by obtaining a detailed understanding of what capacity is currently present. A starting point of any inventory exercise is simply to count the existing resources. For server consolidation and other capacity planning activities, project teams also need detailed information on four core hardware components: processors, processors, memory, storage, and network adapters. Detailed knowledge of applications, services, and shares is equally valuable.  Traditionally, consultants and customers have to collect the  Traditionally, data manually — a step that can be costly. Agent-based management tools can help, but they are seldom seld om fully implemented across an entire enterprise. The gaps may be a result of budget constraints, lack of internal process, or lack of knowledge about where the organization’s servers reside. Unknown servers may exist because of purchases by independent departments, de partments, recent mergers, or refreshes that did not include procedures for disposal, among other reasons. reasons. The average enterprise underestimates the number of servers in its environment by 20 percent. In some cases this miscalculation reaches 50 percent.

Utilization of Current Capacity  Besides knowing what capacity exists, the project team needs to analyze how that capacity is being used. The team might gauge performance in a multitude of ways; however, a core set of performance metrics can identify ide ntify the utilization and throughput for such key server resources as processor processor,, memory, network adapter, and storage. It is also important to capture additional performance metrics for specific applications. Inventory tools focus on inventory data collection. Performance tools focus on performance data d ata collection. Few tools provide both sets of data. If a tool collects both, the tool is typically agent-based. As in the case of inventory information, an agentbased tool's data is accurate and complete only if the agent software is installed on ever y computer.

planning, why Capacity Planner is the ideal tool for addressing those issues, and how it can achieve specific consolidation results. 3

 

VMWARE   W H I T E

PAPER

 To turn the performance data into useful information, information, it must be correlated with inventory data. Many organizations attempt to use a performance per formance monitoring tool by itself. The The performance logs paint a picture like that shown in Table Table 1. Ser ver

CPU% Utilization

A

10%

B

20%

C

30%

D

40%

Overall performance

25%

Table 1: Uncorrelated Performance

For server consolidation and capacity management, the conclusion that utilization is 25 percent is inaccurate. When inventory and performance information are combined, as in Table 2, the results give a more useful picture.

Utilization Variation over Time Calculating average utilization has some value, but it is more important to know peak hour utilization. Peak utilization will differ within a 24-hour period based on the types of users, applications on the server, scheduled maintenance, and so forth. The peak hour is the one hour in the 24-hour period that has the highest average utilization. Average utilization is calculated using data collected between the hours of 7 a.m. and 7 p.m., for example, e xample, or for the entire e ntire 24-hour period. The difference between these two metrics is significant, as can be seen in Figure 1. 30.0% 25.0% 20.0% 15.0% 10.0% 5.0% 0.0% 1

3

5

7

9

11

13 13

15 15

17 17

19 19

21 21

23 23

Ser ver

CPU% Utilization

CPU Capacity

Utilization

A

10%

3000MHz

300MHz

B

20%

1500MHz

300MHz

Figure 1: Comparison of hourly measurement vs. daily average of utilization

C

30%

500MHz

150MHz

D

40%

200MHz

80MHz

5200MHz

830MHz

Most system tools focus on real-time performance statistics or 24-hour averages. Many tools collect the data but then require users to take time-consuming time- consuming steps to extract and manipulate performance data to determine the peak hour average. As noted in other sections of this paper, it is impossible to collect data from servers using these tools unless the servers have the necessary system management tools installed. And it is unlikely the system management tools will be installed on servers that are not included in an official inventory.

Overall capacity utilization

16%

Table 2: Correlated Performance

 The correlated data show more opportunity for server consolidation, revealing that capacity utilization is only 16 percent.  The analysis based on percentage of CPU utilization alone was distorted by the fact that older, slower CPUs were being used more intensively than newer, faster CPUs. VMware has found that 40 percent of the servers at a typical client site are slower than 500MHz. Capacity Planner has found new, state-of-the-art servers running below ideal capacity while older servers are pushed to their capacity limits.  The capacity calculation is further complicated by the fact that particular processors have different capabilities. To provide an accurate picture of resource utilization, the analysis tools must adjust for differences between 64-bit and 32-bit processors and for differences between such new technologies as the AMD Opteron and older technologies such as the Intel Xeon.

Hour

Uti tili lizzati tion on

Aver Av era age

 Analysis  Anal ysis:: Deter D etermini mining ng How to Conso Consolida lidate te

After all the necessary and relevant information has been obtained, the most important step follows: using this information to recommend changes that will consolidate the environment in an effective yet manageable way. Several considerations should be addressed to achieve this goal. First, the performance of the applications in the current environment may not be optimized. For example, any of the following conditions can lead to excessive resource usage: • Poorly configured storage storage resources • Excessive virus scanning scanning on Exchange Exchange systems • Bad index design design on SQL Server systems • Excessive logon processing in Citrix environments environments Although the project team could undertake a performance study on each application and perform per form tests to see where additional efficiencies can be wrung out of a system, this process is

4

 

VMWARE   W H I T E

PAPER

very time-consuming and requires expertise beyond what most IT organizations are willing to budget. A more attractive alternative is to compare utilization rates with industry averages and ISV-provided benchmarks. After your team has identified the

Similarly, you must evaluate the demand placed on various servers at different times of the day. Your Your team needs to combine virtualized servers in a way that balances load during the day, as shown in Figure 3.

applications that have the largest potential for improvement, you can focus tuning efforts on those applications.

In principle, decisions like these can be made by an IT analyst studying inventory and utilization data. In practice, however, however, the mathematical complexity becomes unmanageable in any but the smallest environments environments..

Getting the most efficient utilization in your current environment is the first step. Next, you must decide which servers can be virtualized and run on the same system. This involves careful consideration of such factors as: • Which workloads can be combined based upon the predicted predicted total resource utilization, a combination of CPU, memory, storage, and network. This depends on the resource utilization of each original server as well as thresholds for usage based upon factors such as peak load. l oad. • Which servers are too obsolete obsolete to maintain, and and what server hardware can be purchased in order to achieve a net consolidation gain. For example, by retiring 10 old servers and virtualizing their workloads to run on one new server, the capital purchase cost could be more than offset by savings in power,, cooling, rack space, and power a nd network ports. When you virtualize servers, you must also determine which workloads to consolidate onto a particular host system. Some customers are tempted to stack multiple virtualized servers ser vers running the same application into a single ESX Server system.  This approach limits the consolidation consolidation opportunity, because like applications compete for the same resources. What you want your project team to do is determine what resources each application requires, then match applications that demand different resource allocations to maximize your virtualization opportunity , as shown in Figure 2.

Clearly, a tool that can consider all these factors must be one that is uniquely designed to embody all the principles of capacity planning. While systems management tools from Hewlett Packard, Mercury, IBM, and others are the workhorses that keep systems running, and while these tools will continue to evolve, Capacity Planner provides a complementary software tool to answer the core questions you need to consider for a consolidation project.

Capacity Planner VMware Capacity Planner is an enterprise e nterprise IT capacity planning solution designed specifically to collect and analyze the data you need to plan an effective server consolidation project. It enables faster, more accurate, benchmarked infrastructure assessments and provides project teams the integrated set of analysis, planning, decision support, and monitoring functionality they need to enable and accelerate server consolidation and capacity optimization projects. At the heart of VMware Capacity Planner is a unique Information Warehouse, Warehouse, which houses a constantly growing

B AD B AD 50%

50%

     U     y     r      k     s      P     o      i      C     m      D     e      M

     k     r     o

    w     t     e      N

     U     y     r      k     s      P     o      i      C     m      D     e      M

     k     r     o

    w     t     e      N

     U     y     r      k     s      P     o      i      C     m      D     e      M

     k     r     o

    w     t     e      N

     U     y     r      k     s      P     o      i      C     m      D     e      M

     M      A      1

     k     r     o

     M      P      1

     M      A      1

     M      A      1

     M      P      1

     M      A      1

     M      A      1

     M      P      1

     M      A      1

    w     t     e      N

BETTER BETTER

50%

50%

     M      A      1

     U     y     r      k      P     o      i     s      C     m      D     e      M

     k     r     o

    w     t     e      N

     U     y     r      k      P     o      i     s      C     m      D     e      M

     k     r     o

    w     t     e      N

     U     y     r      k      P     o      i     s      C     m      D     e      M

     k     r     o

    w     t     e      N

     U     y     r      k      P     o      i     s      C     m      D     e      M

     M      P      1

     M      A      1

     M      A      1

     M      P      1

     M      A      1

     M      A      1

     M      P      1

     M      A      1

     k     r     o

    w     t     e      N

Figure 3: Optimal consolidation based on resource utilization over time

Figure 2; Optimal consolidation based on resource type utilization

5

 

VMWARE   W H I T E

PAPER

set of industry reference data that is used to drive intelligent, benchmarked IT capacity recommendations for the enterprise. VMware Capacity Planner delivers compelling value to project teams and the organizations they work for by: by : • Automating and streamlining the capacity planning cycle • Enabling accelerated, more more accurate, benchmarked infrastructure assessments • Guiding intelligent, objective capacity planning and server consolidation recommendations • Driving increased productivity, productivity, reduced complexity, and improved predictability of IT infrastructure With Capacity Planner, your project team will be able to answer questions they might not be able to answer with certainty today, such as: • How many servers do we have?

be evaluated in the decision phase.

Model  During the modeling phase, Capacity Planner enables your planning team to evaluate the various alternatives generated in the planning phase. Your team can use scenario modeling to test alternatives including purchase planning, virtualization, and server consolidation. They can also conduct what-if analyses for consolidation based on different groupings, thresholds, target servers, and other factors. Your Your team can recommend the alternative that best meets the success criteria and represents quick win, high-profile consolidation-success opportunities to build business unit support. Capacity Planner then enables your team to clearly articulate different scenario outcomes and present recommendations in their assessment reports and project pro-

• How well are are we utilizing utilizing them?

posals.

• What is the the right amount amount of capacity?

Monitor 

• What are the options options and what is the best recommendarecommendation? • How else can we better optimize optimize our environment? environment? Capacity Planner provides the tools to assess your needs, plan for and decide on consolidation strategies, and monitor your infrastructure to maintain its efficiency.

 Anal yze Capacity Planner enables e nables your project team — consultants, staff, or a combination of the two — to conduct comprehensive assessments of your existing e xisting infrastructure to assess how much IT capacity currently exists and how well this capacity is being utilized. Agent-less implementation ensures discovery and inventory of all hardware and software assets, providing a complete view of the IT infrastructure. Capacity Planner then correlates key performance metrics with inventory data to generate ser ver load profiles. This information enables your team to analyze and evaluate how well current capacities are being utilized.

Plan

6

purchases, disposing of old servers, migrating applications, redeploying existing assets, or implementing a virtualization solution. And Capacity Planner provides the ability to set success criteria, constraints, and thresholds against which these alternatives can

Capacity Planner helps you continuously compare resource utilization against benchmarked thresholds to ensure e nsure ongoing capacity optimization. Your Your planning team can provide periodic reports that monitor current capacity utilization and compare them to industry benchmarks in order to detect anomalies in utilization. Automated alerts and monitoring capabilities enable your team to detect deviations in utilization trends, predict capacity problems, and make timely troubleshooting and optimization recommendations. They They can continue to take advantage of Capacity Planner capabilities to ensure that you are able to manage unexpected or planned changes in capacity requirements and utilization over time.

How Capacity Planner Pl anner Works VMware Capacity Planner is a Web-based capacity planning application that combines inventory and utilization data. It gathers the data without any need for agents installed on the target systems. It then generates optimization recommendations, enables modeling for server consolidation and virtualization projects, and provides ongoing capacity planning decision support.

Capacity Planner enables your team to develop an effective capacity optimization plan so you can determine exactly how much IT capacity you really need, considering current and future business needs. Capacity Planner enables your team to analyze capacity utilization metrics, predict

 The key components of Capacity Planner are the Data Collector,, the Data Manager, the Data Analyzer, the Information Collector Warehouse, and the Capacity Planner Dashboard. Da shboard.

capacity needs, forecast utilization trends, and compare your data against industry benchmarks. Your Your team can also identify opportunities and alternatives for capacity optimization, whether they involve hardware refresh or new

 The Data Collector discovers and inventories inventories information from all of the computers on your network or from only the subset that you want to evaluate. It uses operating system APIs to communicate with all targeted servers ser vers in order to collect the information

Data Collector 

 

VMWARE   W H I T E

required for analysis. No software or agent is ever installed on any server. The Data Collector runs on a single workstation, and all collections are executed using remote procedure calls. Network traffic impact is minimal, and the CPU load on the server targeted for data collection is less than 1 percent. This minimal impact ensures completely unobtrusive visibility into computing infrastructure.

Data Manager   The Data Manager provides an organized view of the collected information and administrative control for the Data Collector Collector..  This includes detailed and summary views and reports on all discovered objects, collected inventory, and monitored performance information. The Data Manager also includes data synchronization capability that is used to send sterilized, anonymous data in the form of CSV files over Secure HTTP (HTTPS) to the Information Warehouse.

Information Warehouse

PAPER

to clearly visualize and gain insight into infrastructure. Reports provided with the Dashboard range from inventory, performance, and utilization reports to advanced analyses that enable project teams to run what-if scenarios and generate intelligent capacity optimization recommendations. recommendations. In addition, users have the ability to create and save their own custom reports. Users can manipulate data, drill down from summary to detailed data, apply filters, and sort data to gain different analytical views of the information. Security and access control features ensure secure, authenticated, and authorized access to and protection of information from particular customer sites. Users can also continue to monitor data and provide periodic reports and capacity optimization recommendation recommendations. s.

Rigorous Information Collection  The Data Collector component of Capacity Planner installs installs on a single workstation or server at a customer's site. Without the need to install any agents on target machines, it discovers servers and desktops by name and operating system within

 The Information Warehouse Warehouse is a central data warehouse hosted in a remote, secure location in the United States where collected client data is sent and stored. A data upload utility parses the CSV files, then scrubs and processes this data before loading it into the Information Warehouse. The Information Warehouse Warehouse also includes industry benchmark and research data derived from data collected from hundreds of Capacity Planner client sites. This data is not client-specific but instead represents valuable industry averages such as industry performance averages for different types of servers and maximum observed values or thresholds on server resources.

minutes. Over the next few hours, Capacity Planner collects a detailed inventory of all servers and desktops.

Data Analyzer 

Capacity Planner collects data every hour and calculates peak hour utilization for each one-hour increment in the 24-hour day. After several weeks, it identifies utilization for the busiest hour

 The Data Analyzer serves as the core analytical engine that processes all the analysis required for intelligent capacity planning. It includes advanced algorithms that solve capacity optimization problems and supports analysis capabilities such as trending, regressions, scenario modeling, anomaly detection and alerts. The Data Analyzer combines inventory and performance data to develop server load profiles and calculate insightful utilization metrics. It also aggregates data from different collectors across client installations to prepare industry reference metrics (such as averages and ratios) that then serve as benchmarks. Alerting capabilities enable users to define thresholds and set alerts aler ts to monitor any data within the Information Warehouse.

Capacity Planner D ashboard   The Dashboard is a Web-based, hosted hosted application that delivers a rich set of IT infrastructure analysis and capacity planning capabilities to users through a zero-footprint browser interface that requires no proprietary software installations or downloads.  The interface provides a rich set of analyses that enable you

Capacity Planner collects from targeted servers more than 40 core performance statistics plus additional relevant statistics for specific applications. This low-overhead query collects performance metrics from the four main data groups of processor, memory, storage, and network utilization. For memory, for example, Capacity Planner collects not only paging data or what is available in bytes but also specific cache information that affects the overall project decision strategy. This data is then correlated with the previously collected inventory data.

in the week. Capacity Planner also maintains weekly summary statistics that track maximum observed, minimum observed, average, hourly, prime time, nonprime time, and weekend loads. In addition, Capacity Planner maintains a summary for the most recent four weeks of performance per formance statistics on these same criteria. The summary is used to determine peak load for consolidation recommendations. recommendations. Peak load is determined by finding the hour of the day with the highest sustained load. It is not a measure of maximum observed values. After four weeks of per formance collection, trend lines become valuable and performance per formance projections can be made. These projections are very important. For example, you do not want to consolidate multiple loads into a single physical server ser ver only to run out of capacity in a month.

Intelligent Capacity Analysis Capacity Planner tracks and maintains current software vendor benchmarks for servers in its Information Warehouse and 7

 

VMWARE   W H I T E

PAPER

produces an alerts table showing servers ser vers in a customer's organization for which performance data deviates from these vendor values. Capacity Planner tracks and maintains the industry averages for software performance in its Information Warehouse and produces an anomalies table showing the servers in a customer's organization for which performance data deviates de viates from these average values. The industry averages are based on the most recent four weeks of data collected from all Capacity Planner sites. For more information on virtualization and capacity planning, visit www.vmware.com www.vmware.com..

8

 

Item: WP-002-PRD-01-01 Revision: 20060825

VMware VMw are,, Inc Inc.. 3145Port 3145PorterDrive erDrive Pal Palo o Alt Alto o CA 9430 943044 USATe USATell 650650-475475-5000Fax 5000Fax 650650-475475-5001www.vm 5001www.vmwar ware.c e.com om © 2006 VMware, Inc. All rights reserved. Protected by one or more of U.S. Patent Nos. 6,397,242, 6,496,847, 6,704,925, 6,711,672, 6,725,289, 6,735,601, 6,785,886, 6,789,156 and 6,795,966; patents pending. VMware, the VMware “boxes” logo and design, Virtual SMP and VMotion are registered trademarks or trademarks of VMware, Inc. in the United States and/or other jurisdictions. Microsoft, Windows and Windows NT are registered trademarks of Microsoft Corporation. Linux is a registere regis tered d trade trademark mark of Linus Torvalds Torvalds.. All othe otherr marks and names mentioned mentioned herein may be trademarks of their respective companies.