Optimizing Virtualization
to Deliver Best-in-Class Outcomes
An ENTERPRISE MANAGEMENT ASSOCIATES® (EMA™) White Paper
Prepared for VKernel
July 2009
IT MANAGEMENT RESEARCH,
INDUSTRY ANALYSIS AND CONSULTING

IT MANAGEMENT RESEARCH,
INDUSTRY ANALYSIS AND CONSULTING
Table of Contents
Challenges in Achieving Virtualization Goals ................................................................................................1
Optimizing Virtualization with Capacity Management.................................................................................1
Achieving Best-In-Class Outcomes with Capacity Management ...............................................................2
Virtual Machines per Server.........................................................................................................................2
Physical Resource Utilization .......................................................................................................................3
SLA/Uptime Achievement ..........................................................................................................................3
EMA Perspective ..................................................................................................................................................4
About VKernel .....................................................................................................................................................4
Optimizing Virtualization to Deliver Best-in-Class Outcomes
©2009 Enterprise Management Associates, Inc. All Rights Reserved.

IT MANAGEMENT RESEARCH,
INDUSTRY ANALYSIS AND CONSULTING
Challenges in Achieving Virtualization Goals
ENTERPRISE MANAGEMENT ASSOCIATES® (EMA™) research1 has consistently high-
lighted several key drivers for virtualization deployment. For example, in the most recent study,
99% of enterprises cited server consolidation as either critical or important, 97% cited hardware
cost reduction, and 92% cited meeting service level agreements (SLAs).
These are certainly realistic and achievable goals, but achieving them means overcoming specific
and significant challenges, especially:
• Getting ‘over the hump’ – a majority of enterprises typically find their virtualization
deployments stalling at around 25% of their environment, and as a result find that these
goals are only partially fulfilled.
• Improving consolidation ratios – EMA research shows that virtualization in the average
enterprise reduces physical servers by just 17%, with a large number only achieving a 10%
reduction.
• Improving resource utilization – the key to maximizing virtualization value is to maximize
granular resource utilization, yet the average CPU, NIC, and memory utilization is typically
only near half that of the best performers.
• Maintaining service levels – squeezing resource utilization is a key goal, but is only useful if
IT can guarantee that service, performance, and availability are not damaged in the process.
Virtualization success is not just about density, it is also about service.
Optimizing Virtualization with Capacity Management
Capacity Management is an IT management discipline that optimizes resource utilization while
maintaining performance. By monitoring and adjusting the utilization of granular resources
– including memory, CPU, I/O, storage, etc. – it ensures that IT
resources are always sufficient to service business needs, without
Capacity Management optimizes wasting otherwise idle resources.
resource utilization while In a virtual environment, Capacity Management allows additional
maintaining performance. VMs to be deployed (and migrated) using spare capacity on
existing servers, without impacting system performance. By con-
stantly tracking available server capacity, matching that capacity to
dynamic workloads in real time, and predicting capacity requirements for future workloads, IT can
better maximize resource utilization while still meeting SLAs.
It is important to note a small but significant difference between two related terms:
• Capacity Planning provides an ad-hoc or snapshot approach that looks forward to establish
a resource allocation plan (such as for planned P2V/V2V migration, or quarterly capacity
reviews).
• Capacity Management provides ongoing, real-time monitoring and adjustment of resource
utilization, aimed at continuous optimization of resource availability and service.
1 EMA Research Report, Virtualization and Management: Trends, Forecasts, and Recommendations,
http://www.enterprisemanagement.com/research/asset.php?id=721
Optimizing Virtualization to Deliver Best-in-Class Outcomes
©2009 Enterprise Management Associates, Inc. All Rights Reserved. Page 

IT MANAGEMENT RESEARCH,
INDUSTRY ANALYSIS AND CONSULTING
While these terms are used more or less interchangeably in many circumstances, and both pro-
cesses are valuable, the difference is important – even more so in the highly dynamic area of
server virtualization.
Achieving Best-In-Class Outcomes with Capacity
Management
Recent EMA research2 has shown that Capacity Management correlates with best-in-class out-
comes in several areas that directly contribute to server consolidation, hardware cost reduction,
and meeting SLAs, including:
Virtual Machines per Server
Average 6
Best Practice 15
0 5 10 15 20
Figure  – Average vs. Best Practice Outcomes for Virtual Machines per Physical Server
EMA research found that the average enterprise is running around 6 virtual machines on each
physical server. Best performers in this metric, however, are able to run around 15 VMs per server
– between two to three times the average server consolidation rate. Below average performers, by
contrast, run just one or two VMs per server.
This research also found that the best performers in this metric
were significantly more likely to use Capacity Management soft-
Best performers in VM-to-
ware with specific support for virtual systems.
Server ratio are significantly
more likely to use Capacity Of course, many other factors will affect this metric – such as
Management software. the size of the physical servers and the types of workloads in use.
However, Capacity Management is likely to help any enterprise
move from below average or average to best-in-class in VM den-
sity. Understanding actual utilization of resources allows more additional workloads to be loaded
onto fewer existing servers without damaging performance, driving higher utilization, and requir-
ing fewer servers for the same workload.
2 EMA Research Report, Best Practices in Virtual Systems Management (VSM): Virtualization Metrics and Recommendations for
Enterprises, http://www.enterprisemanagement.com/research/asset.php?id=1104
Optimizing Virtualization to Deliver Best-in-Class Outcomes
©2009 Enterprise Management Associates, Inc. All Rights Reserved. Page 

IT MANAGEMENT RESEARCH,
INDUSTRY ANALYSIS AND CONSULTING
Physical Resource Utilization
65%
Average
45%
85%
Best Practice
75%
0% 20% 40% 60% 80% 100%
Memory CPU
Figure  – Average vs. Best Practice Outcomes for Physical Resource Utilization
In the same study, EMA found the average enterprises reported physical CPU utilization of 40-
50% and memory utilization of 60-70%. By contrast, the best performers reported an average
physical CPU utilization of 70% or higher, and average memory utilization of 80% or higher. This
would allow best performers in these metrics to use almost half as many physical servers for the
same virtual workload compared to the average performers.
Again, best performers in these metrics were significantly more
Best performers use almost likely to use Capacity Management software.
half as many physical
servers compared to Capacity Management is specifically designed to give visibility into
utilization of these granular resources, allowing IT to add more
average performers.
workloads, and to maximize utilization of these resources without
damaging performance.
Importantly, capacity management will also make sure these resources are not completely satu-
rated, by detecting and even predicting workload fluctuations, to ensure there is always sufficient
headroom to accommodate these peak loads. This is critical to improving server utilization while
meeting SLAs.
SLA/Uptime Achievement
Average 99.5%
Best Practice 99.999%
99.2% 99.4% 99.6% 99.8% 100.0% 100.2%
Uptime
Figure  – Average vs. Best Practice Outcomes for SLA/Uptime Achievement
Optimizing Virtualization to Deliver Best-in-Class Outcomes
©2009 Enterprise Management Associates, Inc. All Rights Reserved. Page 

IT MANAGEMENT RESEARCH,
INDUSTRY ANALYSIS AND CONSULTING
While not the only component of an SLA, uptime is often an SLA’s most important measure. Best
performers in this category were able to keep systems up at least 99.999% of agreed service levels
– less than one second of unplanned downtime each day. Average performers, however, achieved
less than 99.5% uptime, allowing more than 7 minutes a day of unplanned system downtime
– more downtime every day than the best performers allowed all year.
Capacity Management again correlates significantly with improvement from below average or aver-
age to best-in-class in this metric.
Although many factors contribute to achieving SLAs, the importance of Capacity Management
is easily explained. With limited resources, it is easy to either increase density, or maintain service
levels, but it is hard to do both at once. Capacity management provides visibility into workloads to
ensure increasing workloads do not overload physical resource capacity, allowing both increased
VM density and maintained SLAs.
EMA Perspective
Every IT administrator does capacity management to some extent, but spreadsheets and guesswork
are a path to failure. They invariably lead to resource saturation, application slowdown, and failure
to meet service levels. The alternative – over-provisioning servers to ensure resource availability
– leads to high IT costs and failure to meet virtualization objectives.
To optimize resource utilization while maintaining SLAs, it is important to get real visibility into
resource utilization, to match it with current and future workload requirements. This is especially
critical in a virtual environment, where new workloads are easily deployed, existing workloads are
rapidly migrated, and physical resources are in ever-increasing demand.
Capacity Management provides this visibility – both real-time and predictive – by monitoring,
measuring, and managing resource utilization.
In the first instance, Capacity Planning starts even before virtualization does, ensuring optimal
migration from physical to virtual environments. Capacity Management then provides ongoing value
through continuous optimization of physical and virtual resources, and especially when deploying
more advanced virtualization techniques like load balancing and
live migration.
Capacity Management is The result of good Capacity Management is shown in EMA
an important discipline to research – greater VM density, higher resource utilization,
ensuring achievement of and achieving SLAs. This is why EMA believes that Capacity
key virtualization goals. Management is an important discipline to ensuring achievement
of key virtualization goals.
About VKernel
VKernel Corporation is an award-winning innovator delivering a truly unique and differentiated
management and monitoring approach for dynamic virtual data centers. As an alternative to com-
plex and expensive systems management vendors, VKernel’s “plug-and-play” Virtual Management
Appliances (VMAs) are affordable and simple to use to maintain the health of the virtual data
center, optimize resource allocations, lower and control costs, and achieve an instant ROI. For
more information, visit www.vkernel.com
Optimizing Virtualization to Deliver Best-in-Class Outcomes
©2009 Enterprise Management Associates, Inc. All Rights Reserved. Page 

About Enterprise Management Associates, Inc.
Founded in 1996, Enterprise Management Associates (EMA) is a leading industry analyst firm that specializes in going “beyond the surface” to provide deep
insight across the full spectrum of IT management technologies. EMA analysts leverage a unique combination of practical experience, insight into industry
best practices, and in-depth knowledge of current and planned vendor solutions to help its clients achieve their goals. Learn more about EMA research,
analysis, and consulting services for enterprise IT professionals and IT vendors at www.enterprisemanagement.com or follow EMA on Twitter.
This report in whole or in part may not be duplicated, reproduced, stored in a retrieval system or retransmitted without prior written permission of
Enterprise Management Associates, Inc. All opinions and estimates herein constitute our judgement as of this date and are subject to change without notice.
Product names mentioned herein may be trademarks and/or registered trademarks of their respective companies. “EMA” and “Enterprise Management
Associates” are trademarks of Enterprise Management Associates, Inc. in the United States and other countries.
©2009 Enterprise Management Associates, Inc. All Rights Reserved. EMA™, ENTERPRISE MANAGEMENT ASSOCIATES®, and the mobius
symbol are registered trademarks or common-law trademarks of Enterprise Management Associates, Inc.
Corporate Headquarters:
5777 Central Avenue, Suite 105
Boulder, CO 80301
Phone: +1 303.543.9500
Fax: +1 303.543.7687
www.enterprisemanagement.com 1920.072009