Virtualize Servers

The Past: One Application per Server

Until fairly recently, data center operators typically installed just one application (or ”workload”) on each physical server. When taking into account development/testing environments and disaster recovery, a ratio of three to five physical servers per application was commonplace in data centers.1  Because of the “one workload, one box” approach, most servers run at a low "utilization rate" – meaning that just a fraction of their total computing resources are engaged in useful work. Even though the past decade has seen data center managers adopt server virtualization broadly, a 2014 study by NRDC found that average server utilization was still just between 12 and 18 percent.2


Figure 1: Virtualization

The Present: Many Applications per Server

A "virtual server" is a software simulation of a server and an operating system that executes programs just like a real server. Server virtualization offers a way to consolidate servers: it allows you to run multiple different workloads on one physical “host” server. Virtualization allows for fewer physical servers in a data center, with each remaining physical host server operating at higher total utilization. (See Figure 1, below.)  Virtualization saves energy because a virtualized data center needs fewer servers to accomplish the same amount work as a data center using the one workload, one box approach. Although six is the average, it’s not unheard of to consolidate the workloads of as many as ten separate physical servers (or more) on one physical host server, so the energy savings can be substantial.

Virtualization enables faster deployments, improves scalability, and reduces downtime. It speeds up disaster recovery efforts because virtual servers can restart applications much more rapidly than physical servers. With virtualization, you can move entire systems from one physical server to another in just a few seconds in order to optimize workload performance or to perform maintenance – all without causing downtime. Some virtualization solutions have built-in resiliency features, such as high availability, load balancing and failover capabilities.3

Big Opportunities in Small Data Centers

Thanks to compelling benefits, virtualization has become commonplace in large data centers. According to a 2016 Gartner report, on average most large enterprises report 75% or higher virtualization of their data center, "illustrating the high level of penetration," wrote Michael Warrilow, research director at Gartner.4

Virtualization is less pervasive in small data centers. A 2015 report by Techaisle found that server virtualization penetration had reached just 54% in US SMBs.5  This is mostly due to resource constraints: moving to a virtualized computing environment typically requires an upfront investment in software, hardware, and training. However, even the smallest data centers and server rooms are increasingly turning to virtualization to reduce operating costs, improve disaster recovery capabilities, reduce downtime, and simplify software development and testing.

Savings and Costs

Virtualization enables you to use fewer servers, thus directly decreasing electricity consumption. Reducing the number of servers in a data center also allows for a smaller power infrastructure.  As a result, less energy is consumed by power distribution units, UPS systems, and building transformers.   And because every server produces a lot of waste heat, fewer servers means that the data center needs less air conditioning. As a result of these indirect energy benefits, saving one watt-hour of electricity at the server level typically results in an additional 1.9 watt-hours of electricity savings at the facility-level!6

Note that some utility companies offer rebates to help offset the costs of virtualization projects. Please see the ENERGY STAR Utility Guide for Designing Incentive Programs Focused on Data Center Efficiency Measures for a (partial) list of utilities that offer rebates for energy-efficiency measures in the data center.

Sample savings associated with server virtualization:

  • Virtualization enables the repurposing and decommissioning (removal) of some number of existing servers.  According to the Uptime Institute, decommissioning a single 1U server (a standard-sized rack server) can save $500 annually in energy, not to mention $500 in operating system licenses, and $1,500 in hardware maintenance costs.7
  • In a 2010 study, Southwestern Illinois College performed a detailed 3-year total cost of ownership (“TCO”) analysis for a 35 server upgrade with and without virtualization (see Table 1 below).8  The scenario with 35 virtual servers running on four physical host servers saved over $280,000, comprised of:
    • Nearly $150,000 in direct cost savings (the cost of virtualization software and storage equipment was offset by substantial savings in server hardware and networking equipment);
    • More than $130,000 in indirect cost savings, including substantial electricity, cooling, server provisioning, and procurement savings.
  • A Cisco study found that a virtualized server costs about $2,000 to deploy (compared to $7,000 for a standard physical server with 2 CPUs) due to reduced labor and hardware costs.9

Virtualization Table

Table 1: Virtualization at Southwestern Illinois College: a 3-year Total Cost of Ownership Analysis

Tips and Considerations10

  1. Take inventory. It is easy to lose track of servers as time passes.  Mergers and acquisitions take place, employees come and go, projects are completed, and servers may not be retired or repurposed. The first step toward virtualization (and better management of server utilization generally) is to take inventory of computing resources and their associated application workloads.  Creating and regularly updating a server hardware and application inventory will help you track the applications running on each server. Mapping applications to the physical servers on which they are running helps identify unused servers and opportunities for consolidation.  There are a number of software tools that can help you keep track of data center assets, including a few open source options:
    1. OpenDCIM provides complete physical inventory (asset tracking) of the data center.  It is offered free under the GPL license;
    2. RackTables provides asset tracking for the data center.  It is offered free under the GPL license.
  2. Identify servers that should not be considered for virtualization. Servers that have privacy, security, or regulatory restrictions, or require ultra-high service levels, may not to be the best candidates for virtualization or consolidation.
  3. Group remaining servers by workloads. Group the server inventory into pools of workloads that can work off of the same physical server at the same time. For example, placing “innovation” or “development" workloads on the same server as a mission-critical "production" workload is not a good idea. As previously mentioned, some servers may not be able to share resources due to technical, security, privacy, or regulatory concerns. Answering the following questions allows you to pool workloads into groups and determine the computing resources required for each group:
    1. Who owns the server?
    2. What is the purpose of the server?
    3. What is the service level requirement?
    4. Where is the server located?
    5. Is there a requirement for this workload to remain on a dedicated physical resource?

For example, the exercise may yield a list of 100 "innovation" workloads that all reside in the same location. You can then determine the resource requirements for the entire group. In this example, you may just need 5 to 10 physical host servers to support the resource requirements of those 100 workloads.

  1. Consider shared (networked) storage.  A virtualized environment may require networked data storage, because physical host servers may need access to the same data at the same time.
  2. Keep in mind that application workload can depend on the time of day, week, month or longer if the business cycle dictates. For example, some applications are used primarily during business hours, while others peak during off hours. Servers should be sampled at least once per hour over a 30-day period to develop a complete workload profile. This technique ensures that peak- and sustained-load times are captured in addition to normal usage.
  3. Combine workloads that have complementary characteristics. For example: Workload A consumes 10% processor capacity during business hours and very little at night. Workload B consumes 10% processor capacity during the night and very little during business hours. Combining the two workloads delivers the best possible results.
  4. Keep an eye on cooling. Consolidating workloads onto a smaller physical footprint can create "hot spots" in your data center. You should monitor for possible hot spots and you may need to work with HVAC experts to ensure an adequate and efficient cooling configuration.
  5. Track all virtual machines (VMs). Easy to deploy, VMs can get out of hand if you are not careful. If new VMs are created for specific projects, be sure they are not forgotten once those projects are completed.

1 The Economics of Virtualization: Moving Toward an Application-Based Cost Model, IDC, 2009.

2 Data Center Efficiency Assessment, NRDC, August 2014, p. 13. (PDF, 485 KB)

3 The Economics of Virtualization: Moving Toward an Application-Based Cost Model, IDC, 2009.

4 Virtualization Market Now ‘Mature,’ Gartner Finds, InformationWeek, by Charles Babcock, May 16, 2016.

5 SMB Server Virtualization Penetration Is Increasing But Challenges Remain, Techaisle Blog, May 7, 2015.

6 New Strategies for Cutting Data Center Energy Cost and Boosting Capacity, Emerson Network Power presentation, 2012, p.8. (PDF,  1.4 MB)

7 Decommissioning as a Discipline: Server Roundup Winners Share Success, by Matt Stansberry, Uptime Institute, undated.

8 Implementing Server Virtualization at Southwestern Illinois College, by Christine Leja, 1/15/2010. (PDF, 162 KB)

9 How Cisco IT Virtualizes Data Center Application Servers, Cisco Systems, Inc., 2007. (PDF, 237 KB)

10 Section was developed using the following document as a guide: Using Virtualization to Improve Data Center Efficiency, Green Grid, 2009.