While server virtualization has numerous benefits, server consolidation remains one of the most significant drivers for virtualization projects. Enterprise Management Associates (EMA) conducted recent research on virtualization trends and found that respondents consistently rated facilitating server consolidation as first or second on the list of reasons to adopt server virtualization. About 70% of 627 responding organizations rated server consolidation as a critical goal for virtualization.
As in other areas of conducting and managing server virtualization, virtual systems management (VSM) tools are essential to successful virtualization deployments. In this tip, we discuss how VSM tools can improve server consolidation ratios.
Making gains in server consolidation
EMA research on the best practices in virtual systems management (VSM) indicates major differences in server consolidation outcomes among the average and below-average performers, and the top 10% to 20% that make up the best performers. EMA surveyed 153 firms, and used several key performance indicators to measure real gains in server consolidation, including the following:
- the number of virtual machines (VMs) per physical server, i.e. VM density;
- the number of physical servers that have been retired because of virtualization projects;
- the percentage utilization of physical CPU resources;
- the percentage utilization of physical memory (RAM) resources; and
- the percentage utilization of physical network interfaces (NIC), i.e. bandwidth utilization.
@66395While respondents' average VM density was 6:1 (i.e. six VMs per server), for example, the best performers achieved rates of 15:1 or higher. The average for physical server retirement was about 17% of all servers, while the best performers accomplished more than twice that, at 35%. Physical resource utilization varies, but where average enterprises reported average physical CPU utilization of 45%, memory utilization of 65% and network utilization of 35%, the best performers reported an average physical CPU utilization of 75%, an average memory utilization of 85% and an average network utilization of 75%. Additionally, more than 70% of responding organizations rated conserving data center floor space as critical or important. About half of responding organizations reported conserving space as an effective outcome of their virtualization deployment.
VSM best practices for server consolidation
The key VSM disciplines that reduced server and data center use included the following tactics:
- Configuration management. Tools in this category (such as EMC Ionix Server Configuration Manager, ManageIQ EVM Insight or Veeam Configurator) enable organizations to better detect, track and control virtual resource allocations -- and thus maximize virtual and physical resource utilization. As a result, the best performers in physical resource utilization were up to 41% more likely to have these tools than were average or below-average performers. The best performers in floor space reduction were 13% more likely to have these tools.
- Virtual machine management. VM management involves the ability to dynamically manage VM allocations with automated tools, such as VMware vCenter, Microsoft System Center Virtual Machine Management (VMM) or Citrix Systems Inc.'s Essentials, allows IT to quickly react to available or overused resources while maintaining high server utilization. The best performers in VM density were up 39% more likely to have VMM tools than were average or below average performers.
- Workload automation and job scheduling. Batch processing is still a major part of many workloads, and automation that can dynamically balance resource allocation between batch and online/transaction workloads (such as MBMC Software's CONTROL, IBM Tivoli Workload Scheduler and ASG Distributed Workload Management) effectively allows servers to pull double duty. Organizations with workload automation tools were 28% more likely to be the best performers in VM density, and up to 16% more likely to be the best performers in the physical resource utilization category.
In these metrics, is not just virtualization or management that makes a difference. Recent EMA research, for example, indicates that intelligent, centralized storage systems – such as the NetApp V-Series – can take over many storage-related activities (e.g., storage provisioning, backup migration and replication, image clone management, snapshot management, etc.) without using server resources. This can dramatically reduce server load, which in turn improves resource headroom, enabling a higher server density and improved consolidation ratios. Of course, better management techniques can't solve all your server consolidation problems. The number of cores, processor speeds, memory types and (perhaps most important) the type of workloads all affect how many VMs you can squeeze onto a single server.
Similarly, removing servers from a data center does not instantly shrink the physical floor space. A lot depends on the terms and conditions of facilities leases and the physical flexibility of the organization. Consider the following:
- Organizations that have flexible terms on their data center lease may be able to put up new walls and re-lease some of their regained space.
- Organizations with multiple sites (including multiple floors or buildings on the same site) may be able to re-allocate regained space to other departments, sell or re-lease the newfound space.
- Organizations without major penalties for lease termination may be able to pay out their data center lease and either shut down the empty data center space or relocate to a smaller facility.
But even if these organizations can achieve some server consolidation, many can't realize data center space reductions without these flexible facility terms. As a result, few organizations are able to achieve substantial consolidation rates. Nevertheless, when organizations apply these VSM best practices, it can result in measurable reductions in server and data center use, where best performers show clear and measurable advantages from their deployment of these tools and disciplines.
|Andi Mann, is a research director with the IT analyst firm Enterprise Management Associates (EMA). Mann has over 20 years of IT experience in both technical and management roles, working with enterprise systems and software on mainframes, midrange, servers, and desktops. Mann leads the EMA Systems Management research practice, with a personal focus on data center automation and virtualization. For more information, visit EMA's website.|