Consolidation of servers in data centers through virtualization has clear benefits to companies but the benefits are overstated by vendors promoting consolidation technology to customers, says an Impact Research Report from Info-Tech Research Group.
In their pitches to customers, vendors tend to exaggerate the number of servers that can actually be consolidated which can be misleading for enterprises looking to invest in this technology, said Matt Brudzynski, senior research analyst at Info-Tech Research Group. Vendors are ing high consolidation ratios of 12 or more virtual machines per processor, while in reality only about half that, six virtual machines per processor is the common average at this time. Virtualization is a technique that fools servers into thinking they are unique devices when in fact they become a shared resource, reducing the number of physical boxes needed in the data center. Info-Tech s Impact Research report, The Business Case in Virtualization, provides IT managers with the information they need to make a business case for purchasing decisions. The high consolidation rates touted by vendors aren t actually happening because customers need to leave adequate capacity on existing host servers for load-balancing, emergency backup and ensuring adequate data throughput to disk storage, Brudzynski said. The leading providers of X.86 server consolidation technology are VMWare and Microsoft Corp. The industry is moving closer to the target of higher consolidation rates with introduction of quad-core processors that are at the early stages of deployment in the marketplace, said Brudzynski. These processors provide greater throughput to disk storage and will help vendors more effectively deliver on consolidation promises. Prior to implementing virtualization, IT managers should do thorough analysis of server utilization data over the past year, as well as workload forecasting and future performance modeling to ensure the right technology choices are made to achieve desired consolidation, Info-Tech advises. Source: www.datawarehouse.coma>