memory-sizing-for-server-virtualization

86 mb the average maximum memory consumption for

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 6 GB Hosts 32 Lower 75% 24 We calculated the average potential consolidation level among the lowest 75 percent of servers by maximum memory consumption, 44 Lower 50% and among the lowest 50 percent. Figure 4 also shows the data for 33 22 the lowest quartile for comparison. These less-utilized servers are obvious consolidation candidates 55 Lower 25% 41 because of the potential efficiencies. We estimated the following 27 10 20 30 potential consolidation levels, based on memory requirements only: 40 50 Potential Number of Consolidated Servers Figure 4. Potential consolidation levels for less-utilized pools of servers, based solely on memory constraints. We calculated average consolidation ratios for the lowest 75 percent, lowest 50 percent, and lowest 25 percent of servers by maximum memory consumption.  We then calculated potential consolidation levels among larger pools of the servers in the lower three quartiles, as shown in Figure 4. 16 0 this quartile. This resulted in a ratio of approximately 18 to 1. 60 • Approximately 25 to 1 for the lowest 25 percent of our server sample • Approximately 20 to 1 for the lowest 50 percent • Approximately 15 to 1 for the lowest 75 percent Memory Sizing for Server Virtualization White Paper Consolidation Strategy Based on our analysis, we believe we can achieve very good consolidation ratios using cost-effective dual-socket virtualization hosts configured with Quad-Core Intel Xeon processors and 16 GB of memory. In the near term, we are aiming to consolidate that we will be able to achieve the anticipated servers similar to those in the lower two quartiles consolidation ratios. of our 3,000-server sample. When we virtualize these workloads, we anticipate that each VM will require between 0.6 and 1.5 GB of memory; many will require 1 GB or less. Over time, we expect to fine-tune our consolidation strategy. For example, we may choose to consolidate several server workloads of the same type, such as databases, into a By specifying servers with 16 GB memory, we single larger workload and then virtualize the expect to be able to achieve high consolidation consolidated workload. This approach would levels of up 15-20 to 1. reduce the number of VMs in our environment, We are in the early sta...
View Full Document

This document was uploaded on 01/20/2014.

Ask a homework question - tutors are online