In our recent webinar “Max VM Density Requires Optimal Storage Networking & Operational Transparency” respondents to one of the polling questions indicated that over 40% of the time their servers were supporting less than a dozen virtual machines (VMs). This result is similar to what we have seen in the field; in most cases, data center administrators are reluctant to push their physical hosts to the limit, or anywhere close to it. Why? Most seem to be concerned that stacking too many VMs will lead to unpredictable performance of mission critical applications.
The problem with being overly conservative is that it wastes money and an opportunity. When physical hosts are NOT pushed closer to their potential VM maximum the organization loses an opportunity to increase the return on its virtualization investment. As we discuss in our white paper “Maximum VM Density Requires Optimal Storage Networking” the potential ROI of increasing VM density can be considerable.
Watch our on-demand webinar and download our white paper "Maximum VM Density Requires Optimal Storage Networking" which is exclusively available to on-demand viewers.
The modern physical host has the memory and CPU power to support dozens of virtual machines. For the virtualization ROI to pay off they need to leverage that power and increase VM density, since these systems are often much more expensively configured than the traditional server designed to support a single application.
Not only does the physical host have the horsepower to support very high VM-to-host ratios, the storage technology can now support the random I/O storm that a dense VM infrastructure will certainly create. Flash, in the form of all-flash arrays or flash assisted hybrid arrays, can sustain the I/O storm, as can the storage network thanks to modern Gen 5 and soon Gen 6 architectures.
Safety is the Key
In the webinar we discuss how the performance foundation of high performance hosts, flash assisted storage and high performance networks now need to be complimented with controls so that mission critical applications can be safeguarded against “noisy neighbors“. This safety can be found in two areas; the network and the hypervisor.
In the webinar Andre Beausoleil, a Senior Product Manager and I discussed how modern storage networks can be optimized and controlled. This allows bandwidth to be allocated to VMs based on how critical their applications are.
Also joining us on the webinar was Bill Erdman, Product Line Manager within VMware’s Cloud Management Business Unit. He and I discussed how VMware’s vCenter Operations and Log Insight solution can work with the Brocade infrastructure and further fine tune the allocation of storage performance. It can also provide trending so that future performance conflicts can be averted before they start.
High VM- to-server ratios can be achieved safely, but it requires the right infrastructure and tools in order to provide the appropriate performance and insight. Our webinar and white paper give you the details needed to achieve new levels of return on your virtualization investment. You can get both by registering for our on-demand webinar “Max VM Density Requires Optimal Storage Networking & Operational Transparency“.
Brocade is a client of Storage Switzerland