How to virtualize more?

Taking consolidation to the next level

Server virtualization – it’s no longer a question of whether it is the right choice for your organization; if you’re like most companies, that decision was made some time ago.  The question is how are you going to virtualize more of your computing resources to take advantage of even greater efficiencies, higher availability, reduced deployment times and cost savings.

Today it’s estimated that only 60% of applications are virtualized.  Many organizations are still experiencing impediments to full adoption based on insufficient virtualization & application expertise, inadequate computing infrastructures and budget.

DPSciences helps our clients accelerate their path to virtualization based upon the following design considerations:

  • Go big on memory – with virtualization, memory is king.  Less memory means less performance in a virtualized environment and less applications per physical server.
  • Right size your storage I/O – it seems like a simple concept but often virtualization creates performance problems due to I/O constraints.   When 1 physical server = 1 application there were enough paths to the data.  When 1 physical server = lots of applications, you need a lot more paths to spinning or SSD storage resources.
  • Check your network bandwidth – similar to the issue of getting to physical storage, having enough interfaces from the server to the network is a primary design consideration.   With 1 application per server, 1 NIC was sufficient.  Today with many applications per server, many additional NICs per server are required.
  • Make sure your virtual workloads are compatible – Certain applications don’t play well together on the same physical server.  For example, one may use hyper-threading, the other not.  The more flexibility you have to quickly adjust physical server capacity in the same way you adjust virtual server capacity is the key.

And finally build upon the decision that you’ve already made to virtualize by optimizing your physical computing architecture with Cisco’s Unified Computing System (UCS).

Cisco UCS servers are purpose built for virtualization and offer the most efficient way to address the design considerations discussed above with advanced memory capabilities, virtual storage and network I/O and the ability to add or adjust physical server capacity to handle unique configuration requirements and to more quickly respond to new business requirements.

Take VMware to the next level