In a new study on virtualization, Crucial asked over 350 IT managers about memory use per virtualized application and discovered a surprising finding. While IT pros noted clear benefits of more memory – more VMs per physical server, better quality of service (QoS), reduced disk thrash and improved ability to handle unpredictable workload demands – only 46% of IT pros said the amount of memory installed in their servers was sufficient to meet workload needs.
The pattern held across nearly all industries and countries surveyed, and at companies of all sizes that had anywhere from 30 physical servers to more than 200. Here are the survey’s findings and how you can use them to improve your data center’s virtualized application performance.
On average, IT managers are running about 29 VMs per physical server
Over 60% of IT managers said the amount of memory they allocated per VM was “very” or “extremely” important – for nearly every major virtualized server app
Nearly a third (32%) of IT pros said unexpected/unpredictable workload demands are a challenge their organization currently faces
Only 46% of IT pros say the amount of memory they have installed is sufficient for their current workload needs
Nearly half of all IT managers (47%) plan to upgrade their RAM within the next year
There’s no doubt that virtualized workloads will grow. The only question is ... by how much? And in what ways? Only you know enough about your company to accurately estimate how much DRAM you’ll need. However, one thing is certain: you’ll likely need more than you’re using today, since tomorrow’s bigger workloads will require more memory.
Server memory is a long-term investment that’s never just about today’s workload. It’s also about foreseeable future workloads because memory is often the limiting factor. Get as much as possible out of your IT budget by identifying high-growth, business-critical applications and then installing future-proofed 32GB modules to deliver optimal QoS and scalability at a typically lower price than that of multiple lower-density modules.