Here’s some more work for you. Here’s some more responsibility. Here’s a shorter deadline. Now do it all with less money, less time, less resources, less, less, less!
It seems as though the more efficient we become, the more constrained we are. The current economic climate doesn’t help either. Yes, this is the new norm. So what can you do?
If you’re reading this you’ve probably invested time and money into a virtual infrastructure, or are considering it. Great! Virtualized computing environments squeeze every last drop of performance from hardware and, when properly budgeted, can save you thousands in the long run. But don’t expect a free lunch.
Physical to Virtual Consolidation
Consolidation of physical servers to virtual hosts allows you to break the 1 application to 1 server mold. However the increased density in your server room might create hot spots, especially if you’ve decided on using a blade chassis.
That increased density means you’ll also be pushing your hardware harder. This will likely increase your power requirements, slightly. Newer hardware is indeed more efficient, and technologies like VMware’s DRS Distributed Power Management allow you to move workloads around to less stressed hosts and power off unused resources. The net effect is a possible overall reduction in power usage, but peak times could actually require more.
An Up Front Expense?
Virtualization is a net new expense. Unless you are starting from scratch, you will need to invest in hardware, and software licensing. I was recently asked to vet the cost of a 24 host, enterprise level virtual environment. Assuming a requirement of 10 Terabytes of storage, and going with mid tier hardware I came up with an up-front ballpark cost of USD $225,000. No small change. Amortize your projected savings carefully. Is it worth the up-front investment? Luckily you can grow your virtual environment easily as required with little to no negative impact on the existing services.
Implement Standards
Virtualization has made provisioning services a snap. You’ve heard all the marketing buzz — reduced time to market, provision servers in seconds!, etc. Suddenly that 10T of storage is GONE. But how?
Sprawl. (Yes, up.time can help you with this!)
Back in the days before virtualization, if you needed more resources you had to justify the expense nine ways from Sunday. When it finally arrived you’d spend a week staging it. Then testing and finally implementing it, only to have it completely consumed a few months later! When you planned your virtual infrastructure you WAY over provisioned it, didn’t you? You thought ahead 3 years like you did when you bought a single server for that one application. However now you’re planning for possibly hundreds of workloads. Need another machine? No problem, just clone it and wait a few minutes. Ever have cash burn a hole in your pocket? Budgets prevent us from blowing that spare cash. It’s exactly the same in a virtual environment, except the spare cash is extra CPU cycles, storage and memory. From simply devising a set of rules for managing the virtual machine life cycle, or implementing tools to manage it, the only way you will realize long-term savings is to ensure you’re only using what you need. Don’t run your VM environment like that TV salesperson’s famous oven — “Set it, and forget it!”.
If you keep these things in mind when building and managing your VMware vSphere environment, or any other virtual infrastructure, you will absolutely be able to do more, with less. Of course <shamesless plug>, up.time can solve you VMware monitoring needs with it’s deep VMware monitor and reports.
