Are you on the pre-order list when the latest and greatest i-something is announced? Or, like me, did you only just get your first Android phone a few months ago because, well, the old rock solid reliable clamshell phone was still working just fine? Both strategies have their merits – the former capitalizing on the newest innovations while they’re still new, and the latter waiting for a more mature market that provides more perspective and better pricing.
At this point, virtualization has proven itself in practice and is well past the early adoption stage. If you’re building an IT infrastructure from scratch you’ve probably got some combination of local and remote IT components including cloud,virtualization, SaaS, PaaS, and IaaS. But – what if you’ve got an existing IT infrastructure that is working just fine? Why fix something that isn’t broken?
There are several factors that can drive a business with an existing IT infrastructure to migrate IT components to virtual servers:
1) Server replacement costs
Servers need to be replaced every few years, and that cost is already built into your IT budget. Migrating to a local virtual server (e.g. VMware) means fewer servers need to be replaced. Migrating to remotely hosted cloud based IAAS servers eliminates the need for replacing servers all together – but, for remotely hosted servers, you need to make sure you’ve got enough bandwidth on your WAN to access them.
2) Facilities savings
Replacing a room full of servers with two or three VMware hosts saves significantly on power and cooling requirements. Moving to cloud based servers means only having to maintain WAN connectivity and local client workstations.
3) Legacy applications
If you’ve been in business for a while, you may well have one or two applications that are infrequently used, can’t be replaced, and that are no longer updated by the developers. And, of course, they only run on older, unsupported OS versions. The servers hosting these applications may be long past their replacement dates, but you need the apps, so you patch the servers up as best you can and keep your fingers crossed. Admittedly, nowhere near a best practice, but sometimes you have to work with what you’ve got. Fortunately, VMware supports a pretty wide range of Guest OS versions, which can provide you with the opportunity to migrate your legacy application to a VMware virtual machine (VM) and retire the legacy server.
4) Hybrid configurations
Migrating your servers VMs does not need to be done all at once. Older or under-provisioned servers can be migrated into VMs with more resources, and servers hosting legacy applications can be migrated on to more stable hardware, while other servers can remain on physical boxes and be moved later as required.
One of the first steps when considering moving to VMs is implementing a server monitoring package that will create a baseline of current resource usage. This baseline will provide an estimate of the resources your VMs will need, and determine which physical servers are in need of additional resources, and would benefit the most from virtualization. Server monitoring after the physical servers have been migrated to VMs will guide you in fine tuning how resources are allocated among the VMs.
After the servers have been moved to VMs, you should also monitor VMware itself. VMware performance monitoring can provide alerts for VMware resource problems at both the host machine and VM level, and VMware capacity planning can help you distribute your VMs so that you get the most out of the VMware host resources.