| COLUMN |
Initiating a server virtualisation project |
 |
 |
|
By David Boyd
06 Apr 2009 | SearchStorage.co.UK
|
|
There's no shortage of literature to describe the benefits of server virtualisation. There are countless stories of how companies of all sizes have benefited from virtualising their infrastructure. Virtualisation can increase the capacity of the data centre, reduce hardware support costs, bring electricity bills within digestible limits and even increase performance by removing aged architecture from your estate.
So you're sold on the idea, but where do you start? Initiating any project requires careful consideration. The first step is the business case: investigating the financial prudence of running a project. Even with all of its obvious benefits, virtualisation won't be the only and/or best answer in every case. If virtualising your estate means you don't have to purchase or build a new data centre, the projected savings will be substantial. However, if it's not clear-cut at this stage, estimate the number of potential virtualisation targets. In-house monitoring and reporting tools will come in very useful at this stage. If you only have half a dozen obvious candidates, your chances of seeing a return on your investment are quite low or nonexistent. Don't underestimate the costs of implementing your solution. New hardware will be needed, virtualisation and migration software will need to be purchased, and staff will need to be recruited or trained.
Once the business case has been made, senior management convinced and the capital has been found, it's time to get stuck in.
The introduction of server virtualisation isn't just a change in technology; it's a shift in culture. Your service offerings will need to be adapted to encompass a virtual environment, and you'll have to educate not just your technical teams but your customer base. Virtualisation isn't just a mechanism to drive costs out of the IT department; it can offer real benefits to users. However, it's not a cure-all. A new server can be deployed within minutes of an order being received, but this doesn't mean users should request a new host every second day. Addressing misconceptions and fears early will ease the entire migration process.
As with all technology implementations, the best solutions are the ones that have been comprehensively designed for the target environment and project goals. Having sold the project to senior management, there's likely to be pressure to get the project underway and migrations occurring as soon as possible, but the value of a considered design can't be underestimated. All good virtualisation designs have to be based upon thorough baselining of server performance over prolonged periods of time. Many servers will appear as good virtualisation targets if monitored for short periods, but many applications require considerable dedicated resources at certain times (e.g., month's end). When designing a target infrastructure, you need to keep in mind the number of host-level changes that will be required to facilitate migration. Keeping changes to a minimum will make life easier, but don't let it compromise your adherence to best practices.
The next big challenge is migrating your existing hosts from physical to virtual targets. There are an increasing number of products in the marketplace that will facilitate this process, but the downtime tolerance of your applications should ultimately dictate which method is used.
Virtualisation projects can be difficult and politically contentious. But if managed and designed with business needs and project goals in mind, virtualisation can deliver real benefits to the IT department and its customers.
About the author: David Boyd is a senior consultant at GlassHouse Technologies (U.K.), a global provider of IT infrastructure services. He has more than eight years of experience in backup and storage, with a major focus on designing and implementing backup solutions for blue chip companies.
|
|
|
|
