Bladewatch.com

http://www.pcworld.com/businesscenter/article/147748/seven_tips_for_managing_a_multiplevendor_virtualization_environment.html

Fortisphere CTO John Suit meets with a lot of companies who are trying to get a better grip on managing virtualization. In the field lately, he sees the results of the virtualization downside that IT leaders have been telling us about for months: VMs roll out awfully quickly. This speed is a blessing and a curse, since each VM you create is another one you must track, manage and secure. Unfortunately, it becomes even more complicated to keep track of VM sprawl and related worries when you start playing with more than one vendor’s virtualization technology.

The challenges around virtual server management can be just as significant in many ways, particularly if you consider management from an administration standpoint: licensing, monitoring, security patching and billing.

We also have to consider the marginal cost element, that an extra virtual machine might require additional infrastructure, understanding what we have, what systems/workload is being hosted on which hardware becomes key. Understanding workloads against capacity, thinking about pre-provisioning can be something to consider, that the capacity should grow with demand, that as blade enclosure 7 or the last few rack servers start getting used to host virtual machines, IT should be provisioning or ordering extra capacity.

As with your physical estate though, we need not only someone to manage these machines, but to manage the estate:

• What servers do we have – are they still in use, can they be decommissioned
• What systems/applications have similar workload profiles – could we host them on the same infrastructure
• Can we consolidate server roles – can we not have one large virtual/physical machine running SQL than 6 virtual machines
• What level of patches/hot fixes and layered component levels are we at – to keep everything in line with vendor requirements
• Can we make small changes to improve performance/reliability?
• Reporting on system utilization/use – does that virtual machine need that much disk space/memory/cpu? Does it justify a virtual machine or could we share load with another virtual machine – shared infrastructure

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.