Revera is one of this countrys largest users of virtualisation for its hosting and storage services, having operated since the early 2000s. MD Roger Cockayne has worked with virtualisation as a discipline for more than 20 years.
The idea existed with the mainframe, partitioning machines into virtual segments, which people could leverage independently as if they were independent machines themselves. Then network virtualisation emerged the ability to take a physical network and hardware and split it up into virtual, logical ones.
Then storage was virtualised, so you could connect a virtual service across a number of different storage appliances, but the application connecting has no idea, presenting capacity from anywhere. It was on these principles we built our datacentres, he says.
Local firms were too small and poor to provide the hardware themselves, so Revera uses virtualisation to offer storage they can use.
Server virtualisation has been the final piece. In the past we had physical servers connecting to virtual infrastructure services such as network, storage, backup, data replications and some security services like firewalls.
Now that we have virtualised our servers, we can run multiple virtual systems on a single server. But weve taken this to a whole new level in terms of availability. We deploy server farms on which virtual servers run. So we can move virtual instances transparently across instances that can be moved across physical servers, without customers noticing a thing.
This means we can save the customer money and give them cluster-like availability and resiliency for even the smallest machines, he adds.
Virtualisation 2.0 technologies will allow data centres to provide more services, Cockayne says, and allow customers to buy servers as small or large as they need. Furthermore, its an easier spreading of the workload that gives better and more predictable performance.
However, users of outsourced providers will have to look at their DR capability and that of their outsourced provider, who may not have sufficient redundancy or capacity in their systems.
IT managers will also need to model their behaviour to test what might happen in a disaster, should they move to a virtualised and hosted environment.
Sign up for Computerworld eNewsletters.