Put your feet up and make the virtual data centre work for you
Control every layer - without upsetting your service provider
If you have run a virtualised service in your internal setup you will be used to the orchestration facilities the platform gives you.
When you host in someone else's world these facilities are split into three basic camps.
- Things you don’t get:
What you won't get is server affinity (being able to guarantee that two virtual servers run on the same physical host, and that if one moves the other follows it automatically) because you just won't know what the hardware is.
The things you don't get will simply be replaced by SLAs. In your own world you use server affinity to ensure performance, whereas in a hosted service you have the SLA to fall back on if performance is not as good as promised.
- Things you get but don't see:
If you run your own virtual setup you may well choose to configure the high-availability options that will move servers around automatically in the event of a system failure.
This is generally available in a virtual data centre, but you don't configure it; it is there automatically if you sign up for the right type of service. Someone else takes care of it for you so you don't need to know anything about it.
Some of your automation tasks simply go away
- Things you get:
There will still be features that are exposed to you, though possibly not that many. The one you will definitely crave is server anti-affinity. This may seem odd, given that I've said you won't get server affinity, but it does make sense.
If you are running a resilient application with primary and secondary servers, you need to be able to tick a box that guarantees they will never be running on the same box.
While the orchestration features of the setup deal with the virtualisation-specific functions you need, there is also stuff that you want to automate outside the confines of the virtual setup. Let's go for three categories once more.
- Stuff you no longer need:
Say you enable automated de-duplication on your own SAN, you won't have this capability in the virtual data centre. The point is, of course, that you don't need it because that is not your problem.
You are paying the service provider for a certain amount of storage that runs at a certain speed, and you don't care how it is provided. So some of your automation tasks simply go away.
- Stuff you can't have:
There is every chance that you will have to do without some of the features you are used to. Perhaps you used an automation feature in your own virtual server infrastructure to move systems around, restart them overnight, and so on.
It may well be that you don't get this in someone else's setup. It is one area, then, where you may have to come to terms with losing some functionality.
- Stuff that doesn't change:
Although you can't automate anything on the hardware, you are not precluded from keeping everything else.
If you use a managed backup app you can still use its automation features, and the operating systems of your virtual servers are perfectly capable of running automation engines.
So nothing changes with automation at the operating system level, and perhaps you can even build on that to replace some of the features you lost because the provider doesn't support them.
Love is in the air
The concept of a virtual data centre is a powerful one. You get a chunk of virtual hardware that you can play with knowing that you have an SLA to make sure it has a chance of performing as desired. No matter how badly you configure it, you can't trample on anyone else’s – and more importantly, they can't trample on yours).
If you are used to virtualisation you will probably lose some features you are used to, but you probably won't miss them. Either they are not needed or they are there behind the scenes with someone else looking after them for you.
And if you have never ventured into in-house virtualisation before, your eyes will be opened and you will love it. ®
Sponsored: Global DDoS threat landscape report