You’ve just moved house – you thought the sofa fit the new lounge – you were certain. It doesn’t. It fits in the space alright, but the door now won’t open fully – what a pain!
Time for a new sofa?!
Fortunately when you move your IT assets into a new environment (let’s focus on servers, moving to Infrastructure as a Service (IaaS) in the cloud, for the purpose of this article) there are ways to reduce (or increase) the size of certain resources; without resorting to extensive surgery/rebuilding/
You may have seen the marketing material from the cloud providers about rightsizing or scaling – and they’re not spinning you a line. When you move a virtual server (or convert an existing physical one) one of the first things to check is if the server working correctly; one reason you may have a performance issue is down to rightsizing of the processors; normally too many!
This is completely counterintuitive but there are good technical reasons why less is better for certain applications in a virtualised environment.
Once the dust has settled on your conversion/virtualisation/
Your IT team should be monitoring the actual usage of these (and other) resources in order they can adjust accordingly. Processors are not being used – they’re allocated and costing money, and possibly negatively impact real world performance. The same goes with memory, if you’re not seeing the usage but it’s allocated, this can be a waste of money (check your cloud provider for what they’re charging for).
Speaking of which – whilst monitoring for those statistics, your team should be speaking to the application owners, the service managers, the end-users (critically) and gauging feedback the application performs as it should. You’ll often hear back to say things are working far quicker than they once were – this is a fantastic good news story to share across your program, your stakeholders and the entire organisation. Internal case studies and success stories, shared from within your business, are great for advocacy of change.
I’ve talked about reduction of resources – and moderns operating systems will happily allow you to switch off, remove resources and turn back on and get on with the job in hand; the same goes for increasing resources.
Your legacy hardware was creaking – so you moved to the cloud.
Now you have a massive amount of flexibility limited only by the operating system and application type as to how much processor or memory you can throw at a machine, known as scaling up (scaling out is an article for another day). Additionally, hard disks and networking – no longer are you constrained to waiting on hardware to arrive, a few clicks and an entire multi-tier storage architecture can be made available for example.
This isn’t all plain sailing – there are times certain applications can be unsettled by virtualisation and changing of underlying resources, especially if they’re old but this situation is very rare now. I recall a rightsizing requirement on an ERP system a long time (10+ years) ago that needed updating before we could move it to the cloud. We only found out it was an issue at the point of virtualising it, with multiple attempts to get the configuration right, the business and project team decided to roll back and revert to its physical counterpart. The necessary upgrades took place under normal change control in order to get to the final destination, where a successful conversation and rightsizing took place.
You’ll hear me talk about focusing on the goal and changing the plan a bit – this was one of those occasions, we didn’t so much change the plan as we’d take this scenario into consideration but it certainly wasn’t clicked, click, next, the cloud!
Rightsizing allows you to get the most bang for your buck and gain near-immediate benefit from the cloud without refactoring or rearchitecting your existing solutions. If the application allows it means you can scale up whilst running a refactoring or rearchitecting project alongside and keeping in touch with your business needs; leveraging the flexibility and agility of the cloud.