Once upon a time we had to own, configure, run, update and back up expensive servers in a complex hosting environment.
As a Christmas present to itself and clients, in late 2014, IRM transformed its hosting environment: moving from the original complicated server setup, to a cloud based solution with Amazon Web Services. We are delighted with the result – here’s why…
Yes we are!
IRM was a cloud computing provider before such catchy words existed. We have always hosted and supported client websites and our content management system in a managed services environment. Back in the day, the infrastructure that we provided meant we were referred to as an Application Service Provider, or ASP.
Shared service data centres were the best way to do this when IRM started. We hired a rack in a data centre, and populated it with our own servers. The data centre provider looked after the physical environment, delivered internet capability via a dedicated set of lines from multiple ISP’s and offered 24/7 monitoring and support for the machines. IRM used a shared service like this for 12 years. We delivered very high availability of websites for clients, suffering very few unscheduled outages.
Despite all this, our whole service had physical limitations that were not easy to improve. So we constantly had to capacity plan. We were still responsible for upgrading the operating systems and hardware in our rack. Infrastructure growth was in discrete lumps, and each upgrade was delivered with considerable effort from expensive IT people. And then there was the capital cost of the hardware itself.
Use of infrastructure in the cloud completely changes the management requirements and risk profile of the IRM service. We now don’t own a server, and we don’t have Microsoft server software licences. We buy on-demand server capability, including all the licencing, by the hour. We have no limitations on our internet demands. There’s no dedicated access, we just share the AWS capacity in a fully secure way, which delivers appropriate capacity in peak times and when there are troughs.
We created fresh new versions of all of our servers, with the latest operating system software, on the Amazon Web Services (AWS) service, using one of their Sydney datacentres. Your website data is still hosted in Sydney. We migrated the client websites and databases one at a time to the new servers. We helped each client change their DNS entries to point to the new servers. Overall it took us five months. We ran duplicate infrastructure for that period, and for the most part, things went smoothly.
In the cloud, the new servers are as familiar as our IRM-owned servers – once we log on, they look just like they always did. But if we want to expand the disk space, run up another server, upgrade a server to a bigger one – more memory, more processors etc – we just go to the AWS console and do it there in a few minutes. No mess, no fuss – access in minutes.
Server hardware might still fail. It hasn’t happened yet, but if it does we just re-start the server, and it will seamlessly switch over to running on different hardware. Backups? We still do all our file level backups as before. We can choose to store them in a different datacentre just by specifying the volume location on the console. We do volume backups and take machine image copies. All from the console.
Hardware has become software. And by the way, many of the console functions are available on an AWS mobile phone app. We can run up another server from the bus on the way to work with more ease and confidence now than ever before.
The benefit for clients is that you can even more confidently expect that the IRM service will respond quickly, reliably and will grow seamlessly as your demands change and as IRM’s client base grows.