Cloud Costs Concerning?
I've been interested in the ongoing evolution of the perennial debate around cloud vs on-premise. Last year, you may recall that 37signals, a heavy user of compute services, migrated a huge amount of their workload away from the the cloud and back to on-premise systems.
Now, 37signals are not a tech-unsavvy business - they are the makers of Basecamp and various other web applications, and they created Ruby on Rails for internal use before releasing it as open source back in the day. In short, they know what they are doing with online systems - but they managed to save nearly $2 million in one year. The first year savings also paid for all of the new server infrastructure they needed - and this is kit they will be using for at least five years!
Well, they've gone a step further now, and have also started moving petabytes of data from AWS's S3 service - a staple of many web businesses - to their own hardware spread across two datacentres for resilience. And this time they are expecting to save $1.3 million a year - and get more storage as well!
Now, cloud computing means different things to different people, and not many of us will have as high a storage need as 37signals, nor as much of a compute load. But it is interesting to note that this company have improved their compute and storage provision, saved millions a year, and have done it without changing or increasing their staff.
In my own little corner of the digital world, public sector IT, cloud computing has very rapidly become the preferred route for all services, particularly with the way many of them are bundled up with Office 365. Given local government IT also deals with a bewildering array of niche applications for all the various services councils provide, cloud based services can offer economies of scale. But I think it may be worth us all taking another step back and really getting into the details on this - are we actually saving money? And what are the pros and cons of on-premise we are saying goodbye to?