It was interesting, albeit unsurprising, to hear that IT departments are reading the riot act to Cloud providers when it comes to SLAs. The business is understandably keen on the Cloud model with the cost savings, increased flexibility and efficiencies it can deliver. But these benefits can be easily outweighed by the disadvantages of more frequent outages, lack of control over data and security concerns if these aren't handled up front.
One of the reasons for SLAs becoming more complex is that Cloud services bring an additional layer of considerations. Data can be hosted almost anywhere and is therefore subject to different regulations depending on the host country. Services can be affected by activity completely beyond the provider's control - cloud providers can even change SLAs without informing customers. Take for example the lady in Armenia whose attempts to steal copper led to the entire country being disconnected for five days. Not ideal if that happens to be where your data is stored. It goes without saying it is important to check the small print.
The issue for IT departments is 'SLA multiplication'. Not only are SLAs from service providers increasing in complexity as Cloud brings new issues into play but as the business becomes more tech savvy - thanks largely to the consumerisation of technology - it is demanding tighter SLAs from its own IT department. 99.9% uptime isn't good enough anymore and as access from a growing proliferation of different devices increases so does the expectation of what users expect to be able to do on them. IT departments are now defining their own SLAs with the business which means that they face costly fines or other penalties such as less budget if they're breached.
Add to this the expectations being put on IT departments in terms of measurable business benefit delivered and you can see why automation of areas like SLA management has become so important. On its own however automation isn't enough. It requires an additional layer of intelligence. When it comes to SLAs prevention is far better than cure. Being able to proactively identify problems before they occur is far better than cleaning up after them. Real time monitoring, intelligent updates and flexibility in design means SLA management is no longer about damage limitation, it's about giving IT managers the ability to prevent the damage in the first place. More than that it's about giving them the head space to focus on delivering real value to the business.
This was first published in December 2011