I don’t think cloud SLAs are a thing of the past. On the contrary, I think they are a thing of the future.

One of the things that are probably stopping companies from adopting the cloud is precisely the lack of attractive and clear SLAs. And with the service outages suffered by Amazon and other cloud providers’ costumers recently, it becomes clearer that SLAs will become a key factor when considering moving to the cloud.

I couldn’t agree more with Carlosdl.

SLA’s not only form a foundation for the contract between yourself and the service provider but also for key business stratergies such as your DR (Disaster Recovery) plan. If there is no SLA of any form then any service provider could take weeks about fixing any issue and still be within their contract rights to do so, leaveing the customer without a service that is costing money.

SLA’s protect the customer and the provider. The lack of strong iron cast SLA’s is not helping anyone.

I agree with the consensus so far on this topic.

I wonder if we aren’t seeing a change in thinking from the old break/fix style of Incident management strategy to a cloud based managed services business model. In the Break Fix model, the client paid for response time and the billable hours after the system was up and running. Availability was hardly considered and perhaps customers having yet made the jump in thinking yet.

With an SLA (Service Level Agreement) the Hosting company guarantees a level of service availability. If that availability is below the SLA then a penalty is charged but only for the money paid originally for the service. It’s really the only way to grade services without a complete failure in service.

In the article on Cloud SLA’s the writer implies that the cloud service takes no responsibility for failures or data security. Having worked on various SaaS and Cloud services for multiple companies including BPOS I do know that there is a consequence to the host provider when the SLA is not met.

Still that consequence does not equal more than what the customer is paying to the Cloud provider.

The real problem for the customer is the lost productivity when the system is down. Whether the IT department is on premise or in the cloud, there is always a risk that something will go wrong. Cloud technology tends to be 5 to 15% more available on average than the typical IT Outsource Company or IT Department that has no Problem Management role in the support model.

Conservative estimates on the productivity loss during an IT Failure range between $7,000 / hour and $50,000 per hour for small and medium size businesses. This is just the business loss in productivity. The question then is should a hosting company providing a service at $1000 / month (as an example) pay more than $1000 to the customer for business losses during an IT outage? If the power company has a break in a line and can’t provide power for a day, can the business sue for business losses because there was no power? IT outsource companies don’t provide that type of guarantee on their services.

In the future I imagine that the SLA will become much more important. Perhaps Cloud providers will be rated, perhaps like bonds and credit worthiness is rated, for the quality of their SLA fulfillment. Because the SLA is the simplest way to quantify service that doesn’t require a complete failure, I don’t think the SLA is going away any time soon.

Great, thanks so much for everyone’s input! I’ve added your Knowledge Points.

Keep the insight and thoughts coming!

- Melanie

I agree with Carlosdl. SLA are needed as part of the adoption of Cloud.

Funny, the posting from July 28th mentioned a recent outage at Amazon, and yet here on August 10th, there’s been yet another Amazon EC2 outage. SLA’s are important, but so is having a legal team ready to press matters against your vendor when they fail to mete out the promises of the agreement. And if the data goes down in a puff of smoke (recall Microsoft’s Sidekick/Hiptop data loss in 2009), your vendor’s failure becomes _your_ failure, and your customers will hold you to it.

good answer dcdude

I don’t believe the SLA for cloud is dead. However, as it is somewhat flawed in defining the deliverables guaranteed in a contractual aspect, it would have to revised to be more accurate and better interpreted if it has been met or not.

Service Level Agreements are surely not just hardware dependent. They require many levels of involvement, from the CEO down to the user.

SLAs should be always clearly defined before a physical medium is decided upon. The medium is a tool after all. It is not a fashion accessory.

If the cloud gives a more cost effective way of working, and the SLA guarantees that the service provider is large enough, and more importantly proficient enough to give that service, then how can the SLA be a thing of the past?

SLAs should start locally within the business, then when it is offered out to the best possible provider. Any company that thinks it is big enough to command instant outsider support ,without a formal agreement, is doomed to flounder at the first major disaster. (downtime = profit loss)