Bernard Golden is giving up on the service level agreement (SLA). He finds this once essential component of the vendor relationship has now spiraled into irrelevance in light of cloud sourcing. In an article for CIO.com, Golden explains his rationale for condemning this IT tradition.
When it comes to the cloud, there really isn’t an SLA “solution.” The truth is that cloud providers are constantly working to provide the best availability to their clients. Negotiating an SLA with one will give you zero advantage, says Golden. That’s because there is a dynamic that encourages large cloud vendors to refuse SLAs, while small vendors will sign them simply to get your business. Even if you do negotiate an SLA, at best, you’re getting a small refund for every outage period. Any compensation for loss of business is directed to (a costly) continuity insurance policy:
The biggest problem with a focus on SLA is that it frames the issue incorrectly. It addresses a problem — infrastructure failure — with the wrong solution — legal. Now, don’t expect your legal or finance people to tell you this. As the Sage of Omaha, Warren Buffet says, “Never ask a barber if you need a haircut.” If you ask your lawyer if you should spend time on the SLA, he or she will — of course — tell you yes. That’s their haircut.
The Path to Application Enlightenment
The correct path according to Golden is technical, not legal. He suggests focusing on the following three characteristics when seeking application availability:
- Elasticity and Automation
Redundancy means not relying entirely on hardware or software. To avoid single point of failure (SPOF) one must “implement a redundant application topology.” A little bit of this, a little bit of that. Keeping ever more complex applications online means partitioned architectures. An app is hard to fix or update in a monolithic stack. Instead, what Golden recommends is a set of self-contained components that communicate through a RESTful interface for more rapid improvement. And lastly, elastic applications will be capable of shedding and integrating components to their topology. Therefore, a robust operations management of applications must become automated via standardized components, automation, monitoring, and analytics.
Read the original article at: http://www.cio.com/article/2883770/cloud-computing/the-death-of-the-sla.html