Cloud Computing: Failures and lessons learned

Outages experienced by customers of major and minor Cloud Service Providers, over the past couple of years and especially in the past two weeks, have resulted in down times beyond those offered in Service Level Agreements (SLAs). While some have made headlines, many have gone unnoticed and unreported. IT executives, including those already embarked on the Cloud or intending to, are scrambling to find reliable means of capitalizing the benefits of Cloud while mitigating risks posed by them.

Hybrid Cloud infrastructure is the key to ensuring minimum disruption to business. Mission critical data should reside in both the Cloud and the company’s internal or co-located data centers. Access to this data should be provided through multiple routes, including Internet and WAN service providers, for both the customers and business partners. Technologies such as Continuous Data Replication/Protection should be deployed to replicate data needing high availability. SLAs from service providers and IT should be in sync and set realistic and acceptable expectations to the consumers. IT infrastructure should be designed, configured, deployed and thoroughly tested to validate the SLA. Mock drills with varying degrees of impact should be performed by teams including IT and representatives from different facets of the business.

As businesses venture into this relatively new medium of computing, expected hiccups are bound to occur that can violate the SLA. Over time, with proactive monitoring, testing and continuous improvement, a credible SLA can be enforced that is acceptable to the business.