Netflix Christmas Eve outage (2012): How to better avoid disruption of services in the Cloud
As the Cloud becomes ubiquitous and coherent, most companies rely on public Cloud infrastructure as part of their operations because it is convenient and inexpensive. Likewise, there is another side to a public Cloud that is just beyond the customer’s control: the Cloud infrastructure. I do believe that it is the obligation of the customers to make sure that their Cloud systems are up and running all the time to avoid service interruption.
These are some proactive steps to take to prevent a public Cloud’s service disruption.
1. Companies need to define what aspects of their businesses are more critical that address business continuity. This process should be one of the top priority that most CIO and CEO need to understand before adopting the public cloud.
2. Establish DR (disaster recovery plan) or automatic failover process.
3. There is a need to have updated documentation on recovery process during an outage, as most companies rely on manual recovery procedures; however, I do believe that company should document both manual and automatic recovery plans.
4. Data backup is also necessary. It is crucial that a backup and recovery plan must be in place for most critical data. You can choose to backup in the Cloud, disk or tape.
5. Companies should understand their cloud provider’s service-level agreement (SLA) and how it is relevant to their business structure.
6. Also, spreading workloads across different regions can also reduce service outage such as setting up cloud computing in two separate areas. Most prominent cloud providers like Amazon and Rackspace have multiple data centers across different regions. For example, you can set up systems in East Coast and West Coast and the West Coast, so if one of the regions goes down, you still have the other location running. While the region is down, you can move your entire workload to the site that is up and running and switch back once the region comes back online.