6 Ways to Protect Your Enterprise from a Cloud Outage
As more enterprises turn to the cloud, there is more public attention on data downtime and outages. There are some tactics you can take to keep some enterprise control and protection.
Here are six ways to protect your enterprise data from a cloud computing outage.
Recent high-profile outages shouldnt scare enterprises away from deployments, but they should lead to more scrutiny over service level agreements, according to Gartner VP Robert Desisto. Just because a vendor has had a good track record for uptime does not mean they will have good uptime in the future, the distinguished analyst wrote in a recent blog. Desisto recommends starting with a review of best practice SLAs.
Regional weather disasters or data center equipment malfunctions have a hard time knocking your cloud capabilities offline if you spread out your deployments. In his commandments on cloud failure, cloud writer and OpenStack specialist Sriram Subramanian says enterprises should divvy service providers and load balance across more than one region and periodically perform data backups in separate geographic areas.
If an outage occurs in connection with a hacker incident, additional encryption on your part goes a long way toward keeping data safe. In a review of financial industry cloud best practices, Tsion Gonen, chief strategy officer at protection company SafeNet said it is crucial to encrypt all data to prevent any further damage after a cloud outage. "We're in an age where by definition you cannot assume you will be able to completely seal the perimeter. Someone will get through the door, and if he doesn't get through the door he'll get through the window. He will get to the asset you're trying to protect. That asset in most if not all cases is a piece of data."
As the cloud matures, so have innovative ways to address it. For instance, Netflix found a silver lining in its cloud portals failing often. The movie provider released its cloud infrastructure software, Chaos Monkey, to the open source community over the summer as a way to advance and strengthen on-demand testing.
The cloud itself may give you a chance to get at data before an outage puts it out of reach. Automated cloud bursting can include notifications on provider outages so data can be manually handled at another data center site. Solid cloud on-ramping can keep from additional backlogs or unwanted deployments.
Along with the downtime and coverage in your SLA, its important to gauge your own cloud usage levels to avoid unnecessary deployments and, subsequently, vendor outages, writes virtualization solution architect Mike DeNapoli. Ask if you can buy capacity in very small chunks and change usage on the fly. Check about billing based on the amount of cloud you consume, or any long-term commitment to usage patterns.