The world of cloud computing is new enough that it is still tackling basic issues of security, like data recovery, contractual liabilities and resiliency infrastructure for cloud servers. The important questions coming from cloud clients though are simple; most importantly, do you still need to back up your data? After all, wasn’t outsourced IT supposed to take care of this for you?

Backup

Backups may be a grinding chore but they fulfil a basic principle: don’t put all your eggs in one basket. Store your information, and copies of it, safely and reliably, where you can access it when you need it, and never put all of your copies of anything in the same location. Redundancy is resiliency. This applies even to cloud servers because no one can guarantee that cloud servers will not go down simultaneously and that their back-up systems (if any) will not fail either.

Nothing is Totally Safe

When Amazon’s Elastic Block Store (EBS) and Relational Database Service (RDS) went down in its US-East-1 region on April 21, 2011, it incapacitated some websites, slowed others down and showed the world that even with all the efficiencies of cloud computing, nothing is fail-safe when it comes to data loss and recovery. Amazon’s business IT support found themselves able to do little to deal with the situation, and showed cloud computing to indeed be vulnerable.

If you were a business running on Amazon’s EC2 at the time, especially in the US-East-1 region, you would have been glad had you kept an up-to-date backup of your data somewhere other than on the EC2. The cloud itself was unaffected, but users with their data in the affected availability zones of US-East-1 region found themselves unable to access it, making their websites and operations on the EC2 useless. Had businesses had their data running across different regions and availability zones, or on a local emergency server, they could have had the resiliency to weather the downtime.

Cloud Coverage

Some companies are going the route of running two clouds and interfacing them carefully, which introduces greater resiliency across the system and minimises a repeat of what happened at Amazon’s EC2. One of the problems that led many of Amazon’s clients to be vulnerable to data loss or slowdown due to the outage were overly tight restrictions and fess on data transfers across regions, as well as the physical set-up underpinning the availability zone structure. That means that some of the problems can be fixed at the policy or SLA level.

Lessons

If anything, this experience taught people a valuable lesson: the basic information saving rules apply even when exploring bright new horizons like cloud computing. Cloud servers are on a different level of abstraction and efficiency than hard drives and data centres but ultimately they rely on physical hardware, which is vulnerable to things such as power outages, human error, old age, natural disaster and the like. That means that there is always the possibility that the servers will go down and huge sections of data will be rendered inaccessible or lost. When that happens, will you have an updated backup of your data to fall back on?