Some recent good and bad experiences with clients have lead me to believe there are ten crucial rules when working with the Cloud here is the first, relating to Disaster Recovery :
One of the 10 commandments of Cloud Computing
1. Thou shalt implement Disaster Recovery for your Cloud services:
A recent experience with one of our clients highlights the need to have a Disaster Recovery solution as part of your cloud strategy in place.
We have found, it is essential for any service that you are migrating to the cloud, including virtual machines and applications. Our client made a decision not to implement DR for their cloud service against our strong recommendation they do so. Their decision was based on cost and the 99.9% financially backed availability service level agreement provided by the cloud vendor.
Unfortunately two weeks ago the cloud vendor had a four-day outage and could not provide an ETA for when the service would be available. The other issue was that the cloud vendor posted on its portal that customers had a four hour window to back up their data and strongly urged all its clients to do so, not good.
Luckily we were monitoring this portal closely and were able to back the client’s data up for them. From what I understand some of the other customers of this cloud vendor were not so fortunate.
This outage was unacceptable to the client as this was a production (but not critical) service and the 30% credit in fees offered by the cloud vendor was cold comfort to our client. Luckily we had implemented the cloud solution on a “true” cloud platform, which meant there were no contracts and utility billing.
We were able to migrate the client to a different cloud provider within 24 hours and get the client up and running again quickly. I can only imagine the issues we would have had if the solution were on a hosted platform with the typical monthly, one, two or three year contracts. The client would have had no option but to suffer through the outage.
The amazing thing is that the cloud vendor we migrated the client to, also had a 72 hour outage two days after we moved the service over.
Needless to say we are now implementing a DR solution for the cloud service on the client’s site to ensure the service is available no matter what happens in the Cloud based environments, they work with.
I am sure many IT managers will see this scenario as an excuse to say the cloud is not secure or the cloud isn’t ready. However Cloud specialists see this as a positive story and a precautionary tale.
Here are some good lessons to learn from this situation:
• Make sure that when you sign up with a cloud service that it is a true pay as you go service. This is usually where you pay by the hour. If a vendor wants to lock you in by the month or year, it is not a cloud service but a managed or hosted service. The by the hour pricing model ensures you only pay for what you use and also that you can move elsewhere when (and if) you need to.
• If the solution is not tolerant of outages and data loss; ensure that you architect Disaster Recovery and Backup solution as part of your cloud solution.
• Many clients get dazzled by the ‘do it yourself ‘ease that is marketed as part of the benefits of the cloud. This may be correct in the consumer space, however in the corporate and enterprise… only fools rush in. This is where data loss, security breaches and outages occur. Therefore engage with skilled partners, understand the cloud space and that can architect, build and migrate these solutions.
If you are thinking about Cloud planning and moving towards Cloud migration please contact us at Cloud Solutions Group to discuss your cloud projects and requirements:
By Josh Rubens