Cloud backups: Where's my get out of jail card?

Quis custodiet ipsos custodes?

It is time for your reminder to review, test and otherwise care about backups. Today's twist is probing the cloud to see if we need to worry about cloudy backup ourselves. If so, how exactly do we do it?

The more I delve into the state of cloud services the more I am disturbed by what I see. Every vendor has a completely different backup policy; some have none at all.

Infrastructure-as-a-service (IaaS) vendors seem to be the worst for this. We are often lead to believe that virtual machines and storage on a public cloud are invincible.

When giants fall

This is far from the truth. Even the mighty Amazon permanently lost customer data just over a year ago.

Think about that for a second. Amazon lost customer data. We are not talking Ma and Pa's Old Time Co-Lo and data hosting. This is the company that beat out IBM for the new CIA datacentre.

Amazon is the best of the best at cloudy tech and even it does not offer a default 100 per cent guarantee that your data won't go walkabout and never come back.

You can host your data in multiple regions of Amazon's cloud, but multi-region outages can occur – at which point you're hooped.

You may find a public-cloud vendor that you are sure has so many layers of backups that your active data is indestructible, but this is a guarantee only of the working set and not a true backup.

Working set technologies, such as high availability, fault tolerance and the like, protect against infrastructure failure. They do not protect against Oopsie McFumblefingers making a huge mistake during data entry or accidentally deleting something.

We're only human

In the real world, protection against infrastructure failure is not enough. You will inevitably run into those moments where you must pull up a copy of the data as it existed at a previous point in time and work on it.

You can never completely engineer people out of the data collection process. If it isn't McFumblefingers punching information into a form and hitting submit, then it is a script somewhere talking to an API or reading input files from a hotfolder. APIs evolve. Output formats are tweaked.

In some cases changes are made to address security problems. In others it could be a politician revising daylight savings times or some nation declaring independence. There is always something that needs an update. And updates are ultimately made by humans so eventually something will go wrong.

This means there is a need for version control – incremental backups as in days of yore. Backups are easy as pie if you are running your cloud on your own infrastructure. vSphere can take snapshots of running virtual machines for you and Unitrends (or similar) can back up pretty much anything under the sun.

You can even back up that data (or copies of that data) to the public cloud using software such as Asigra. I pick Asigra because it appears to offer several different types of cloudy backup and it is Canadian.

I have issues with my data living in the US. Asigra provides tech that allows me (as a managed service provider) to be the backup location for my clients.