Showing results for 
Search instead for 
Did you mean: 
Partner    VIP    Accredited

Over the past couple of days, I've been watching the meltdown and ensuing anger arising from it of a cloud provider in my country (no names mentioned at all in this article!).

Long story short, admins and users woke up on Friday morning to find all their VMs gone, and a message stating the provider would begin the process of reprovisioning all VMs. That statement in itself should have been major cause for concern, but I doubt many thought of the actual fallout that was still to come. Users and admins logged onto new VMs expecting the same settings, only to find a vanilla VM with a changed IP address (problems if you had to reconfigure applications for this), data! Not only that, but if some users were backing up to separate instances on the same platform and simply had nothing to restore. All VMs gone, and for many, their companies (possibly) and data from years gone by gone too.

The fallout began with people asking where their backed up data was, and the cloud provider simply stated: If you don't back up your data, it's your problem, and then pointed them to a clause in the cloud contract stating this.

This got me thinking: Not many people know that placing your workloads on a cloud platform DOESN'T involve having that particular cloud provider backing up your VMs, but even if they did, they still place their eggs in 1 basket, so to speak. This would either be due to faith in an environment never failing, what with redundancy expected everywhere on a platform, but that backups would always be available.

And if they're not? Problems, business closure, law suits, general disgust with Cloud which may/may not be finding its way to better acceptance, and for the Cloud provider, loss of credentials in the industry and public eyes.

So, what should you be doing to protect your workloads sitting on a cloud platform?

Firstly, read the fine print...if you're placing your livelihood on a Cloud platform, make sure that you read the contract before you sign it so that you know whether or not you've got automated backups for your monthly costs you pay. Cloud providers may, or may not, offer this as a packaged deal. I've seen that the big providers don't, meaning it is an optional extra.

Secondly, choose a reputable provider. Many cloud providers are popping up now stating they cater for specific markets and able to scale according to the size of your business and requirements. But do they have th expertiese from a technical management and operations perspective, a back-end solution design perspective, and functionality perspective, or are they simpy trying to cash in on the hype around Cloud? All of those considerations have to be taken into account before deciding where to put your workloads.

And for me, because I come from that environment, backups! Backups! Backups! Backups! What do they offer? Is it local to the same array? If so, why do that as you risk total data loss in the event of the array being hacked or experiencing a catastrophic hardware failure causing it to default to fatory settings (it happens!)? Do they have a separate array? What sort of redundancy do they have for backups, ie. backup to disk and then create a second copy on another disk array? This offers multiple copies from which to recover from.

However, in this case, another option is to take backup matters into your own hands. Backup Exec, for instance, allows the option to use AWS as a backup target, and even targetting another Cloud provider, you would be able to use the capabilities of Backup Exec to duplicate data out of your cloud provider's platform. An example would be backing up your data locally to a dedupe store for instance on your current Cloud provider, and then using Backup Exec's optimized dedupe functionality with CAS (through Enterprise Server Option) to duplicate the data to another Cloud provider in Europe. A good solution which protects your recoverability.

For me personally, I would make sure this is either in another data center in the same country but a different location, or even another geo location. This allows you to spin up VMs in the target Cloud provider for instance, and recover a lot faster than you may be able too on a crashed platform.

Protecting your data is key to keeping your business open, your employees in a job, and hair it's natural colour. I don't think that data protection should have a value attached to it. It's when it's too late to recover that hindsight becomes 20-20, and the "should've", "would've", "could've" thoughts start.

Consider what you want to do to protect your data, investigate what your current or potential provider does to cater for your requirements, and give good thought to redundancy with your backups! It might just save your business! Don't get caught out when it's too late!


Version history
Last update:
‎07-23-2015 02:26 AM
Updated by: