Tears of a cloud: Don’t be let down by backup and disaster recovery

For many, the advent of cloud storage was a blessing. Cherished pictures and videos, contact lists, documents and more could be automatically put online and saved (theoretically) forever.

Enterprises took notice as well and now, cloud backups are fairly standard practice. However, business and individuals have one significant difference: the amount of data they have to back up.

For an individual user, backing up their archive of photos takes no more than a few minutes.

For a business, this process can take days or even weeks to establish redundancy – and you can add more days to pull that back down.

Between you and your target data centre there are several variables at play, including both your ISP and those your backup service gets their connectivity from. Maybe the data centre burns down during the transfer, leaving you high and dry.

Perhaps the ISP on the other side of the country is getting swallowed by a hurricane, or maybe someone drove into your node and your ISP can’t get out to repair it until after the weekend.

If anything goes wrong somewhere on the way, not only is it out of your control as the administrator, but you’re also at fault for not planning for that possibility. That’s not healthy for your career, nor is the downtime good for productivity.

Maybe some day the net will be fast and robust enough that an entire lifetime of business data can be uploaded to the cloud and backed up in a small enough window of time that nothing can conceivably go wrong, but that is simply not the case today – at least not for everyone.

Fortunately, disaster-proof recovery offerings do exist that let you to maintain an internal focus of control in your data centre. There is, for example, ioSafe that makes several different NAS devices with the explicit purpose of being disaster-proof. I’ve tested ioSafe 214 and can report that it is indeed possible to fireproof (and waterproof) your backups.

Disaster-proof devices integrate directly into your extant network by becoming another backup point, exactly like you would expect from a NAS. From there, administrators can back up that data to the cloud, on to tapes, or to another data centre. Should anything go sideways during that backup process, you can count on having intact hard drives in the disaster-proof device – even if a dissident employee burns down the building.

Smaller businesses can even run their backups on the disaster-proof device and forgo more expensive offsite storage or having someone transport tapes out of the building. Sadly, this just isn't a realistic option as you move upmarket: disaster-proof storage doesn't currently scale into the petabyte range.

Larger businesses with a hybrid cloud solution will maintain a local copy of their data and toss backups (or active-passive failover copies) up to a service provider or public cloud provider. While this works well under most circumstances, the existence of the backup gap becomes especially crucial during internet outages as the difference between what's on the local backup systems and the off-site solutions grows.

Some businesses with sensitive data cannot store their confidential data on the public cloud due to compliance regulations, and may not be able to go the hybrid route; they may need to maintain their own secondary sites.

The more critical the data, the greater the importance of cost-effective and reliable disaster recovery solutions, and the lower the likelihood that "cost effective" is actually feasible.

The more critical the data, the higher the chance that some form of automated failover (and failback from) secondary sites will be required in the event of a net outage.

"Backup backpressure" builds over time. As soon as the internet comes back on not only must current changes and backups be sent, but all those that have built up during the outage must as well. And if your data centre should burn down before you're all caught up, well...

The largest of the large enterprises can solve this problem by paying lots and lots of money for redundant network links, backup generators, secondary sites and/or public cloud services. (If they legally are allowed to use them.) The SMB can use existing disaster proof storage solutions to mind the gap.

But what of the midmarket — those millions of companies that can't afford enterprise-level connectivity but can't squeeze their data into disaster-proof backup buffers? There is a backup gap in the market waiting to be bridged. ®