Tape sucks, move on. So goes the slogan of Data Domain – or rather EMC Data Domain. The market was briefly puzzled by Joe Tucci’s expensive acquisition of a product which was, on the face of it, a competitor to an established and well-liked EMC product. That puzzlement rapidly evaporated as Data Domain was integrated into the new Backup and Recovery Solutions group. Long-term EMC and Data Domain customers are still debating who took over who in this space; DD’s culture suffuses the BRS group and increasingly EMC is “playing nice” even with companies and products with whom they go head to head on a daily basis. If EMC can’t own your backup software then they will own your dedupe target. We are EMC of Borg and you will be assimilated.
It’s been said before that backup is dead, dedupe and replication is everything. I disagree with that, for all kinds of pragmatic reasons, but I have to say: backup sucks. So I have a problem with the name of this new division.
You see, we don’t want to back up our data. We don’t want to restore it either. We just want it to be available somewhere we can use it, ideally without the input of the techies who are currently poring over the smoking ruin of whatever it was the data used to be on. Backup was only ever a means to an end, and there are better ways to achieve that end now – ways that were expensive but are getting cheaper all the time.
Tape sucks. Backup sucks. And very definitely recovery sucks. So the backup and recovery division of EMC will I am sure not be long in changing its name to the something-else-that-does-not-suck division. And we’re having exactly the same conversations internally and with our business units.
What sucks most of all is paying huge sums of money to software vendors to write data to special storage in a form only their software can understand. We’re all building private clouds right now, part of the idea of cloud is to remove yourself from lock-in. Why add it in the continuity space?
Instead of restore, why not just present the data in usable form at the target site? Recovery of data, recovery of service, recovery of performance. What we need to do is converge recovery of data and service, make them close to immediate, and then offer self-service staging to recovery of performance for the services which are important to the business on that day. Quarter end you want the financials to be prioritised, mid month it might be the build services required to meet the next ship date for your software product. The business wants to decide, and ideally decide based on the actual situation. Is the primary site a smoking ruin or will it be back in 48 hours? The decision may well have different outcomes for different scenarios.
So I no longer want backup software, and I especially don’t want restore software. I want to replicate my virtual infrastructure in a cost-effective way, maybe a once a day snapshot rather than write by write to keep the bandwidth costs down but for some services it could well be write by write. The business wants to decide, and be free to change its mind, so that’s what I have to deliver. Companies that do Backup and Recovery with capital letters are behind the curve on this.