On March 24th, Duncan Epping posted a new blog entitled “Startup intro: Rubrik. Backup and recovery redefined” and subsequently tweeted said post. On that same day in another part of the world (my office), we had paperwork in hand, waiting to be inked, to refresh aging EMC Avamar Gen4 nodes with an Avamar/DataDomain combo. We had looked at several other options from HP, Dell, and Veeam, but it was all just more of the same with a minor pro or con, but nothing worth writing about (including Avamar/DD). No one had really advanced what VCB (VMware Consolidated Backup) brought to the market in 2007.
Then I saw Duncan’s tweet, and I thought to myself, “Hey! This sounds like what we were trying to get when we bought Avamar in 2011!” So I hopped over to rubrik.com, which pretty much consisted of the Aurora Borealis and a button to click for “Early Access”–simplicity from the start! :) The next day, Mike and the guys at Rubrik walked through a demo that confirmed the revolutionary impression I’d started to gather from Duncan. Sign me up!
On April 29th, it hit the floor in two data centers with Eric and Ray shepherding the process (we’re talking beta here, so it’s only prudent to have some authorities on hand to ensure success). Lunch and driving the 15 minutes between sites took the longest part of the install. Seriously. The installs were complete and protecting VMs before the clock struck noon.
As of the GA announcement today, May 26, we’ve been running on Rubrik for 28 days, and I can confidently say that this is something you want to get your hands on. It’s Google’s search & insight married to Apple’s beauty & intuitive design.
Paradigm Shift: Backup & Boot
As I walked through the product with the guys, I had to intentionally lay aside the trappings of old backup and the caveats of the past. The biggest thing I had to let go of was my concept of “restore”. In my previous platform, I had to run through periodic exercises restoring numerous VMs to spare primary storage in order to prove out disaster recovery capabilities. If you’re familiar with Avamar, you know this is an arduous process. Even with the advancements that come with DataDomain integration, it doesn’t take many steps forward.
In a Rubrik world, “Restore” is still an option (don’t get scared), but it’s just for circumstances when you want to roll back the original VM to a point in time (like Apple Time Machine). The action word for DR validation and even Test/Dev is “Mount”. I know, I know. DataDomain allows you to boot VMs directly from itself, too. Just ask anyone technical, though, and they’ll tell you the truth: it won’t perform and it’s one at a time (excellent ingest, slow recovery). Not so with Rubrik.
Mounting VMs in Rubrik automatically creates the NFS mounts on ESXi hosts and spins up VMs–with 1000s of IOPS of performance–from its hybrid of SSD and near-line disks. Once those VMs are running and servicing the business, storage vMotion them off NFS to primary storage and keep rolling. Or if it’s just a test, shut them down after login and delete them from vSphere. I’ve done it.
Time Machine: To The Future
Rubrik (the organization) is challenging the status quo, and the backup space is just their first step. If you get the chance to spend five minutes with Bipul Sinha, Rubrik’s CEO, you’ll know what I’m talking about. He is a man with both eyes to the future. We live in the age of the start-up, where all kinds of innovation is popping up, but Bipul and team are looking beyond all that. I laugh thinking this, but our discussion made me want to push “Pause” on all our infrastructure projects and just wait for what they’ll invent. Of course, when that happens, I’ll need a new job, because IT will be too easy.