Pure Storage are just after running their first community event, Pure //Accelerate, and by any measure it can be classed as a success. Pure made a number of announcements about new products in their portfolio but the one that caused the most excitement was the announcement about their new FlashBlade product. More on that in just a moment. Pure made announcements about a new baby flash array, the //M10, to bring flash storage to the masses at under $50k and more details about their collaboration with Cisco on the UCS based FlashStack. Both of these announcements by themselves are substantial enough for most storage vendors but Pure stepped it up a notch with the FlashBlade.
I’m not going to go into the components of the FlashBlade, more information on that can be found on the Pure Press Release, a run-down by Alex Galbraith or Enrico Signoretti’s blog post. As with just about every other storage geek out there I got extremely excited when I heard the announcement and imagining the use-cases for the FlashBlade pretty much made my head explode. After a bit of time letting it sink in over a coffee and digesting the cost per GB on the FlashBlade, the scalability and the form factor size I decided that I hadn’t over-reacted in my excitement.
Pure developed the FlashBlade over the past two years with a development team that couldn’t even post on linkedIn that they worked for Pure. At least that’s the rumour. It’s a masterful feat of engineering. In the GA release expected in Q3 of 2016 FlashBlade is expected to have more features than just those announced at //Accelerate. NFS and S3 compatibility are already confirmed with SMB, HDFS and more on the roadmap but not confirmed to be ready at time of GA. It is surprising however that Pure has gone the direction of manufacturing their own proprietary hardware. Just about every other Flash vendor is going the way of commodity hardware. Another proprietary flash/memory based storage array vendor, Violin, is struggling to make a profit. I think Pure with the M series has more strings to its bow and shouldn’t run into similar problems but it’s worth keeping in mind that others have previously tried to plough their own furrow and not have the level of success needed to keep the company afloat.
I have some concerns over the replication factor of 2. Ideally it would be higher. Given that this is a first-gen release and Pure have said that they planned on getting the product out the door and will fine-tune the features with time then I think the replication factor will be dealt in time.
I’m looking forward to catching up with the local Pure reps to get more detailed information on the FlashBlade. I’m keen to know what the roadmap is for FlashBlade, particularly around Intel 3D NAND. The future looks promising for Pure and it looks like they are going to stick around for quite some time. They definitely can’t be accused of not trying to innovate and turn the storage arena on its head.