Disclaimer: I recently attended Storage Field Day 8. My flights, accommodation and other expenses were paid for by Tech Field Day. There is no requirement for me to blog about any of the content presented and I am not compensated in any way for my time at the event. Some materials presented were discussed under NDA and don’t form part of my blog posts, but could influence future discussions.
For each of the presentations I attended at SFD8, there are a few things I want to include in the post. Firstly, you can see video footage of the INFINIDAT presentation here. You can also download my raw notes from the presentation here. Finally, here’s a link to the INFINIDAT website that covers some of what they presented.
If you’ve been around the traps, you’ve likely heard of Moshe Yanai. He had an awful lot to do with EMC’s Symmetrix for an awfully long time, and went on to fund and lead XIV, which was sold to IBM. He knows a lot about storage, particularly at the big end of town, and he’s back for another crack.
INFINIDAT states that their “modest goal [is to] store humanity’s knowledge, forever”. These are some pretty lofty ambitions, to be sure. Someone at the presentation referred to this as “Moshe v3.0”. I thought that was a nice idea, but wondered what it really meant in light of what I knew about Symm and XIV.
Infinibox Architecture
INFINIDAT have stated that their design principles are as follows:
- provide the right infrastructure for the current and future data explosion;
- add onboard capabilities to leverage and optimise the use of data;
- provide superior mechanisms for data protection;
- overthrow existing paradigms of storage-application integration; and
- support unlimited use cases by eliminating cost constraints.
Great. So that all seems like a reasonable list of goals, and if you’re going to have a shot, why not go big?
INFINIDAT uses commodity hardware to power its offering, with Dell PowerEdge R730s being the choice of platform here. The architecture employs the following key specs:
- Triple-redundant power (N+2 redundant on everything) and data path
- Three active-active-active nodes (Linux servers clustered together with InfiniBand point-to-point connections)
- 24 x 8Gbps FC ports
- 6/12 x 10 Gbps Ethernet ports
- Up to 3.2TB RAM
- Up to 48TB SSD
- Up to 480 disk drives
- Support 3, 4 and 6TB NL-SAS drives
Also noteworthy is that the disk enclosures are connected to DC power, not the internal system UPS. There are eight disk enclosures. These have two PSUs, two I/O modules (SAS expander) and house 60 top-loading drives. You can have up to 480 6TB drives deployed with one of these things. This seems like a lot of spinning disks. In addition to this, every node can see every disk.
Not only is there a lot of available performance in these systems, they’ve also given some thought to the type of resiliency you’d expect to see in a high-end storage array. You can read more about INFINIDAT’s storage architecture here (registration required).
Closing Thoughts and Further Reading
While every man and their dog is working for a storage startup that’s based on an architecture that provides massive scale-out, cloud-optimized, analytic-laden, cheap-as-chips storage services for the masses, INFINIDAT have taken a different approach entirely, and really gone for the high-end of town. Given Moshe’s background, it’s really no surprise that this is the sort of product they’ve come out with. Nonetheless, it will be interesting to see how they compete in the VMAX and XIV space, and whether the approach will pay off. El Reg has a pretty decent write-up on them here that I recommend checking out as well. Stephen also has a post here that is worth your time.
If you’re in the market for a new approach to big-iron storage, INFINIDAT might be just the ticket and would be worth looking into. I’m really looking forward to seeing how they go over the next 12 – 18 months.