Could Rozo squeeze into the scale-out NAS-object scalability gap?

Rozo Systems is a scale-out NAS startup using Mojette transform erasure coding with its RozoFS software, and claims its software provides high performance at a fraction of the cost of competing systems. There is no adoption cost from the point of view of applications using its software-defined product. It enables, Rozo says, …

[Platypus]

Why do these articles only ever seem to compare against *proprietary* solutions? Another basis of comparison for semi-open-source RozoFS would be truly-open-source Gluster (on which I work) or truly-open-source Ceph, both of which already have erasure coding too. Based on experience with that, I'd say *it doesn't matter* which erasure-coding algorithm involves more addition or multiplication because those calculations are only a minor factor in overall performance. The amount of data that must be transferred, either during normal I/O or during repair, matters far more. The coordination overhead matters even more than that. If you have two clients trying to write overlapping blocks, and they don't coordinate properly, then half of the servers get erasure-coded pieces of one write and half get erasure-coded pieces of the other. This isn't even "last writer wins"; anyone who tries to read that data subsequently gets *garbage* back. The #1 determinant of performance in such systems is how they avoid this issue for every kind of operation (including both data and metadata with all of the atomicity/durability guarantees that must be met to keep users from screaming).

If the Rozo folks want to brag about their erasure-coding efficiency, let's see some actual performance data. While we're at it, let's talk about the scale at which things have really been tested. Anybody can claim hundreds of nodes and multiple exabytes but AFAIK no project in this space has ever successfully run at that scale on the first try. They *always* run into new failure modes and performance anomalies that never appeared at smaller scale and that often require substantial new subsystems to address. Then they find out that customers at this scale are going to want tons of other features as well. Some of these are still only on Rozo's roadmap, after having been shipped years ago by competitors. Others, especially related to multi-tenancy, are still missing entirely.

I think what Rozo is doing is very cool, and I wish them all the success in the world, but let's not lose sight of the fact that there's a *long* row to hoe before even the best ideas turn into a competitive storage solution. They sound a lot like the Ceph folks did *five years ago*, but Ceph (with far more resources at hand) is just now making the transition from bleeding-edge to enterprise-ready. It's not because they lack talent, I can assure you of that. It's just that these problems are *hard*, and solving them takes a lot longer than Evenou and Courtoy seem to think. I'd love to hear from the RozoFS developers about when *they* think RozoFS will be competitive with what's already out there.

[random_graph]

Not a better mousetrap, maybe a better spring

There are so many things that make this story incomplete. Although fast erasure coding is a neat trick, there's no reason to believe that Isilon or Qumulo's (or even Rozo's) performance is going to be dramatically bottle-necked or accelerated by the EC calculation.

When I see so many generalizations and superlative claims (performance, scale, etc) and denigrating all things that have come before, I rapidly become skeptical. If they're indeed courting the market where people care about individual file-systems >20PB and massive performance, that means technical HPTC; The customers are going to be seismic processing, national surveillance initiatives, and national labs (NCAR, LLBL, LHC, etc). In this tiny market space they'll have to deal with feature-rich incumbents, parallel client implementations (GPFS, Panasas, Lustre, pNFS), exotic interconnects, and extreme price sensitivity. There's a massive opportunity cost of going down that path, as illustrated by Panasas's difficulty bridging into the enterprise, and Avere's walking away from the NFS caching market.

Everyone in the NAS space likes rendering and DNA sequencing ("commercial HPC") as targeted use-cases because feature requirements are relatively lean, the workload is deterministic, and buyers are anxious to protect themselves from over-dependency on Isilon. But it's a crowded space that already includes new entrants such as Peaxy. With Isilon/EMC slogging through M&A mud, Qumulo is probably best positioned to take share over the next decade. Hopefully Rozo can take their better spring, build a better mousetrap around it, and demonstrate at the solution level they really have some differentiation.