My clients at WibiData:
- Think they’re an application software company …
- … but actually are talking about what I call analytic application subsystems.
- Haven’t announced or shipped any of those either …
- … but will shortly.
- Have meanwhile shipped some cool enabling technology.
- Name their products after sushi restaurants.
Yeah, I like these guys.
If you’re building an application that “obviously” calls for a NoSQL database, and which has a strong predictive modeling aspect, then WibiData has thought more cleverly about what you need than most vendors I can think of. More precisely, WibiData has thought cleverly about your data management, movement, crunching, serving, and integration. For pure modeling sophistication, you should look elsewhere — but WibiData will gladly integrate with or execute those models for you.
WibiData’s enabling technology, now called Kiji, is a collection of modules, libraries, and so on — think Spring — running over Hadoop/HBase. Except for some newfound modularity, it is much like what I described at the time of WibiData’s launch or what WibiData further disclosed a few months later. Key aspects include:
- A way to define schemas in HBase, including ones that change as rapidly as consumer-interaction applications require.
- An analytic framework called “Produce/Gather”, which can execute at human real-time speeds (via its own execution engine) or with higher throughput in batch mode (by invoking Hadoop MapReduce).
- Enough load capabilities, Hive interaction, and so on to get data into the proper structure in Kiji in the first place.
More specifically, 5 of the 6 Kiji modules are:
- KijiSchema, the oldest of the bunch (not that any have been out for long). A “direct wrapper” for HBase, KijiSchema provides HBase with a DDL (Data Description Language), metadata (I haven’t asked details), and serialization (via Avro).
- KijiMR, released in February. KijiMR is the batch part of Kiji data processing — notably MapReduce integration (as you might have guessed from the name) and bulk import.
- KijiScoring,* released very recently. That’s the (human) real-time part of Kiji processing.
- KijiHive, which is — you guessed it! — Kiji’s Hive adapter.
- KijiREST.* I bet you can guess what that does too.
*It’s a roadmap item to “blur” KijiScoring and KijiREST together, and have them run in the same process.
And then there’s the confusing one — KijiExpress, which has something to do with running models. Right now KijiExpress seems to be mainly a Scalding/Cascading interface to the rest of Kiji. However:
- Just about anything else you might imagine that could help with specifying, running or managing analytic models sounds like it’s at least a possibility for eventual inclusion in KijiExpress …
- … but it seems like an exaggeration to say that any particular direction for the project is already on the roadmap.
Finally, on the business side:
- WibiData has very few production customers to date, in each case for a rather custom system.
- WibiData knows what its first application vertical market will be, and will probably tell you under NDA if you’re in that segment and you ask.
- WibiData will gladly support Kiji developer seats in the usual commercial open source way.
- WibiData will not exactly support Kiji runtimes in the usual commercial open source way. Rather, it will support entire applications — or perhaps “applications” — built in/on Kiji. Presumably, WibiData professional services would be involved in building those apps, but that’s not a strict requirement.
- Based on mailing list activity and so on, WibiData believes there are serious Kiji users with whom it has no financial relationship.
- I counted 30 distinct headshot photos on WibiData’s team page, in a 9:1 primate:canine ratio.