Big Data

Basho's database downfall highlights danger of bringing too-niche tech to market

Basho built a great database, Riak, but its demise has far less to do with bad investors than with bad timing for narrow tech.

The Basho lawsuit may be popcorn worthy, but looking at it on the surface doesn't provide the whole story.

As reported in The Register, Basho, the company behind once-popular NoSQL database, Riak, fell into serious disarray when its Series G investor, Georgetown Capital Partners, effectively took control of the already sinking ship by "maneuvering the company into a position of maximum financial distress." The key to remember, however, is that the only reason Basho was in dire need of a Series G (which, for those who need help with the alphabet, came after Series A, B, C, D, E, and F rounds of funding) was because Riak was no longer able to sustain a business around it.

Too simple a DB

While the lawsuit does make for entertaining reading, with statements like "'I want the power to crush him" and "harsh and vile tongue lashings" strewn throughout, the primary court to seal Basho's fate was the free market. There, Riak saw its popularity plummet as the world moved beyond simple key-value stores to more sophisticated databases.

Don't get me wrong. A key-value store is highly useful, whether for caching, rich media stores, or other uses. Of the 343 databases tracked by DB-Engines, key-value stores account for 51 of them, with one, Redis, making the top-10 list.

Key-value databases are great where high performance and massive scalability are required, and when you need simultaneous indexing and persistence. They start to lose their luster, however, if you need to filter on the value field or don't want the performance hit associated with having the whole value returned.

SEE: Special report: Harnessing IoT in the enterprise (free PDF) (TechRepublic)

In general, key-value stores are great as complements to other databases, which is why Amazon's DynamoDB incorporates both document and key-value stores, why MongoDB is paired with WiredTiger, etc. It's also why exactly no standalone companies of any size exist to support a key-value database.

Riak, then, was doomed from the day that MongoDB, DataStax (the company behind Apache Cassandra), Amazon, and others joined the fray of next-generation databases.

The business value

Even without such full-featured database alternatives, Riak also suffered from its single-vendor syndrome. With just one company behind it, and that company somewhat on the skids for years, Riak didn't see the kind of engineering investment needed to keep pace.

SEE: Has the time finally come for PostgreSQL? (ZDNet)

Redis, meanwhile, has a standalone company behind it, but has a real community of contributors from Pivotal, Mozilla, and more. As such, companies like Amazon have felt comfortable building commercial services around it, further driving its popularity. But, again, that popularity comes as a complement to other databases, and not as a single database offering from any sizeable vendor.

True enough, Georgetown Capital Partners and its lead partner Chester Davenport did Basho (and Riak) no favors, likely accelerating the company's demise. But let's not forget that they were accelerating Basho's demise, not causing it. The roots of that demise were actually in the all-too-niche database that Basho was trying to sell.

Also see

database.jpg
Image: iStockphoto/phongphan5922

About Matt Asay

Matt Asay is a veteran technology columnist who has written for CNET, ReadWrite, and other tech media. Asay has also held a variety of executive roles with leading mobile and big data software companies.

Editor's Picks

Free Newsletters, In your Inbox