Luiz - Fotolia

Cockroach Labs looks to accelerate cloud-native database

Spencer Kimball, co-founder and CEO of Cockroach Labs, talks about the distributed SQL database vendor and its future, with new financing and a new cloud-native database release.

Cockroach Labs said Tuesday that its CockroachDB 20.1 update is available now to users, providing improved PostgreSQL compatibility as well as new security and multimodel database features. The new release comes after the 2015 startup made public an $86.6 million Series D round of funding on May 5.

Building a new relational database platform is challenging, but it's an endeavor that Cockroach Labs has been tackling since its inception.

The vendor's flagship CockroachDB distributed SQL database works inside cloud-native deployments to help organizations scale global workloads. While CockroachDB is a newer database, it has increasingly added support to enable compatibility with the widely deployed open source PostgreSQL database.

In this Q&A, Spencer Kimball, co-founder and CEO of Cockroach Labs, outlines the direction his company is taking and where the market for databases is headed in the future.

Why did you decide to raise more money now?

Spencer KimballSpencer Kimball

Spencer Kimball: Given the potential for the macroeconomic situation worsening with COVID-19, that certainly accelerated our interest in doing this round of funding.

If you step take a step back, we actually decided to raise this money before COVID-19 became the obvious problem it is now. We actually strategically took time to think about what we need to do in the next two years, and we wanted to have the capital to do that.

What impact has COVID-19 had on Cockroach Labs, and how do you see it affecting how enterprises use databases?

Kimball: It's clear that COVID-19 is an accelerant for all kinds of businesses moving into the cloud more aggressively. With everyone staying at home there is a dawning realization, which was already underway but is accelerated by COVID-19, that organizations want to benefit from other people's economies of scale and expertise. I think COVID-19 is sort of underscoring the reason and opportunity for more efficient operations and having someone else running the database for you.

It's clear that COVID-19 is an accelerant for all kinds of businesses moving into the cloud more aggressively.
Spencer KimballCo-founder and CEO, Cockroach Labs

Another factor here that we've seen come into play dramatically is that certain businesses are seeing incredible surges in demand: things like delivery services, entertainment and gaming. When something like COVID happens, there's a huge change in the dynamics of how to scale a service and that has really put a lot of people's back up against the wall.

There were some severe outages for some of our prospects, which accelerated them making the decision to move. CockroachDB is very well suited to these kinds of dynamic surges in demand, you can really just give it additional hardware to scale up very dramatically.

CockroachDB 20.1 provides improved PostgreSQL. Are you aiming for feature parity?

Kimball: We've implemented more features to almost be a drop-in replacement for postgreSQL and that really matters for developers. We work with virtually every schema and all of the standard SQL queries. But PostgreSQL has a lot of extensions. There's a lots of bells and whistles that PostgreSQL has acquired over 30 years.

So I think the right way to think of it is that we're always aiming to approach PostgreSQL 100% compatibility, though honestly, I don't think we'll ever get there. There are things in PostgreSQL that probably shouldn't be reimplemented. For example, we don't have stored procedures as a feature, and it might not be one that we ever end up doing because many application developers are moving away from that.

Where do you see the trend of multimodel databases headed and is it an approach that you might want to take with CockroachDB?

Kimball: We do have some multimodel capabilities. For example, in the new 20.1 release we support a feature called recursive common table expressions, which is typically used if you want to have a relational SQL database do graph queries.

CockroachDB is really not a graph database at heart and it's not optimized to do graph queries the way that a true graph database would be. But it kind of expands CockroachDB's capabilities to satisfy a marginal graph use case.

We already had JSON [JavaScript Object Notation] support in CockroachDB, which means that you can declare a schema that essentially makes CockroachDB very similar to a document store.

But I can't stress enough that the market for OLTP [online transaction processing] relational databases is so massive and it has incredible opportunities. We're never going to be pushing as hard as we could on multimodal aspects, because there's such an unbelievably ripe and gigantic market in the relational space.

How do you see security features evolving for CockroachDB?

Kimball: With security, there's just a mountain to climb there. We're finding that there's lots of different features that are requirements when you get into bigger enterprise accounts. In the world of other databases like PostgreSQL, they get added by various vendors over time, over decades, even.

One of the things we've done in 20.1, is our backup/restore capability wasn't encrypting backups, but that has now changed. It was a feature that was requested by some of our customers to implement. We're implementing security features that are necessary to fit into the SQL database category which has a very big footprint.

With Cockroach Cloud you have a database-as-a-service capability. Is that a general direction and trend that you see for the overall market going forward?

Kimball: Our cloud business is growing rapidly, but revenue is still quite small. We have really big customers, and those big customers, they use CockroachDB in a self-hosted fashion. Based on my experience, the bigger a company is, the more that they are interested in self-hosting but that is now changing.

The big companies that I'm aware of, in three to five years, in almost all cases are moving into the public cloud. I think in three to five years, it will be true of all kinds of infrastructure services and certainly database services, that they run in the cloud. It's going to be hybrid for a while, that's just reality, but bursting to the public cloud? The writing is already on the wall.

Dig Deeper on Database management