Cockroach Labs updated its namesake CockroachDB distributed SQL database with new capabilities that enable more control over global deployments and data access.
The new version, CockroachDB 21.1, became generally available on May 18 and is the first release from Cockroach Labs since the vendor made public a $160 million funding round on Jan 12.
With the new CockroachDB version, a key focus is enabling multi-region cloud availability, so administrators have table- and row-level control over where data is placed in cloud deployments around the world.
Matt Aslett, research director for data, AI and analytics at S&P Global Market Intelligence, said Cockroach Labs is now well established in the database space and is forging new ground in distributed data processing.
Aslett noted that the vendor offers both CockroachDB for self-hosted deployment on cloud infrastructure or more rarely on premises, as well as the CockroachCloud managed service. He added that CockroachCloud has some well-known brands as customers, including Baidu, Bose, Comcast, DoorDash, eBay, Lush Retail and SpaceX.
While Cockroach Labs has found success with those brands, Aslett noted that not every organization requires the capabilities of a distributed SQL database. Only slightly more than a quarter of the respondents to a recent survey by research firm 451 Research said they have a globally distributed database in production.
"It could fairly be argued that the extremes of what is possible with distributed SQL databases serve relatively niche requirements compared with the many established relational database products and cloud services," Aslett said.
Aslett noted that his firm has found that while few vendors actually have a globally distributed database, supporting globally distributed data processing -- including the ability to control data residency to meet regulatory obligations -- is a growing requirement for modern applications.
Improved multi-region support
Cockroach Labs co-founder and CEO Spencer Kimball explained that the new multi-region capability in the new CockroachDB version is a feature the vendor has been working on for five of the six years it has existed.
Cloud providers have different regions -- physically separate, geographically dispersed data center locations. Within each region a cloud provider typically also has availability zones, which are additional segments within a region that are provisioned separated to enable high availability.
Many modern cloud databases enable deployment across multiple availability zones. Deployment of a synchronous, highly available database across multiple regions, however, is another issue, Kimball said.
There are several key reasons why any organization would want or need to have a multi-region database capability.
The first is for disaster recovery if an entire cloud region were to go offline, providing a path for business continuity. Having data in multiple regions can also help with speed, because data is physically located closer to end users.
The other key issue is compliance with regulatory requirements that mandate that data for a specific group needs to remain within a specific geographic area.
How CockroachDB 21.1 enables distributed SQL global deployment
With the new database version, Cockroach has simplified how developers configure multi-region database deployment with different table types.
Database administrators can now define different table types that enable the multi-region database deployment. The simplest table type is the regional table, when the entire table is going to just be stored in a single region, Kimball explained. Going a step further, it's now also possible to lock a specific row within a table to a specific region.
For high availability applications, CockroachDB 21.1 has a global table type that is replicated and available across all the regions an organization is operating in.
"The sort of sweat equity that's been put into the new CockroachDB 21.1 release isn't just about reducing the complexity," Kimball said. "It's about adding that capability with global tables that can actually deliver consistent reads that are very low latency, globally around in all the regions that your clusters currently operating in."