PingCap is aiming to make it easier for organizations to use the open source TiDB distributed hybrid transactional and analytical processing database with the general availability of the TiDB Cloud service.
As an HTAP database, TiDB combines functionality that is commonly found in online transaction processing and online analytical processing databases in a single platform. TiDB also provides distributed SQL capabilities and is intended to be compatible with the MySQL database.
Founded in 2015 and based in San Mateo, Calif., PingCap provides commercial support and services for TiDB. With the new TiDB Cloud, the vendor is now providing users with a fully managed database-as-a-service option for deployment.
While PingCap refers to the TiDB database as HTAP, Forrester Research refers to the combination of analytics and transactional capabilities as a translytical database.
Noel YuhannaAnalyst, Forrester Research
"Traditionally, we have kept different databases for various use cases, such as transactional, operational and analytical workloads," said Noel Yuhanna, analyst at Forrester. "Translytical databases such as TiDB can enable all of these workloads in a single database platform, eliminating the need to move data around to support real-time analytics and operational insights."
TiDB Cloud enables HTAP and distributed SQL
Ed Huang, co-founder and CTO of PingCap, said the TiDB project was originally inspired by the Google Cloud Spanner distributed SQL cloud database.
The idea of a distributed SQL database has also inspired other startups, including CockroachDB from Cockroach Labs, as well as Yugabyte and its YugabyteDB database, which are both compatible with the open source PostgreSQL database.
With TiDB, Huang said that instead of PostgreSQL, he and his co-founders chose to be compatible with MySQL, as it was the database he was most familiar with as a user.
"Users can just use TiDB like MySQL, but it is scalable," Huang said. "The scale-out process is totally transparent to the application, which helps to reduce maintenance as users don't need to know about the complexity of maintaining clusters in a large distributed system."
The TiDB database stores data in both row and columnar database formats. Huang said the row format is well suited for transactional operations; columnar, by contrast, is a primary approach of analytics databases.
Cloud service aims to ease TiDB adoption
The TiDB Cloud, unveiled May 10, uses the same open source database that is available on premises as the foundation for the service.
Huang said the vendor is planning on adding capabilities over time to make the cloud service more elastic, by using cloud-native technologies such as microservices.
He noted that for on-premises deployments, users have to manage all the components of delivery, including the operating system, hardware and configuration. With the cloud service, that complexity is managed by PingCap.
"The whole point of our cloud is that we can do a lot of things to create a better user experience," Huang said.
TiDB 6.1 is coming
The TiDB Cloud is based on TiDB version 5.2, a production-ready release. The latest version of the open source project is version 6.0, an experimental release with new features that will be considered production-ready in version 6.1 later this year.
Huang said TiDB 6.1 will bring new features for cloud and on-premises users. One feature he highlighted is a data placement capability that will enable users to set policies on where data is stored across different data centers or cloud deployments.