Couchbase adds support for multi-statement SQL transactions

Let the OSS Enterprise newsletter guide your open source journey! Sign up here.

Couchbase today delivered a major update to its namesake database that adds support for the multi-statement SQL transactions driving transaction processing (OLTP) applications.

Version 7.0 of Couchbase extends existing support for SQL into the realm of applications that were mainly the preserve of relational databases. Now IT teams can process high-performance transactions alongside analytics queries that can run across large quantities of unstructured data on a document database, said Couchbase CTO Ravi Mayuram.

IT organizations can for the first time run multi-document SQL ACID transactions in microseconds, Mayuram said. The goal is to reduce the number of databases organizations need to deploy and maintain to support different classes of application workloads, he added.

Beyond the multi-statement SQL

Couchbase Server 7.0 adds schema and table-like organizing structures, called Scopes and Collections, within the document database. IT teams can now add a table (Collection) while transactions are running without having to add or modify the schema (Scope) or take the database offline. Via what the company refers to as a dynamic data containment model, it also becomes easier to migrate existing transaction processing applications to Couchbase 7.0. “We now have one-to-one mapping,” Mayuram said.

The company has also added support for partitioning and index isolation to boost overall performance, along with a configurable backup service that indexes data in a way that’s more portable.

Finally, Couchbase has added a cost-based query optimizer to its query service that replaces a previous rules-based approach.

Fresh off becoming a public company this month, Couchbase traces its lineage back to an open source project and data protocol called Memcached that organizations employed to make data available in memory. Memcached was then combined with the JSON data format and SQL to create the Couchbase database. Couchbase today claims to have more than 500 enterprise customers, including American Greetings, Domino’s, Intuit, Experian, Western Union, USAA, Avis, Carnival, and Emirates. Couchbase claims 30% of the companies listed on the Fortune 100 are among its customers.

Why you should care

Competition among providers of document databases is fierce. Most document databases initially found their way into enterprise IT organizations through developers, who typically downloaded an open source platform they could provision themselves without having to ask for permission or money. However, as the number of applications based on document databases proliferates across the enterprise, it becomes apparent IT operations teams will assume more responsibility for managing them as developers move on to other projects.

A database administrator (DBA) may make a case for refactoring an application to run on a relational database platform that an IT operations team already manages in a production environment. Most relational databases have added support for JSON to make it easier to migrate those applications. Couchbase is effectively countering that argument by making it possible to run multiple types of SQL applications on a document database as part of an effort to reduce the total cost of database management using platforms that are considerably less expensive to license than a commercial relational database.

It’s not clear how many IT teams are going to migrate from one database platform to another, due to the complexity involved in such efforts.

However, new application deployments make it more feasible than ever to run multiple classes of applications on the same database platform. In many cases, transaction processing applications are generating a lot of the data that might be fed into an analytics application that incorporates structured, semi-structured, and unstructured data. The challenge now is determining which of those databases can run all those classes of applications without compromising performance.

Source link