Facebook, Google, LinkedIn, and Twitter share expertise to launch WebScaleSQL, a brilliant-DBMS built on Oracle’s MySQL Community Edition.

In a move which could shake up both the industrial and open source database management system markets, Facebook announced on Thursday that it has worked with fellow Internet giants Google, LinkedIn, and Twitter to develop WebScaleSQL, an open source, web-scale branch on top of Oracle’s publically available MySQL Community Edition.

“Our goal in launching WebScaleSQL is to enable the size-oriented members of the MySQL community to work more closely together which will prioritize the aspects which can be foremost to us,” wrote Facebook software engineer Steaphan Greene in a draft blog post shared with InformationWeek. Greene explained that the crowd desires to “help companies leverage the nice features already present in MySQL 5.6, while building and adding more features which are specific to deployments in large-scale environments.”

[Want more on Facebook’s contributions to high-scale hardware? Read Open-Source Hardware: Prepare For Disruption.]

Over the previous few months, the collaborators have built a community structure for sharing and reviewing WebScaleSQL contributions, and engineers from all four companies have contributed code and provided feedback to one another “to develop a brand new, more unified, and more collaborative branch of MySQL,” Greene wrote.

MySQL is the most well liked open source database management system (DBMS) on the earth, accounting for one quarter of all multi-client DBMS deployments, according Evans Data Research. In choosing to construct on MySQL, the WebScaleSQL community has spread out an avenue for thousands of MySQL developers to grow their deployments to unprecedented scale.

The move also poses a threat to certain NoSQL and NewSQL DBMS upstarts that experience gained a lot of their customers among organizations frustrated by the problem of managing MySQL at high scale. In keeping with Greene, the WebScaleSQL community has already developed components, including:

• Features that make operating at Web scale easier, equivalent to an ideal read-only feature and the flexibility to specify sub-second client timeouts.
• Changes to enhance the performance of WebScaleSQL, including buffer pool flushing improvements, optimizations to certain query types, and support for Non-Uniform Memory Architecture (NUMA) interleave policy.
• An automated framework that can run and publish the result of MySQL’s built-in test system for every proposed change.
• A suite of stress tests and a prototype automated performance testing system.

WebScaleSQL is a hard and fast of patches on top of Oracle’s publicly available MySQL Community release. The plan is to follow changes to MySQL Community Edition to make sure the newest features are available WebScaleSQL. Like MySQL, WebScaleSQL is offered under GPL licensing.

Conspicuously absent from WebScaleSQL’s announcement was any mention of Oracle, which owns MySQL and provides Enterprise and Cluster editions of MySQL aimed toward high-scale needs. WebScaleSQL is building on publically available Community Edition software. “So long as the MySQL community releases continue, we’re committed to remaining a branch — and never a fork — of MySQL,” a spokesperson told InformationWeek in an email interview.

WebScaleSQL will welcome other scale-oriented members of the MySQL community to contribute, in keeping with Greene. The gang is currently engaged on:

• Developing an asynchronous MySQL client that will not should wait to glue, send, or retrieve while querying. This non-blocking client developed and in production use at Facebook is being code-reviewed by the alternative WebScaleSQL teams.
• Adding Facebook production-tested compression and versions of table, user, and compression statistics.
• Adding Facebook’s Logical Read-Ahead mechanism for “as much as 10x” speed improvements in full table scans, akin to nightly logical backups.

The introduction of WebScaleSQL isn’t a direct threat to commercial or open source products within the enterprise market. For starters, Facebook, Google, LinkedIn, and Twitter are all self-supported IT shops, whereas smaller enterprises would need a commercial support provider. It is also unclear where “Web-scale” begins and whether MySQL Enterprise and MySQL Cluster Carrier Grade Edition could be more effective and value-effective products for current customers seeking to scale up. Commercial database vendors including IBM, Microsoft, and Oracle, meanwhile, might learn new tricks on easy methods to scale up their very own enterprise-oriented database products.

As for the NoSQL, and especially the NewSQL, crowd pursuing Web-scale deployments, WebScaleSQL will undoubtedly force them to tell apart their products beyond the baseline promise of offering performance at big data scale.

Improved platforms, standards, bandwidth rates, and information models mean more IT shops are taking a “cloud-first” solution to new services, keeping only select jobs in house. But what’s often unnoticed of the calculus is the impact at the end user once you outsource most services. Register for this InformationWeek editorial webinar and find out how to avoid the pitfalls of outsourcing IT. The Performance From The User’s Perspective webinar happens March 28. Registration is free.

Doug Henschen is Executive Editor of InformationWeek, where he covers the intersection of enterprise applications with information management, business intelligence, big data and analytics. He previously served as editor in chief of Intelligent Enterprise, editor in chief of … View Full Bio

More Insights