MongoDB Inc. leaders at MongoDB World 2015 last week focused on enhancements due with the next edition of their...
leading NoSQL data platform, but just as much attention was devoted to the recent MongoDB 3.0 release, which introduces swappable data storage engines to the company's platform.
MongoDB 3.0, launched in February, was the first such release to include the WiredTiger storage engine. That software was obtained along with its namesake company, WiredTiger Inc., late last year. This store improves MongoDB performance, which for some parallel applications is a marked upgrade from the company's original default MMapv1 engine.
At the New York City event, the company said WiredTiger is now its default storage engine. Company strategy is focused on supporting the ability to switch storage engines depending on user needs, said Kelly Stirman, vice president for strategy and product marketing at MongoDB.
He noted the availability of the Facebook-backed RocksDB engine -- a write-optimized, persistent key-value store -- for Mongo DB. Also, an upcoming encrypted data at rest version of WiredTiger was discussed at the conference.
Yet another MongoDB plug-in takes the shape of the "fractal tree indexing" TokuMX storage engine, developed by vendor Tokutek, which was acquired earlier this year by open-source database provider Percona LLC.
The role of MongoDB data stores going forward will be "very much geared to the type of workload," said Tony Baer, principal analyst at Ovum. "With pluggable data store support, MongoDB potentially widened its addressable market and ecosystem for different types of jobs."
Riding the WiredTiger
Baer noted this growing body of MongoDB storage engines has precedence in MySQL, the widely-used open-source database system that can employ InnoDB, NDBCluster, Blackhole and other engines.
Different types can address different application needs, and those needs met by WiredTiger could prove important to big data NoSQL wunderkind MongoDB.
Ray Smith, CEO and co-founder, Datahug Inc.
Among improvements to MongoDB, the biggest one for Ray Smith is the WiredTiger storage engine. On hand at MongoDB World, Smith, CEO and co-founder of Dublin, Ireland- and San Francisco-based Datahug Inc., said "moving to the new storage engine allows document-level write-locking as opposed to collection level locking."
Locking, as found in implementations based on MongoDB's former default MMapv1 engine, has hurt performance for some jobs. MongoDB's data engine swap will "allow faster parallel processing" for Datahug, according to Smith, whose firm offers cloud-based services that deliver historical analysis of customer relationship management and e-mail header data to help companies' sales teams pinpoint optimal sales prospects.
MongoDB 3.2 update moves into the enterprise
With its released scheduled later this year, MongoDB 3.2 carries forward the vendor's efforts to go deeper into the enterprise application space. These capabilities include a new connector for BI and visualization tools offered by Tableau, Qlik, SAP and IBM. The connector allows MongoDB's aggregation framework to process analytics data within the database.
Among other capabilities due in 3.2 are:
- A new graphical interface to MongoDB that is code-named MongoScout. This software gives a view into NoSQL data structures.
- The debut of document validation rules, which give the enterprise architect more sway in database design. This feature has been desired -- and in some cases, required -- by enterprises both somewhat eager and somewhat reluctant to embrace NoSQL, which arrives with development flexibility but loss of data model control.
The "go to work" ethic was noticeable at this year's MongoDB convocation, the first to be led by new CEO Dev Ittycheria, who left the ranks of venture capitalists to guide the heavily funded NoSQL database star in its next stages of growth. The company has transitioned from the "entrepreneur" stage to the "build a stronger business" stage, said Baer.