This article is more than 1 year old
MongoDB 4.4 aims to be a dev crowd-pleaser, but analysts say it's still short of 'general-purpose' database territory
New features for core product as well as enhancements to Atlas DBaaS
Against a backdrop of mounting losses, NoSQL document store database MongoDB has pushed out its 4.4 iteration with a slew of new features that it expects to improve analytics, ease scaling and smooth performance.
At the same time, the database-wrangler reckons it has sped up query access to data across current and historical data sets in its DBaaS product, Atlas, which now accounts for 42 per cent of total sales.
Analysts told The Reg that while the array of new features would appeal to developers, MongoDB still struggles to substantiate its claims it will become the "general-purpose" database suitable for a broad class of enterprise use cases.
MongoDB 4.4 includes the Union feature, an operator in the MongoDB query language MQL.
Chief product officer Sahir Azam told us the idea was to offer the ability to run analytical queries on the database, instead of spinning analytical data out to data lakes or warehouses.
"It cuts out all of that sort of ETL process and allows you to have that analytical process, run directly in the core database," he said.
MongoDB is a distributed database, and administrators have to make decisions about sharding it at the offset, which might not make sense as the application grows. Another new feature sets out to redefine shard keys and modify data locations at any time as the business requirements change, without taking down and rebuilding the database.
With performance in mind, the hedged reads feature selects the best performing nodes from which to retrieve data, Azam said. "We submit read requests to multiple nodes in that cluster simultaneously and automatically retrieve results from the node that has the lowest latency. If there's a problem with a portion of the cluster, the database is already optimising for the fastest response time."
MongoDB said it has also added features to Atlas, including the general availability of its Atlas Data Lake, which was announced in beta last year. Azam said it will let users create federated queries to run on multiple data sources, both offline data lakes and transactional systems.
Stephen O'Grady, principal analyst with RedMonk, said the feature would fulfil a need organisations have for a single layer of abstraction to address multiple back-end data stores.
"As the number and type of databases and data sources incorporated into modern applications has exploded over the last decade," he said, "the challenge for developers has been scaling their understanding of the multitude of interfaces. This is the opportunity that MongoDB is built for."
In 2019, MongoDB bought open-source mobile database and synchronization platform Realm.io. In the new release of its cloud platform, MongoDB Realm has introduced Realm Sync to synchronise data both ways from mobile client and Atlas cloud database.
Carl Olofson, research vice president for data management software at IDC, said the addition of mobile support was important since a large percentage of MongoDB instances were used to manage data on the edge.
As MongoDB's sales model was focused on developers, there was much to please that crowd in the new release, he said, but added that MongoDB was still some way off realising its claim to be a "modern general-purpose database platform."
"I am not sure what 'modern' means in this context. Every few years, a vendor comes out with something, calling it 'modern'. I remember when relational was 'modern'. I remember when object-oriented was 'modern'. Some are saying that graph databases are 'modern'."
MongoDB sits among a family of NoSQL or non-relational databases and is gaining popularity for product catalogues, real-time data integration and internet of things – anywhere where scale and flexibility are an advantage. But it is unlikely to challenge relational database systems, which have a greater depth of functionality for transactional and analytical enterprise systems.
"In cases where the data model is fairly stable… there will always be relational," Olofson said. "On the other hand, there are plenty of cases where the sophisticated query of operational data is not required, and where the need is greater for speed, simplicity, and the agility of schema-less operation. Document databases fit the best in those spaces."
According to recently published results for MongoDB, Q1 revenue for the three months ended 30 April bounced 46 per cent to $130.3m, with the Atlas segment growing 75 per cent to account for 42 per cent of total revenue. Expenses also went up by 46 per cent to $134.69m, leaving the company nursing a net loss of $54m compared to a net loss of $33.2m a year earlier. ®