MongoDB 5.0 expands open source document database operations
At the MongoDB.live 2021 virtual conference, the vendor unveiled its next-generation database platform with more time series data support and live re-sharding capabilities.
MongoDB on July 13 unveiled the next major update of its database technology at the MongoDB.live 2021 virtual conference with the general availability of MongoDB 5.0 and a series of new features for the MongoDB Atlas Cloud database-as-a-service.
The open source MongoDB platform is primarily known as a document database, though in recent years it has moved toward enabling a multimodel database approach.
With MongoDB 5.0, the vendor, based in New York City, added to the database platform enhanced support for time-series data, enabling support for more applications, including internet of things.
MongoDB also advanced its MongoDB Atlas Cloud Database platform with improved search that enables administrators to optimize results as well as updated data lake capabilities that integrate MongoDB Charts data visualization.
The new set of capabilities and direction the vendor is taking got a largely positive review from Carl Olofson, research vice-president at IDC, who noted that the next step is for MongoDB to appeal more to enterprises.
By providing more advanced management capabilities that make it easier to run MongoDB at scale, as well as with data visualization for data lake data, MongoDB may well become a more attractive option for executive management.
"When we think of document database systems, we think of user-intimate applications including CRM, ecommerce, and gaming applications as well as session data management for apps on smart mobile devices," Olofson commented. "MongoDB has certainly excelled in these areas and won a huge amount of mindshare in the developer community, but their next big challenge is to offer more to the C-Suite."
MongoDB 5.0 features ease database management
Among the key new features that MongoDB 5.0 brings is a live re-sharding capability.
Sharding is a common process for databases -- distributing data across multiple nodes for more resilience and performance.
Mark Porter, CTO of MongoDB, explained that with the new re-sharding features, users can now adjust how sharding is configured, which to date has not been an easy process. With live re-sharding, Porter said a database administrator can set what they want the new shard key to be and MongoDB will shuffle all the data to match the new desired distribution.
MongoDB is also introducing what it refers to as a versioned API. When a developer builds an application that accesses MongoDB they often do so via an API.
Previously, the problem was that as each new version of MongoDB was released, there could be a required API change that would also force developers to also update their applications. The promise of the versioned API is that developers can stick with the same API, even as the underlying MongoDB database is upgraded to new releases.
"With the versioned API we will guarantee you that your app will keep working year after year," Porter said. "What we're trying to do is help customers get on the leading edge of software so that they get all the features all the security fixes just naturally and fluidly, without forcing all of their application development teams to be disrupted by upgrades."
How MongoDB 5.0 enables a multimodel approach
The addition of native time series data support on MongoDB 5.0 isn't the first time that time-series data has been used with the open source database.
"People have been using MongoDB for time series data forever, in fact it's a common workload on MongoDB, " Porter said.
Porter added that MongoDB 5.0 enhance support for time-series data with automated clustered indexes as well as faster data ingestion. He said the document database model that MongoDB is built on is already a multi-model platform that can support different types of data including key value stores, graph data and even relational data.
"You can take most relational workloads, and you can throw them into a document model, and they work just fine," Porter said. "Now, I'm not going to say that we're the perfect database for all of these things and all of the edge use cases, but we believe that we are a general purpose database that you can run graph workloads, time series workloads, key value workloads and OLTP [online transaction processing] workloads, all in one data store."
