analysis At the recent MongoDB conference in New York, the company demonstrated its ambition to take on workloads from other databases.
The company has made a significant entry into the database market with a developer-friendly distributed document database that helps developers build modern, web-based transactional systems.
Time series and search have become targets, with promises of supporting secondary indexes in the former and search facets to help developers build search experiences faster in the latter.
But it was the continued push into analytics that impressed commentators, who also wanted to show the limits of what can be achieved with a document database.
Column Store Indexing, available later this year, will help developers create and maintain a purpose-built index that dramatically speeds up many common analytical queries without requiring changes to document structure or moving data to another system , the company said.
Speak with The registrySahir Azam, MongoDB Chief Product Officer, said developers are often forced to aggregate data in a third-party system and then bring it back into their database to operationalize complex analytical queries as part of their application.
“We added a number of features to the database and [DBaaS] Atlas to make it easier to enable in-app experience,” he said.
“We’re seeing much richer or smarter application experiences where what would normally be a human making a decision outside of a Tableau dashboard is now something that a development team automates in software.
“But these queries are often very different from what you would think of as a traditional type of transaction. They are much more like an analytical column query than a transactional query. So we’ve been working on performance improvements in our query engine, a new type of indexing called Column Store Indexing, which is all about improving performance for complex analytical queries so they can be embedded in the application experience,” he said.
Tests with synthetic data and real customer workloads have shown that MongoDB improves its performance by 5x to 200x on complex analytical queries, he said. Applications could include fraud analysis in financial services, the next best deal in e-commerce, or supply chain management, he said.
“They’re trying to ensure that database performance isn’t negatively impacted by analytics.”
Kimberly Wilkins, MongoDB technical lead at database consultancy Percona, said that developers have previously noticed negative performance impacts when running extensive analyzes in MongoDB – even when running in a separate node.
She said there have been significant improvements in synchronization capabilities and it also now allows larger parse counts in replica sets than other replica set numbers used for writes and infrequent reads. “That’s a great thing they were able to do. They try not to negatively impact database performance with analytics, so you can run your extensive analytics against MongoDB,” she said.
Still, developers and data architects building a cloud data warehouse like Snowflake or AWS Redshift for analytics alongside MongoDB are unlikely to change their minds because of MongoDB’s improvements. However, it could affect future decision-making, Wilkins said.
“If people start thinking they need a little bit of analytics, but have a document database that’s going to affect their writing and reading performance, that won’t be the case anymore if they get it right with MongoDB. It’s really, really impressive,” she said.
Tony Baer, director of analyst firm dbInsight, described MongoDB’s entry into analytics as “small steps” — enabling lightweight queries without impacting operational performance, with important usage limitations.
“The first tenet of operational databases is that you don’t want to slow them down. To do all that complex modeling that you would do in Databricks, or complex analysis that you would do in Snowflake, you really don’t want to burden the operational database with it, and that’s not what it’s designed for, although you can partition the load in it [MongoDB DBaaS] Atlas and have separate nodes. It’s meant for you to make a smart decision on the spot,” he said.
In his speech on SiliconAngle’s The Cube, he said similar ideas were behind Oracle’s move with MySQL Heatwave and Google’s AlloyDB.
Matt Aslett, VP and Head of Research at Ventana Research, said that cluster-to-cluster synchronization for synchronizing data across clouds and on-premises clusters and data federation for querying data across multiple clusters were among the key innovations from last year’s event Week included the momentum that MongoDB had maintained with developers of modern applications.
“The company has done a good job of collaborating with developers who create new web applications using the document model and JSON format, especially for web applications.
“Although much of the company’s early success was fueled by internet and application startups, it is gaining traction with established companies in industries such as financial services, insurance, healthcare and government, including adoption for workloads historically the domain of relational Databases were ,” he said The registry.
However, the company has managed to muddy the waters on what it supports commercially in the current release. “Some of the announcements related to generally available features, and some of them were previews of upcoming features. This allows users to start developing to take advantage of upcoming features, but the large number of announcements can create confusion as to whether individual features will be commercially supported or not,” Aslett said.
MongoDB has ambitious expansion plans and held a glamorous event in New York last week, although it continues to make losses. However, investor confidence is boosted by the company’s rapid growth, Aslett said.
“MongoDB’s revenue not only continues to grow, but quarterly year-over-year revenue growth accelerated in fiscal 2022. I would expect the company to retain investor confidence as long as it continues to meet or exceed expectations,” he said.
If it stays that way, it could be one of the few NoSQL startups with a vision to bring analytical and operational workloads closer together. ®