ServiceNow Hastens Database Ops on Its Now Platform
The company has recently announced its intent to acquire Berlin-based Swarm64, a maker of database acceleration technology, in an effort to provide increasingly larger and more intelligent workflows for its customers.
Swarm64 develops Swarm64 DA, a PostgreSQL database extension that accelerates SQL query performance and simplifies scaling, said Joe Davis, senior vice president of Platform Engineering at ServiceNow, in a blog post. That extension significantly improves the performance and scalability of a standard Postgres database.
ServiceNow uses PostgreSQL as one of its platform databases today and intends to use it more broadly moving forward, Davis said.
“Swarm64’s expertise in database performance and scalability allows us to plan for the future by helping us deliver increasingly larger and more intelligent workflows for our customers,” Davis told The New Stack. “The capabilities acquired with Swarm64 are important to the future scale and performance of the Now Platform — i.e., leveraging CPUs with massive growth in the number of cores for performance. The current platform capabilities have and will continue to support the current and immediate needs of our customers.”
Swarm64’s technology focuses on improvements in two key areas: one, adding a hybrid transaction analytic processing (HTAP) capability via a columnar index that supports hybrid analytical and transactional workloads and, two, enabling extremely high performance and scale by improving query optimizer and parallel execution by leveraging the aforementioned massive number of cores for parallel processing, Davis said in an interview.
“Swarm64’s open-source-based database technology works extremely well for large scale datasets; we plan to integrate Swarm64’s open-source-based database technology into our core data platform to provide a seamless addition to our data processing capabilities and APIs,” he said.
While ServiceNow claims there is not another company that does exactly what Swarm64 does — that is, provide a Postgres extension for both HTAP and parallel execution improvements to significantly increase performance and scale, “There are other companies who provide extensions on Postgres and have different architectures and solve slightly different problems, but they are in a different domain; we aren’t competing against those companies and rather are adding this as a technology to our core platform to support our industry-leading workflows.”
Another company building atop PostgreSQL is Timescale, which offers a time-series SQL database providing fast analytics and scalability with automated data management.
“There are many interesting projects on top of Postgres,” said Prashant Sridharan, vice president of marketing at Timescale. “Obviously, TimescaleDB is one of them. Microsoft bought another one a while back (Citus Data). It’s sort of the beauty of Postgres: it’s a platform that enables people to build many different types of specialized databases on top of it. We’re kind of in the era of specialized databases and choosing a Postgres-based option enables you to tackle new problems while maintaining compatibility with the rest of the business data.”
Digitally transformed organizations require a hybrid database model that leverages both transactions and analytics in order to deliver intelligent workflows.
“I am biased because of my SQL Server background, but, basically, I look at MySQL and Postgres as commodity databases that have a breadth of functionality, but which are not as optimized for performance as the commercial RDBMS platforms,” said Andrew Brust, CEO of Blue Badge Insights, a NY-based IT consultancy specializing in BI, data analytics and database wrangling.
As a result, Brust said, there are various independent solutions out there to create enterprise distributions for them and/or accelerate their performance, “because there is plenty of room for improvement,” he noted.
Moreover, “If a company like ServiceNow has standardized on Postgres, then obviously they’re going to benefit from acceleration technology as part of their stack,” Brust said. “And it probably made sense just to buy a company instead of continually licensing its product and having to worry that it could get acquired by someone else that might merely incorporate their platform instead of continuing to sell and support it as a discrete offering.”
With the addition of the Swarm64 technology, ServiceNow will help customers manage data across many different use cases to execute complex, high-speed data analytics at a mass scale.
Users expect and demand fast user experiences and managing large datasets is critical to this.
“With Swarm64, ServiceNow customers can query more data sources faster than ever,” Davis said. “We are continuing to bring consumer-like experiences to the enterprise. Swarm64 offers a combination of the analytical and transactional database capabilities that are needed to not only support intelligent workflows, but also support best-in-class performance and scalability for our customers. We expect that Swarm64’s technology will better enable customers across industries to quickly scale large workflows on the Now Platform.”
For example, telecommunications organizations are experiencing explosive growth in the amount of network assets — switches, routers, devices, and everything dealing with distributing telco networks and bandwidth — which comes with a need to store and query massive amounts of consolidated data. Effectively leveraging this data is fundamental to inventing new revenue streams, creating highly personalized experiences for customers and employees, and unlocking productivity across the enterprise, Davis said.
The acquisition is expected to be completed in the third quarter of this year. And ServiceNow will “natively re-platform” the Swarm64 technology so that ServiceNow developers, customers, and partners will be able to tap into the technology without having to refactor their applications or make any changes to their workflows, Davis said.
“The platform will transparently take advantage of this technology for the appropriate workloads dynamically,” he said.