Today, InfluxData is releasing the managed cloud service to support its 2.0 platform. The managed service, initially going live in AWS’s US West Oregon data centers, is not exactly a surprise, as it was one of the promises made when the company unveiled the 2.0 platform roadmap last spring. And it’s not the first time that InfluxData has offered a cloud service, but this one is a much different creature.
Unlike InfluxData’s original cloud service, the new one will be serverless. It’s actually the first specialized time series database cloud service out of the gate to do so. Customers pay for use of compute, network and storage; in an upcoming release, InfluxData will add the ability to put rate limits so customers can put upper limits on monthly costs. It will be based on the recently-announced Influx DB 2.0 platform that is still in alpha. By contrast, Amazon, which announced its own Timestream time series managed serverless database service at re:Invent last year, has not yet entered preview.
InfluxData is one of an emerging breed of time series database providers. Since its 2013 release, it has become arguably the most popular time series database with hundreds of thousands of implementations and several hundred customers. And it did offer a cloud service for its original InfluxDB 1.0 platform, but it only encompassed the core database and was not a serverless offering.
But, with the 2.0 release, the company is engaging in a riverboat gamble, changing the APIs and query language as part of a plan to unify the entire stack behind a single, common set of APIs. It’s a common story amongst open source projects, where the 2.0 refactors and rationalizes APIs; that’s previously happened to Spark and Hadoop, amongst others.
The design direction of the 2.0 platform is to make it more visual and, in some cases, codeless (although through APIs, InfluxData will still allow developers to write more sophisticated scripts through the command line). The crux of the simplification strategy is Flux, the new query language that is highly visual. It is part of an offering that also bundles data collection and stream processing along with the GUI and underlying database. By contrast, with AWS Timestream, you would also have to activate adjoining services including Lambda for developing background processing functions, Kinesis Firehose for landing streaming data, and QuickSight or third party tools for visualization.
While the company is hardly taking eyes off its installed base, it is looking to a wider base by simplifying and unifying the platform. The new managed cloud service being released today is the cornerstone of that strategy. The serverless nature of the new cloud service, in conjunction with the visual development environment and unified APIs, should broaden the target audience. But it will be a more developer-oriented audience; by contrast, rivals such as Timescale, are targeting the classic SQL developer base.
Time series use cases are potentially quite broad, addressing almost any process that requires real-time processing. Examples include IoT, server monitoring, weather and climate monitoring, stock tick feeds, and so on. In some cases, these use cases have drawn specialized platforms that in some cases are handled by time series databases, such as Kx, for real-time capital market feeds. Other cases, such as server monitoring, have been addressed by search-based data platforms such as Splunk or Elastic.
InfluxDB instead is positioning itself as an easy to use, general-purpose time series application development engine that can be molded to the desired use case. For now, the focus on Influx Cloud 2.0 is drawing new users to a relaunch of the entire InfluxDB platform; in time, we expect that the company will look to third parties to go that last mile. That’s where it’s budding relationship with Google Cloud Platform (Influx Data was one of the charter members of GCP’s open source database partner initiative) could provide an ace in the hole. Because Google itself has not made any noises about developing its own time series database, it could be motivated to through joint marketing to make InfluxDB a more compelling draw to prospective third party partners.