Last week Confluent announced a "cloud-native" Apache Kafka service, with elastic scaling and consumption-based pricing. Today, at its Accelerate event near Washington, DC, DataStax is making an analogous announcement around Apache Cassandra. It's also introducing a suite of tools for monitoring and optimizing the performance of Cassandra clusters.
Also read: Confluent makes Apache Kafka cloud-native
In a briefing with ZDNet, Robin Schumacher, DataStax Chief Product Officer and SVP, explained what the new offering, to be called DataStax Constellation, will include. Constellation will, first and foremost, include Cassandra as a Service, featuring elastic scaling and consumption-based pricing. Schumacher explained that the service will be based on Apache Cassandra but will include elements of the DataStax Enterprise (DSE) distribution, including certain security features; self-healing capabilities, like node sync and traffic control; as well as 2x-3x improved performance.
Constellation will also include DataStax Insights which, in turn, will consist of Recommendation Insights, Application Insights and Performance Insights. The three sub-components will identify cluster and query issues and bottlenecks; Recommendation Insights will provide AI-driven recommendations to solve cluster issues. DataStax Insights will work in a hybrid cloud/multi-cloud manner, compatible with both Constellation's Cassandra as a Service as well as DSE. Given the DSE compatibility, DataStax Insights will also supercede DataStax OpsCenter.
Schumacher said that DataStax views the Cassandra market and ecosystem through a temporal lens, as a series of eras. First came the era of making Cassandra work with top-notch performance and functionality; next came the era of evangelizing Cassandra and expanding its popularity; following that came a focus on providing customer value through delivery of proprietary functionality in DSE. Schumacher says DataStax Constellation represents a fourth Cassandra era: making the platform easy and obvious, with "push-button ease" for hybrid/multi-cloud implementations.
That seems a direct response to first-party database services form the public cloud providers, which often provide Constellation-like deployment, scaling and pricing options by default. Microsoft's Cosmos DB does this and provides a Cassandra API interface that allows the service to compete with Cassandra, be it DataStax's distributions or the vanilla Apache bits. It accommodates numerous other APIs, works in a geographically distributed fashion by default and, in functionality announced two weeks ago at Microsoft's Build conference, integrates Apache Spark and Jupyter notebooks.
Maybe this competitive front is what inspired DataStax to point out to me that it believes Constellation's on-demand pricing to be 10 times less expensive Cosmos DB's. While I'm never sure what it means to be a multiple less than something, it's clear that DataStax believes Constellation will be competitive on ease of deployment and ease of operation, while beating Cosmos DB on price.
One database, multiple clouds
As in other platform battles, we see two sides -- in this case, the public cloud providers and data/analytics software vendors -- trying to commoditize the other. So while Cosmos DB offers numerous database APIs, exclusively on Microsoft Azure, Constellation will offer a single database that will eventually work across multiple public clouds.
Specifically, DataStax says the service will launch on Google Cloud Platform in the fourth calendar quarter of this year, and that availability on Amazon Web Services and Azure will follow. It will also offer an early access program this summer.