Aerospike Connect integrates streaming data, AI and legacy systems

Aerospike, makers of the in-memory NoSQL database of the same name, which has found success in Internet of Things (IoT) and edge applications, is upgrading and expanding its Aerospike Connect product line. This follows hot on the heels of the May release of its new Aerospike Database 5 version. 

Back in March of 2019, Aerospike announced Connect for Spark and Connect for Kafka. Now it's enhancing those two offerings and adding support for JMS 1.1 and Apache Pulsar 1.0. These upgrades and new connectors mean that Aerospike Connect will now integrate legacy systems, including mainframes, and expand the range of streaming data platforms and use cases it supports.

Also read: Aerospike, ThoughtSpot, Alteryx and AI-inspired integration

What's inside

ZDNet sat down with Aerospike founder and Chief Product Officer Srini Srinivasan, who explained what the enhancements to the Aerospike Connect offering enable. Srinivasan began by explaining that Aerospike Connect for Spark 2.4 supports streaming APIs that leverage Spark Structured Streaming. When combined with Aerospike's connectivity with Apache Spark's DataFrames, as well as Spark's data engineering and machine learning capabilities, Aerospike can drive new real-time AI use cases fed by streaming data.

Aerospike Connect for Kafka 3.0 will target the optimized protocol used by Aerospike Cross Datacenter Replication (XDR) in Aerospike Database 5, resulting in better performance and integration with Aerospike change notifications. Speaking of which, Aerospike Connect for Pulsar 1.0 offers an outbound connector to shipchange notifications to an Apache Pulsar consumer, which Aerospike says will facilitate streaming pipelines for log analysis, IoT and other applications.

And lest we think all of this is about new technology, Srinivasan explained that Aerospike Connect for JMS 1.1 brings in legacy systems, too. That's because JMS (Java Message Service) is often the preferred option when integrating and synchronizing data with mainframe applications. And since the JMS support can stream data both into and out of Aerospike, the database can act as a hub, bridging old and new systems, bi-directionally.

Silos be gone

The trend toward in-database machine learning, streaming data capabilities and integration with open source analytics technology is becoming prevalent throughout the industry, which might make Aerospike look as if it's merely jumping on that bandwagon. But bringing in-memory NoSQL processing together with multiple streaming technologies, mainframe systems and machine learning applications, means Aerospike's move doesn't just achieve parity, but rather adds value.

Integrating legacy, streaming and batch platforms, as well as IoT/Edge and systems of record, means these Aerospike Connect enhancements could bust a lot of data silos, across use cases, platforms and technology generations.