Rockset enables real-time analytics for Oracle users

Oracle makes industrial-strength databases, servers and storage devices that are among the fastest and toughest in the business, but they don’t necessarily perform all the functions users are now seeking. 

Enter six-year-old startup Rockset, which has developed a real-time analytics platform for cloud apps, including those from Oracle. The San Mateo, Calif.-based startup announced on June 1 a new integration with the giant database company, one that it claims enables developers to run real-time search, aggregations and “joins” on data from Oracle databases.

A growing number of enterprises are investing in real-time analytics as a foundation for their digital upgrades, particularly in areas such as ecommerce, logistics and delivery tracking, fraud detection systems, health and fitness trackers and recommendation engines. 

Oracle databases, while becoming more user-friendly in recent years as the company has moved its basis of operations to the cloud, have never been easy to operate and maintain. Using analytics on the databases has been a particularly challenging set of duties for a long time.

“Enterprises looking to use data from their Oracle databases to power user-facing analytics have traditionally relied on read replicas that are too expensive, or loaded it into warehouses that are too slow,” Rockset Chief Product Officer Shruti Bhat said in a media advisory. “Rockset now ingests real-time change data capture (CDC) streams from Oracle, and indexes every field to enable low-latency, high-concurrency analytics, without the complexity of traditional approaches.”

As the IT world moves from batch to real-time analytics, Bhat said this release enables developers using Oracle to: 

  • Run sub-second analytical queries, including joins with other databases, lakes, or event streams;
  • Achieve better database performance by isolating analytical queries on Rockset, which in turn scales horizontally in the cloud; and
  • Power fast microservices using developer SDKs or deliver real-time reporting using one of Rockset’s integrations with visualization tools, including Tableau, Retool, Redash and Superset.

Rockset CEO Venkat Venkataramani told VentureBeat that these connections aren’t supplied by APIs but are fully managed cloud-service entities.

“So, let’s say you’re managing your system of record database, which happens to be MySQL, Postgres, MongoDB, or what have you,” Venkataramani said. “You create an account with Rockset. You point us at your MongoDB database, and we have built-in connectors that will ensure that we will replicate the data from your model to Rockset in real time. You don’t have to lift a finger; you don’t have to write any piece of code. And that data will be one to two seconds behind the source of truth. This will be true for all of your operational databases. This is also true for Kafka and Kinesis and whatever data streams that you’re processing and accumulating. 

“And this is also true if you happen to accumulate data in your lake, whether it is S3, GCS or what-have-you. And so from all these sources, we have built-in connectors at the push of a button. When you bring the data into Rockset, what it automatically does is index the data into fast SQL tables. And we have a full-feature SQL engine on top of it.” 

Rockset was founded and built by the team behind the online data infrastructure that powers Facebook (Meta) Newsfeed and Search. Rockset is inspired by the same indexing systems that power real-time analytics at cloud scale, Venkataramani said. 

As new data is ingested in real-time, Rockset allows continuous transformations and rollups without the need for batch ETL jobs. The data is automatically indexed in a Converged Index, delivering fast search, aggregations and joins on real-time data, with cloud-native speed and scale, Venkataramani said. 

Gartner Research has Rockset competing in the same market as:

  • Redis Enterprise Cloud
  • Cloudera Enterprise Data Hub
  • Databricks Lakehouse Platform
  • Vertica Analytics Platform
  • Neo4j
  • CockroachDB
  • YugabyteDB

Originally appeared on: TheSpuzz