The duality characterizing Aerospike SQL Powered by Starburst, the most recent release from real-time data platform vendor Aerospike, is significant. The solution, which symbolizes the partnership between Aerospike and analytics engine specialist Starburst, enables organizations to run Starburst SQL queries on Aerospike’s NoSQL database.
Those queries run on what Aerospike Chief Product Officer Lenley Hensarling called “Presto Trino”. In addition to granting SQL access to Aerospike’s NoSQL data, the offering effectively brings Aerospike into the modern, distributed data architecture Starburst supports.
The solution’s benefits extend to end users and the respective vendors via “a win-win setup where sometimes we sell this and that leads to people saying ‘I can use this for other data sources,’ which generates business for [Starburst],” Hensarling said. “When they go in, sometimes people say, ‘would this work against Aerospike because I’ve been thinking of using it but my compliance people can’t get access to the data.’”
Aerospike SQL Powered By Starburst democratizes data access so anyone — from compliance teams to analysts — who’s SQL-savvy can use Aerospike’s real-time NoSQL data. The product ushers in an entirely new user base to Aerospike’s low-latency database, substantially broadening its value throughout the enterprise.
The Power of SQL Access
Aerospike’s real-time NoSQL database was primarily limited to developers or data scientists requiring low-latency applications for financial trading, brokerage firms, adtech, or edge deployments. The new release (which involves Aerospike’s connector to Trino, a SQL-based federated query engine, via a license limited to Aerospike data) expands the reach of that data throughout the enterprise. The first advantage is that any business role unfamiliar with programming for Aerospike can now access this data with “their favorite tools like Tableau or DBeaver,” Hensarling noted.
Thus, this previously difficult-to-access data now becomes useful for business reporting, analytics, auditing, or any other purpose organizations might have for it. Secondly, this broader user base can utilize SQL to access NoSQL data like key-value pairs, graph models, document structures (JSON), and more. “When you use Presto Trino with this, we introspect the structure of the data in our database and put that into what looks like a SQL schema,” Hensarling commented. “That’s part of the magic of Starburst.” Subsequently, SQL becomes the gateway for semistructured and unstructured data, not just structured data.
Distributed Data Architectures
Aerospike assists in the above process with support for secondary indexing in its Aerospike Database 6 which, in conjunction with pushdowns for Trino, increases its performance. “Where the indexes are available, Starburst makes use of them, so we’re more responsive to these SQL-like queries,” Hensarling said. Moreover, Aerospike SQL Powered By Starburst transfers the real-time applicability of Aerospike into the distributed data architectures Starburst champions. The latter’s data virtualization, query federation, and data materialization capabilities underpin some of the fundamental developments in data fabric and data mesh deployments.
Users can then be federate Aerospike at will — with SQL — alongside “other data sources like Oracle, PostgreSQL, MySQL, and Cassandra, while allowing us to be part of the modern data mesh,” Hensarling explained. Federating Aerospike’s real-time data with these additional sources enhances the value of all enterprise data so users can query across it. “What’s happening right now in the world of large enterprises is they have local ownership, but they want global visibility. Maybe for a given application they have data. But Starburst allows you to make that available to other people. Even if you have multiple applications on Aerospike, each with its own cluster, Starburst allows you to see that data together.”
Although SQL-based access to NoSQL data sources, formats, and structures isn’t new, it’s something that has enduring value to the enterprise. By enabling a much more diverse set of users to access, in this instance, the real-time data populating Aerospike’s data as though it were structured, this capability will likely add appeal to the Aerospike platform.
Moreover, even the biggest skeptic about the long-term utility of data fabrics and data meshes must acknowledge the merit of providing uniform access across data sources. The fact that the data landscape continues to become more decentralized simply reinforces this need. “That’s a big step forward in the value of the data that’s accreting in Aerospike,” Hensarling concluded. “It allows us to have SQL access to our data. But it also brings us into that ecosystem, if you will, of Starburst, so we can be federated.”
Aerospike is a sponsor of The New Stack.