News

Hazelcast Adds SQL Support in Latest Upgrade

• By John K. Waters
• September 17, 2020

In-memory computing platform maker has Hazelcast is adding new features and enhancements to its namesake Java-based in-memory data grid. The Hazelcast IMDG 4.1 update will include preview support for managing distributed data using SQL, out-of-the-box support for Kerberos, additional tuning options for Intel Optane DC Persistent Memory Modules, and quicker cluster rebalancing.

San Mateo, Calif.-based Hazelcast has deep Java roots (its in-memory data grid is written in Java), but the company has spread its support over the years, thanks in part to the efforts of its open source community, to include several clients and programming languages, including .NET, Node.js, Python, C++ and Go.

To complement its existing APIs for languages, the latest version of the in-memory data grid will come with new SQL support. Adding this capability to Hazelcast IMDG enables digital integration hubs to retrieve data using a common well-known API, the company says. Usage of indexes to speed up the queries, along with the ability to filter results on attributes extends Hazelcast IMDG beyond the query-by-key capability of basic key-value stores. "The familiarity equates to simplified development and reduced implementation costs as well," the company said in a statement.

"Given its widespread familiarity among developers, the addition of SQL support significantly reduces the learning curve and opens up new opportunities for enterprises to tap into the benefits of an in-memory computing platform," said David Brimley, chief product officer at Hazelcast, in a statement. "With SQL, it's even more straightforward to let Hazelcast IMDG do the heavy lifting of retrieving data from multiple nodes, in parallel and ensuring correctness without introducing extra development costs."

In this release, the SQL support is intended for "select" queries on Hazelcast IMDG maps already populated with data, the company says. Additional capabilities, such as joins, aggregations and sorting, will be added in future releases.

The company argues that "the combination of business leaders requiring tailored views of data and the proliferation of data sources is stressing legacy architectures and infrastructure," the emerging solution for which "a new architecture, digital integration hubs, that is a data architecture that provides a single access point and standardized API that can be called upon by multiple applications." These hubs are often deployed to reduce workloads on backend systems, the company says, to accelerate access to data hosted in backend databases and mainframes, and to provide a common API to a variety of data sources  to integrate new technologies into legacy architectures.

Hazelcast empowers proven, in-memory digital integration hubs by providing object storage in RAM, write-through caching, distributed processing and predefined connectors to many popular data sources.

And in this release, the company introduces parallel partition migrations, which Hazelcast says accelerate cluster rebalancing and reduces the time spent in a suboptimal state "by an order of magnitude." For example: "Should a network disconnect one node in a 10-node cluster responsible for storing four terabytes of data, Hazelcast IMDG will be able to complete partition rebalancing in approximately 2 minutes when it would have previously required at least 33 minutes."

The upgrade also introduces out-of-the-box support for Kerberos. By integrating support for the network protocol into Hazelcast IMDG, the engineering time and risks associated with manual integration are greatly reduced, the company says. The Kerberos protocol complements the enterprise-grade security suite of Hazelcast IMDG Enterprise.

The Hazelcast IMDG is open-source and distributed under an Apache 2 license that allows developers to include the grid in their applications.

Hazelcast partnered last year with Intel to optimize for Intel Optane DC Persistent Memory. Intel Optane is workload-optimized technology that comes with Intel Xeon Scalable processors. The combination of Hazelcast in-memory solutions with Intel scalable processors with memory persistence amps up speed, reliability, and scalability, the companies said at the time, because organizations can run more workloads in-memory.

Hazelcast IMDG 4.1 beta is available today and general availability (GA) is expected later this year.