News

Hazelcast Enhances Its Jet Real-Time Event Streaming Engine for the Edge

• By John K. Waters
• 06/01/2020

In-memory computing platform provider Hazelcast is upping its edge-computing game again in the latest release of its real-time event streaming engine, Hazelcast Jet. In addition to some platform performance improvements, the company has added two new operators to simplify deployments within containers.

Jet 4.1, announced on May 13, introduces the Hazelcast Jet Enterprise Operator for OpenShift, which has been certified by Red Hat and is available now from the Red Hat Marketplace, and the Hazelcast Jet Kubernetes Operator, which is available for the open source and enterprise versions of the streaming engine.

Computing at the edge is fraught with potential disaster, because of the time it can take to process the data streaming from sensors and other edge devices. Microseconds matter, explained Kelly Herrell, CEO of Hazelcast.

"We view the blink of an eye as a minuscule amount of time," he said in a statement, "but in reality, it takes 300,000 microseconds. "By placing processing power at the edge and turbocharging it with Hazelcast's ultra-low latency stream processing capabilities, new applications will be enabled that will innovate our business capabilities and improve our daily lives."

These updates simplify integrations with Red Hat OpenShift and Kubernetes, and bring customers closer to achieving "near real-time insights in edge computing environments," the company stated.

Hazelcast announced new support in Jet for artificial intelligence (AI) and machine learning (ML) deployments of mission-critical applications in the March release of Jet 4.0. In the previous release, Hazelcast gave Jet the ability to call out to Python routines from within a Jet Pipeline. Now it’s possible to invoke external gRPC (Remote Procedure Call) enabled services during the processing of a Jet Pipeline. Introduced by Google and donated to the Cloud Native Computing Foundation (CNCF), gRPC is a standards-based framework for exposing and calling services. It is now possible to have Jet Pipelines interact with services written in languages other than Java via this mechanism.

"Successful edge solutions cannot rely on distant datacenters and cloud services to shoulder the burden of high-volume data processing," said David Brimley, chief product officer at Hazelcast, in a statement. "Increasingly, edge deployments have to take on this role. Not only does this approach reduce latency, it makes budgetary sense when network costs become prohibitive, especially when sending data to the cloud. Hazelcast Jet has the power to solve this challenge."

Hazelcast also announced that it has joined the IBM Edge Ecosystem as a featured partner for the IBM Edge Application Manager. IBM unveiled new edge-computing services and solutions in early May designed to help enterprises make the most of their transitions to 5G. The Edge Application Manager is an autonomous management solution that makes it possible for AI, analytics, and IoT enterprise workloads to be deployed and remotely managed. The solution enables the management of up to 10,000 edge nodes simultaneously by a single administrator.