Hazelcast Gets Edgy: Lowering Latency with IBM Edge Application Manager

Kelly Herrell | May 5, 2020

It takes 300 milliseconds to blink. In that time a car has entered an intersection, a manufacturing robot has missed its mark, a video camera image has lost its usefulness.

The digital era has created the new potential to innovate the world around us. In the process, it has also shrunk the concept of response time. To capture new value, microseconds matter. Latency is the new downtime.

Edge computing is about placing compute power as close as possible to wherever data is being generated and the service is being rendered, to instantly create value from situationally-aware insights. Defeating latency is a requirement at the edge; latency creates missed opportunity, massive cost, and extreme danger. It’s probably not a coincidence that “latency” rhymes with “wait and see.” At the edge, the world waits for nothing.

Drilling operations in the energy industry are a case in point, and the solution delivered by our customer SigmaStream is a perfect example of edge computing. The locations are remote and the rigs are covered in sensors that broadcast specific performance data the rig is experiencing. To process dozens of high-frequency information streams – the moment in which the data is generated – SigmaStream deployed Hazelcast Jet, a real-time in-memory event stream processing engine. This enables them to process the data and feed insights back into the controls immediately, adjusting drill speed and avoiding catastrophic failures – all in real-time. Defeating latency enables more efficient and safer operations whereas the alternative is rig downtime and a negative impact on the business’ bottom line.

Edge computing is enabled in part by the latency-busting advances of 5G. However, that advancement applies only to the latency between an antenna and the cell tower. The remaining challenge is the data processing latency on the edge compute nodes themselves, which can be substantial. Why push infrastructure to the edge only to have the data processing software engine be constantly behind? In addition, in the industrial world, operational infrastructure often operates on secure local networks, demanding secure local edge compute.

That’s why Hazelcast is extremely excited to be a part of the new IBM Edge Ecosystem. From a central point of control, IBM Edge Application Manager automates and enables AI, analytics and IoT workloads to be securely deployed and remotely managed, delivering real-time analysis and insight at scale – up to 10,000 devices simultaneously. This is a breakthrough in activating compute power close to where the data is being generated. The policy-driven deployment and configuration of IBM Edge Application Manager, when combined with Hazelcast’s zero-downtime job upgrades and reliable job management, enables the continuous deployment of real-time IoT applications.

“The convergence of 5G, Edge computing and AI will spark a level of innovation never seen before,” says Evaristus Mainsah, general manager, IBM Cloud Pak Ecosystem, “This, in turn, will enable and fuel a broad ecosystem of partners to co-create for a growing set of Edge opportunities. We are excited about the value that our collaboration with Hazelcast can bring to our joint clients needing in-memory data grid computing for edge systems to run edge analytics.”

IBM Edge Application Manager provides an autonomous way to manage a wide range of distributed compute resources. Coupled with Hazelcast’s in-memory computing platform as the data plane running on the edge compute devices, Hazelcast delivers the most flexible, low-latency edge compute solution in the industry.

Hazelcast is uniquely designed for this digital-age requirement. Whether the data is cached or streaming, the processing response time is immediate. And Hazelcast’s small software footprint allows it to fit in the widest variety of hardware and endpoints. At less than 15Mb, and with the ability to run on Kubernetes, bare metal, or even embedded in Java applications, Hazelcast processes cached or streaming data at light speed within a wide range of edge hardware.

Making use of the flood of data being produced at the edge is a seemingly endless opportunity. The innovation will be found in connected vehicles, smart grids, wireless factories, connected cities, health care, and numerous telco use cases – and that’s just the beginning. The future winners in every industry will be determined by the speed at which they leverage data in ways that immediately produce insights and actions.

When time is money, customers choose Hazelcast. We have proven that in clouds and data centers with the world’s largest organizations. Coupled with the deployment simplicity of IBM Edge Application Manager, it’s now true for edge computing as well.

About the Author

Kelly Herrell

Chief Executive Officer

Kelly has led the growth of four innovative companies from early stage to market-leading entities, covering a broad span of compute and networking. At Hazelcast, Kelly brings his unique experience of driving high-value innovation, including open-source models, into the infrastructure of the world's largest customers. Prior to joining Hazelcast in July 2018, he was SVP & GM of the Software Business Unit at Brocade Communications, the result of Brocade's acquisition of his former company, Vyatta, which pioneered software-defined networking and delivered the most widely used software networking operating system in the world. Kelly has a Bachelor's degree from Washington State University where he is an advisor to the Honors program and an MBA from Cornell University where he is a guest lecturer.