Riding the Streaming Analytics Wave

Two years ago, when Forrester last published its Wave for Streaming Analytics, our streaming capabilities were still in the R&D phase. We were working on a technological challenge based on the belief there was a new generation of stream processing that was possible – one steeped in our ethos of defeating time.

For the world of streaming and events, that meant an architecture that leapfrogged the old batch-oriented methodology upon which Spark and others are based. In the true spirit of Hazelcast, we were pushing technology limits into continuous real-time stream processing while also delivering the lowest latency.

Fast forward to today when Forrester’s latest Wave for Streaming Analytics, The Forrester Wave™: Streaming Analytics, Q2 2021, places Hazelcast in a Strong Performer position. We believe it’s highly unusual for a vendor to be ranked in such a position the first time it’s included in analyst research of this nature, but our hard paddling helped us catch this Wave.

Along with the recognition by Forrester, the acknowledgment of our innovation has been growing rapidly:

• Hazelcast’s stream processing capability is in production in mission-critical environments among the world’s largest customers, including leading banks around the world.
• We published a performance benchmark that demonstrated Hazelcast processing one BILLION events per second at only 20 milliseconds of latency.
• The world of academia is quickly acknowledging Hazelcast’s breakthroughs.
  • An academic paper on Hazelcast’s system architecture for stream processing has been accepted to the Very Large Database Conference in August 2021.
  • Another academic paper from Potsdam University was recently published, demonstrating Spark’s inability to complete a continuous processing benchmark, and further showing Hazelcast’s latency advantage of nearly 5,000x over Flink, the Alibaba project.

As with many meaningful innovations, however, it begs the question: “What is it good for?” How can customers take advantage of it to create business value? The range of uses is extensive:

• Power AIOps through instantaneous insights into automated events
• Real-time monitoring of industrial inputs for instantaneous anomaly detection
• Continuous processing for real-time telemetry
• Turn slow batch jobs into continuous processing for immediate, actionable insights

Hazelcast now simplifies the creation of real-time intelligent applications to capture the continuous flow of value that you are missing today. We have only scratched the surface of the business potential.