Hazelcast aims to democratize data access with low latency with their projects, the In-Memory Data Grid project and the Hazelcast Jet Project. They wanted to address the problem of getting a sub-millisecond response from a query where traditional databases struggled to keep up. The projects are primarily aimed at three verticals: financial services, retail and medical.
In this episode of TFiR Let’s Talk, Swapnil Bhartiya sits down with Manish Devgan, Chief Product Officer at Hazelcast, to introduce us to the company and its offerings. He goes into detail about the need for real-time data and how we have evolved from the batch system to processing as the streaming is coming in. He discusses the announcement of their serverless platform Viridian which has hit beta.
Key highlights of this video interview are:
- Hazelcast has two open source projects, the In-Memory Data Grid project and the Hazelcast Jet project. Although the company started out in the in-memory data grid space with a platform that would sit between the disk-based databases and the application. Devgan goes into detail about the company and how the space has evolved.
- Hazelcast’s projects are primarily aimed at three verticals: financial services, retail, and medical. Some of these industries are highly regulated and customer data is very important. Devgan discusses how they ensure security in their platform, shortening the window of time to market by getting the business to the customer quickly.
- Devgan tells us that as workloads have been moved to the cloud, Hazelcast’s understanding for how to effectively run scale-out data platforms takes the operational complexity away from their customers so they can spend more time doing a business application. He explains about their serverless option which helps developers build their application lowering the skills required to have such a scale-out data platform run behind the scenes.
- Moving from on-prem to cloud presents a number of complexities, which in the industries Hazelcast works with are primarily because customer data needs to be safe and secure. Devgan explains that they focus on the experience developers have when building applications. He goes into detail about their three offerings: open source, their self-hosted enterprise offering, and Viridian.
- Devgan explains the concept of function-as-a-service and the benefits of developers not needing to worry about the infrastructure.
- In the early days of Kubernetes, it was mainly serving stateless workloads but nowadays we are seeing more stateful workloads. Devgan goes on to tell us that one of the reasons they have been so successful, even through COVID, is because they serve most of the mission-critical applications in their customer base.
- Data is not created in batches; it is created in streams, however, developers have been using the batch system since that is what has been available. Devgan explains that now you do not need to store and then process but can process as the streaming is coming in. He feels that streaming is misunderstood by many and how their data platform addresses this.
- Devgan believes that most people do recognize the power of streaming and real-time has become a business necessity. However, in certain cases, the data cannot be stored in a Hadoop cluster and later a Spark job is run. Devgan goes into detail about some of the architectures he has seen trying to deal with the real-time data processing, and why a data platform supporting streaming is needed.
- Devgan takes us through how people can get started with their offerings.
- Hazelcast will continue to invest in serverless to make things easier for their customers. Devgan discusses what the company will be focusing on in the future and shares details of its roadmap.
The summary of the show is written by Emily Nicholls.