Product Architecture & Roadmap

This video was captured at the Inaugural Hazelcast User Group London (HUGL) event at the Shoreditch Town Hall.

This event included over 100 Meetup Group members and was kicked off by James Governor, Principal Analyst and founder of Redmonk.

This video covers two main topics, firstly the Hazelcast product architecture, then the product road map. Since version 3.x, Hazelcast has a new architectural model where the underlying infrastructure layers are abstracted from the distributed data structures by an SPI (Service Provider Interface). This enables developers to easily create their own distributed data structures on top of Hazelcast. Hazelcast developers can create a new data structure in a few days, for a new developer it could be a matter of a week. This powerful abstraction layer is made up of multiple clean abstraction layers such as the networking layer, node engine, data partitioning services and cluster management services. In addition the boundaries between Hazelcast and Hazelcast Enterprise are clearly laid out. Also the pluggable serialization forms another powerful extension point that enables multiple languages to talk to the same data set. For now, clients include C++ and .Net/C# (and of course Java) but in the future could include many language bindings.

The Hazelcast Roadmap is also covered, and principal topics include more advancements in the Off-Heap memory implementation in Hazelcast. Also discussed will be the JCache implementation (JSR 107) in an upcoming Hazelcast release. Since Hazelcast hired JSR-107 co-spec lead Greg Luck as its CTO it has redoubled its commitment to the JCache standard.