Alparslan works for Hazelcast as a Senior Solutions Architect. He is a passionate Java developer, and loves to think in distributed and object-oriented way. Prior to joining Hazelcast, Alparslan worked in several Java projects including a web search engine and enterprise financial anti-fraud solutions. He also contributes to open-source projects like Apache Nutch and currently a PMC member in Apache Gora. Alparslan holds a MS degree in Software Engineering of Distributed Systems from Kungliga tekniska hogskolan in Stockholm, Sweden.
Hazelcast is the one of the most popular distributed in-memory data grids in software world. It is designed to scale up to hundreds and thousands of nodes. Simply add new nodes and they will automatically discover the cluster and will linearly increase both memory and processing capacity. Moreover; the data is restored from the backup and the cluster will continue to operate without downtime on a node failure.
All these cool features are based on the remarkable data partitioning architecture in the backend. The data which is stored in Hazelcast is sharded among cluster members with their backups. In distributed environments, a thousands of different problems can arise like split-brain syndrome. When these type of problems are faced, Hazelcast triggers its re-partitioning mechanism automatically.
In this session, you will go deep into the details of Hazelcast’s data partitioning architecture. Real-life problems about data partitioning and the solutions provided to them in Hazelcast will be lively demonstrated with simple examples.