Skip to content

hazelcast.md

Latest commit

 

History

History
58 lines (42 loc) · 1.68 KB

hazelcast.md

File metadata and controls

58 lines (42 loc) · 1.68 KB

HazelCast

https://hazelcast.com/products

Key Points

  • In-Memory Data Grid

  • Apache licensed open source

  • HA, scale out, resilient

  • hash-mod partitioning on the key with replicas (XXX: what about list partitioning, data skew etc?)

  • compact library < 3MB with no deps

  • distributed Java collections (map/set), caching, session state sharing,

  • distributed locks, queues, topics

  • x10 faster than Cassandra

  • Memcache

  • dynamic discovery (multicast) or TCP list, ZooKeeper, Consul, etcd

  • Management Center 2 nodes open source, rest Basic subscription upwards

  • WAN replication in Enterprise only

  • REST + JMX API per node

  • Clustered REST + JMX only in Enterprise

  • Java client - C#/C++ in Enterprise

  • Enterprise 7k, Professional 4k, Basic 2.5k dollars, Open Source

  • Hazelcast founded in 2008, 75 staff inc 30 engineers

  • In Memory Data Storage (TBs, not PBs)

  • In Memory Data Messaging (pub/sub)

  • In Memory Data Computing

  • JCache Provider (JCP) - vendor neutral cache API like JDBC for databases

  • Hazelcast Simulator on Github for provisioning and running stress tests, benchmarking etc

Compute - can run function on every node or select nodes

  • Pub-Sub - has ringbuffer to allocate slow/disconnected consumers to catch up - ringbuffer size is configurable, can overwrite or block

3.6 will have a disk restart option

Code

HazelcastInstance hazelcastInstance = Hazelcast.newHazelcastInstance();

Client only connection not part of cluster memory:

HazelcastInstance hazelcastInstance = HazelcastClient.newHazelcastClient();

Ported from private Knowledge Base page 2014+