Geek Answers Handbook

Ehcache Performance on a Large Cluster

I would like to use the replicated Ehcache cache, first as a backend for Hibernate's second level cache, second as a cache for any data.

I know how a distributed cache works, such as memcached, and I know that it can scale for large clusters, but I cannot find how Ehcache replication behaves on large clusters.

  • Does anyone have a pointer to some information or some kind of benchmark?

I have found that many replication strategies can be used, such as RMI, JGroups, JMS or Terracotta, and RMI and Terracotta seem to be the most popular.

  • How do they compare with large clusters?

Will replication kill my actions when I add many nodes (for example, a few dozen)?

+5
source share
3 answers

A fully replicated cache will only work if your application is read. The replicated cache cannot scale; transferring updates to other nodes will result in loss of your productivity. You need a partitioned cache with backup replicas. Segmented caches will scale linearly even for intensive recording applications.

Try Hazelcast ! it is an open source transactional solution (Apache), peer-to-peer caching for Java. It comes with a second level cache plugin for sleep mode.

? . Hazelcast 100 node - .

+4

" ", node ( ), . .

, ehcache . JBossCache , Hibernate , ehcache.

+3

ehcache?

:

+1

More articles:


All Articles