Cassandra and Terracotta are "distributed" and can both store data, but I think the similarities end there. My understanding of Terracotta from recently evaluating their improvements to quartz is that their design goals are more focused on distributed and fault tolerant JVM resource sharing with persistence being more or less based on Ehcache writing to disk (please correct me if I missed something here). This approach makes a lot sense when considering storage of "job status" in a cluster, by I want my persistence of domain data to be in a different layer than my applications.
Cassandra is more appropriate for a fault tolerant, distributed key value store because it is a decoupled service that, via Thrift, can be used with different technologies easily so is a better fit for persistence in my architecture than Terracotta would be. -Nate On Fri, Feb 26, 2010 at 9:34 AM, Erich Nachbar <er...@nachbar.biz> wrote: >> Can we compare Cassandra to Terracotta ? > > I used Terracotta on a previous project and you can think of it as a > persisted/fault tolerant cache. > The free version doesn't come with sharding, so your data has to fit on one > machine. > Other people I know have used it as a persistence layer (i.e. replacing > something like an RDBMS/Cassandra/HBase/CoucheDB/etc.) and it failed > miserably. > > >