JBossCache 1.0 Released

Discussions

News: JBossCache 1.0 Released

  1. JBossCache 1.0 Released (18 messages)

    Distributed transactional cache JBossCache 1.0 has been announced. JBossCache consists of a TreeCacha, a replicated transactional tree-structured cache; and TreeCacheAop, an "object-oriented" cache using AOP to dynamically manage Plain Old Java Objects (POJOs).

    Checkout JBossCache 1.0 and the announcement.

    NOTE: Bela Ban will be talking about JBossCache at TheServerSide Symposium in May.

    Threaded Messages (18)

  2. JBossCache 1.0 Released[ Go to top ]

    Congratulations!

    Sorry for the lazy question, but I figured this would probably be of interest to TSS folks anyway:

    I want to investigate using JBossCache in a heterogenous (J2EE/.NET) environment for certain uses where concurrent writes are necessary, but the throughput is low enough to suffer the overhead of a distributed invocations.

    What are my out-of-the-box options for remote invocations?

    - MBean invocation? (Looks like it from the overview)
    - Web Service Invocation? (Does this require the JBoss .NET integration)
    - Other options?
  3. Access to JBossCache[ Go to top ]

    Congratulations!Sorry for the lazy question, but I figured this would probably be of interest to TSS folks anyway:I want to investigate using JBossCache in a heterogenous (J2EE/.NET) environment for certain uses where concurrent writes are necessary, but the throughput is low enough to suffer the overhead of a distributed invocations.What are my out-of-the-box options for remote invocations?- MBean invocation? (Looks like it from the overview)- Web Service Invocation? (Does this require the JBoss .NET integration)- Other options?
    Hi Corby,

    yes, since the cache is an MBean, you can use any client access mechanism available to invoke MBeans. If you don't use RMI, however, I'm not sure about the notifications emitted by the cache (CacheListener callbacks). RMI definitely works (if your client is an RMI object as well), I'm not sure about .net.
    Bela
  4. JBossCache 1.0 Released[ Go to top ]

    Can anyone get to this page? It's blank for me.
  5. JBossCache 1.0 Released[ Go to top ]

    Can anyone get to this page? It's blank for me.

    Check your BrowserCache 1.0 or GoogleCache 1.0 (http://www.google.com/search?q=cache:5MXuLiw8_9YJ:jboss.org/wiki/Wiki.jsp%3Fpage%3DJBossCache+&hl=fi&ie=UTF-8).
  6. JBossCache 1.0 Released[ Go to top ]

    > Can anyone get to this page? It's blank for me.Check your BrowserCache 1.0 or GoogleCache 1.0 (http://www.google.com/search?q=cache:5MXuLiw8_9YJ:jboss.org/wiki/Wiki.jsp%3Fpage%3DJBossCache+&hl=fi&ie=UTF-8).
    Our WIKI shit the bed...Keep trying, we're working on fixing it.

    Bill
  7. JBossCache 1.0 Released[ Go to top ]

    Gavin,

    If / when you read this, have you done any performance comparisons yet of the different cache implementations under Hibernate? I'd love to see such a comparison for read-mostly (with no or limited transactional needs) to see how transactional caches, like JBoss cache, compare with distributed invalidation caches like SwarmCache or OSCache. In our current usage, distributed invalidation is good enough, but I'm wondering what the overhead is of upgrading to a transactional cache.
  8. Performance of JBossCache[ Go to top ]

    Gavin, If / when you read this, have you done any performance comparisons yet of the different cache implementations under Hibernate? I'd love to see such a comparison for read-mostly (with no or limited transactional needs) to see how transactional caches, like JBoss cache, compare with distributed invalidation caches like SwarmCache or OSCache. In our current usage, distributed invalidation is good enough, but I'm wondering what the overhead is of upgrading to a transactional cache.
    Hi Jason,

    note that JBossCache is not per se transactional: it can also be used as a simple efficient low-overhead asynchronous cache, ie. replication goes on in the backgroups, for example every 500 ms (when items are available).

    JBossCache can be configured to be transactional, but that's just one way of running it.

    Bela
  9. Performance of JBossCache[ Go to top ]

    Hi Jason,note that JBossCache is not per se transactional: it can also be used as a simple efficient low-overhead asynchronous cache, ie. replication goes on in the backgroups, for example every 500 ms (when items are available).JBossCache can be configured to be transactional, but that's just one way of running it.Bela
    Can it also be used where it just invalidates the other caches in the cluster on a put, like SwarmCache or OSCache do, with no replication?
  10. Performance of JBossCache[ Go to top ]

    Can it also be used where it just invalidates the other caches in the cluster on a put, like SwarmCache or OSCache do, with no replication?Yes and No.

    Yes if you keep your data in the cache, and remove() will 'invalidate' across the cluster.

    No if you just want remote cluster RPC. I do provide a callRemoteMethods() in TreeCache, but that's just a supporting method. If you intend to use it in that way, you may be better off accessing the JGroups chanel underneath directly.

    Bela
  11. It would be nice if some people could give some examples for which a cache like this can be used. In what parts of a J2EE application is this useful? Could it be a replacement of a JMS queue that is used to pass messages around?

     Stefan
  12. I'm not sure what the best way to use it is. But we're thinking of trying to put our entire LDAP tree into memory, and JBossCache sounds like it'd be a prime way to go.

    We use it for user auth and roles, as well as keeping track of customer/site info. Currently we're not caching anything, so short-term I'm putting it in the HTTP Session, medium term I may put it in a stateful session bean (each user's info is fairly unique), and long-term, maybe JBCache.

    Steve
  13. So what do you use JBossCache for?[ Go to top ]

    I'm not sure what the best way to use it is. But we're thinking of trying to put our entire LDAP tree into memory, and JBossCache sounds like it'd be a prime way to go.We use it for user auth and roles, as well as keeping track of customer/site info. Currently we're not caching anything, so short-term I'm putting it in the HTTP Session, medium term I may put it in a stateful session bean (each user's info is fairly unique), and long-term, maybe JBCache.Steve
    LDAP, JNDI are good examples, partly because JBossCache *is* essentially LDAP. I mean, the structure of LDAP/JNDI and JBossCache are almost the same.
    Bela
  14. So what do you use JBossCache for?[ Go to top ]

    It would be nice if some people could give some examples for which a cache like this can be used. In what parts of a J2EE application is this useful? Could it be a replacement of a JMS queue that is used to pass messages around? Stefan
    Hi,

    We have envisioned that the HttpSession replication can be an interesting application for JBossCache. Especially, if the session data is a huge hashmap, then TreeCacheAop can have a performance advantage since it only replicates the modified keys.

    Thanks,

    -Ben Wang
  15. So what do you use JBossCache for?[ Go to top ]

    It would be nice if some people could give some examples for which a cache like this can be used. In what parts of a J2EE application is this useful? Could it be a replacement of a JMS queue that is used to pass messages around? Stefan
    I'm working on HTTP session replication for Tomcat5 (integrated with JBoss). All sessions/attrs are stored in the JBossCache MBean.

    Another aspect we're looking at is the replacement of HA-JNDI.

    Bela
  16. "Distributed transactional cache JBossCache 1.0 has been announced. JBossCache consists of a TreeCacha, a replicated transactional tree-structured cache; and TreeCacheAop, an "object-oriented" cache using AOP to dynamically manage Plain Old Java Objects (POJOs)."

    That it means that there is a 'Replicated Transactional Tree' component available in the latest JGroups release?

    Is the implementation of the aspects done by the JBoss AOP framework.?

    So, if I´m right, if there is such component (Replicated transactional tree), we should code our own POJOs manager, may be using AspecJ. This manager would basically perform a caching algorithm that updates the cached data? (server push, client pull, hybrid push&pull). If we do not need to use an application server.
    Alfonso
  17. JBossCache 1.0 Released & JGroups[ Go to top ]

    Is the implementation of the aspects done by the JBoss AOP framework.?
    Yes, it's based on the latest JBossAop. Bela can speak more but this release has nothing to do with JGroups so far.

    -Ben
  18. JBossCache 1.0 Released & JGroups[ Go to top ]

    Is the implementation of the aspects done by the JBoss AOP framework.? So, if I´m right, if there is such component (Replicated transactional tree), we should code our own POJOs manager, may be using AspecJ. This manager would basically perform a caching algorithm that updates the cached data? (server push, client pull, hybrid push&pull). If we do not need to use an application server.Alfonso
    JBoss AOP works outside of the application server as well as JBossCache.

    Bill
  19. JCache compatibility?[ Go to top ]

    I am investigating different cache implementations to use in my apps and interoperability and standardization are of course to be considered. What is the actual status of JSR-107(JCache) and is JBossCache related with it in any way?

    Thank you,
    Calin Lupa