such a feature is not included in Infinispan, but should be very easy to integrate yourself; you would only need to implement the CacheLoader interface and plug it into your Infinispan configuration file.
The only limitation is that Infinispan understand key/value access to data only, so it won't be able to speedup any query, as long as you don't store full query results in it.. up to you.
Thanks for your quick response. I saw a presentation by Manik positioning infinispan as a full data service. We are looking out for a data grid based solution - I would assume for database backend, we will need to implement read/write back to database backend? We are looking for a solution which is less impact on our current data access any pointer appreciated
1 of 1 people found this helpful
for raw access to any JDBC based database, there already are several CacheLoaders implementations, documented at the link I posted above. What I meant is that if you need something more specialized, implementing that interface is trivial, but maybe not needed as any of the JDBC cacheloaders should work.
A better answer depends on what your current data access is; if you're using JPA or Hibernate, Infinispan provides a 2nd level cache integration for Hibernate which is completely transparent to code using it. A more complex integration is coming as well as http://www.hibernate.org/subprojects/ogm.html, but will serve you only in case you're using JPA/Hibernate.
Also Teiid can be pretty fast, isn't it more likely that you want to have Teiid -> Infinispan -> database (the other way around) ? Just wondering, as since Teiid has an engine to decide how to best perform each specific query, it might help you out with a better integration.
Thanks for the input. Yup having teiid to query all the data on infinispan would work as infinispan provides the scalability while teiid provide the jdbc API for app access so there is less impact for code change. I assume I will need to implement a Infinispan connector for teiid.