0 Replies Latest reply on Nov 17, 2011 4:54 PM by agnihere

    Achieving locality optimization on huge datasets

    agnihere

      Hi,

       

      we are using Infinispan 5.0 to build an excel-pivot like functionality on extremely large dataset: to the tune of 650 mill. We plan to use filesystem as a backing cachestore (potentially SAN) and plan to keep about 20% of data in-memory across the server cluster. While we will try to achieve as much locality optimization as possible but we might be limited by RAM available to JVM processes. Any suggestions?