3 Replies Latest reply: Jun 14, 2013 9:39 AM by user639604 RSS

    Cache Distribution failure

      Hi all,

      What is the optimization features available in the Distribution Cache Mechanism.

      Currently Im running two clusters in 2 nodes and observed the performance is degrading when the clusters (from different nodes) are added.

      Can any one help me on this?
        • 1. Re: Cache Distribution failure

          Can you elaborate more on the issue? If you are adding more clusters to the same node, it is obvious that you are trying to squeeze more juice from the same hardware so it is expected.

          • 2. Re: Cache Distribution failure

            Hi NJ,


            Im having different clusters on different nodes (different hardwares).


            Cluster 1 is on node

            Cluster 2 is on node

            Cluster 3 is on node

            Now performance has got degraded the moment 2 more clusters got added


            I'm not able to get conclude on this why its happening.


            Please let me know if any other information is required.



            ~Ravi Shanker

            • 3. Re: Cache Distribution failure

              Was the performance degrade very significant?


              If you are not using extend client (running the test code within the cluster node), the performance will degraded if you are using distributed cache.


              Looks like you only run one node on each server.


              When running on only one node, all cache requests are running on same process locally.   The response will be extremely fast.

              When running on two nodes, 50% of the cache requests will need to go to a process running on remote servers.   Those request will have network penalty.

              When running on three nodes, 66% of the cache requests will need to go to a process running on remote servers.


              etc etc.