5 Replies Latest reply: Mar 8, 2013 3:55 AM by 984647 RSS

    Is Their Any Support For Data Aware Routing Through The Coherence DataGrid?

    984647
      Hi all,
      Following paragraph describes the feature i m looking for(*Data Aware Nodes*)......
      The Function Service can transparently route behavior to one or more nodes in the cluster. In the case of Data Aware Functions, the behavior can be directly routed to the node(s) that hosts the data needed to execute the function in parallel, stream and aggregate the results substantially reducing the time taken to execute complex data intensive tasks.

      So please tell me is their any support for above said question. If this feature is their in coherence please help me with it's configuration and implementation as well.

      Thanx
        • 1. Re: Is Their Any Support For Data Aware Routing Through The Coherence DataGrid?
          Jonathan.Knight
          Hi,

          Yes you can route behavior to nodes that own a specific piece of data. You might need to expand a bit more on exactly what you atre thinking about doing.

          JK
          • 2. Re: Is Their Any Support For Data Aware Routing Through The Coherence DataGrid?
            984647
            Hi JK thank you for your reply
            I want to achieve Data Aware Routing in Parallel Execution (Parallel Function Execution).
            Consider a case where i have three nodes (say Node A,B C) in coherence cluster with partitioned scheme. Now suppose we have Employee related objects stored in cache of Node A and Node C only, but not on Node B. Then i want to perform some operation on Employee objects only in parallel fashion.
            In coherence i ll achieve this using map.invokeAll(collectionKeys, agent);_ from client program.
            Now my question to you is, as my operations are based on Employee objects only which are stored on Node A & c. Then When i request parallel Execution from client program which will connect to Node A from our cluster, will it go to Node B where their is no Employee objects ??????

            Edited by: 981644 on Mar 6, 2013 7:52 PM
            • 3. Re: Is Their Any Support For Data Aware Routing Through The Coherence DataGrid?
              984647
              Hi JK thank you for your reply
              I want to achieve Data Aware Routing in Parallel Execution (Parallel Function Execution).
              Consider a case where i have three nodes (say Node A,B C) in coherence cluster with partitioned scheme. Now suppose we have Employee related objects stored in cache of Node A and Node C only, but not on Node B. Then i want to perform some operation on Employee objects only in parallel fashion.
              In coherence i ll achieve this using map.invokeAll(collectionKeys, agent);_ from client program.
              Now my question to you is, as my operations are based on Employee objects only which are stored on Node A & c. Then When i request parallel Execution from client program which will connect to Node A from our cluster, will it go to Node B where their is no Employee objects ??????
              • 4. Re: Is Their Any Support For Data Aware Routing Through The Coherence DataGrid?
                Jonathan.Knight
                Hi,

                If you use map.invokeAll(collectionKeys, agent) then the "agent" will only go to the nodes that own the keys in collectionKeys and will only execute on those nodes. The agent will execute in parallel across the nodes - i.e. it will execute on all the nodes at the same time. It will not be in parallel on an individual node though, so if say collectionKeys contained 10 keys and Node A contained five of those and Node B contained five, then Node A would process its five keys one after the other and at the same time Node B would process its five keys one after the other.

                JK
                • 5. Re: Is Their Any Support For Data Aware Routing Through The Coherence DataGrid?
                  984647
                  Hi JK,
                  Thanx for your reply again.
                  Information provided by you is very much helpful for us.