• Icon: Task Task
    • Resolution: Unresolved
    • Icon: Major Major
    • 5.5
    • None
    • None
    • False
    • False

      • Backport to 4.x (BatchMessage a subclass of Message?)
      • Measure performance
      • Refactoring, e.g. do we need a separate timer, or can we send when (a) the batch is full or (b) no more messages are received?
      • Documentation
      • Test with Infinispan
      • Only multicast messages have been tested (MPerf); test with UPerf
      • Make default in sample protocols

            [JGRP-2591] BATCH: productize

            Bela Ban added a comment -

            Merge BATCH and BATCH2, if possible

            Bela Ban added a comment - Merge BATCH and BATCH2, if possible

            Bela Ban added a comment - - edited

            See if we can turn BATCH into a latency-friendly protocol. Currently, the timer going off at given intervals has an adverse impact on latency. Perhaps see if a solution similar to what's done in the bundler (send as soon as a batch is full, or no more message is sent) is applicable. It would be ideal if we had a single version of BATCH that could be used in both latency sensitive and throughput-friendly scenarios.

            Bela Ban added a comment - - edited See if we can turn BATCH into a latency-friendly protocol . Currently, the timer going off at given intervals has an adverse impact on latency. Perhaps see if a solution similar to what's done in the bundler (send as soon as a batch is full, or no more message is sent) is applicable. It would be ideal if we had a single version of BATCH that could be used in both latency sensitive and throughput-friendly scenarios.

              rhn-engineering-bban Bela Ban
              rhn-engineering-bban Bela Ban
              Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

                Created:
                Updated: