[HAMA-794] When aggregator is enabled, resending message in each message receiving round impacts the iteration performance.

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Fix Version/s: 0.7.0
• Component/s: graph
• Assignee: MaoYuan Xian
• Reporter: MaoYuan Xian
• Created: 12/Aug/13 06:32
• Updated: 17/Sep/13 16:17
• Resolved: 17/Sep/13 16:17

Description

When aggregator is enabled, in each message receiving round, all received message have been resent to put them behind the aggregate message. doAggregationUpdates method of GraphJobRunner Class:

    if (aggregationRunner.isEnabled() && iteration > 1) {
      // in case we need to sync, we need to replay the messages that already
      // are added to the queue. This prevents loosing messages when using
      // aggregators.
      if (firstVertexMessage != null) {
        peer.send(peer.getPeerName(), firstVertexMessage);
      }
      GraphJobMessage msg = null;
      while ((msg = peer.getCurrentMessage()) != null) {
        peer.send(peer.getPeerName(), msg);
      }
      // now sync
      peer.sync();
      ...

Should we do some improvement here? Record the original receiveQueue before the sync, and after sync operation, read the aggregator updated value firstly, then restore the original receiveQueue such that avoiding the message resending operations.

Issue Links

• blocks: HAMA-734 Hama Message Manager should be able to delegate the ownership of internal message queue on request for future superstep
• relates to: HAMA-734 Hama Message Manager should be able to delegate the ownership of internal message queue on request for future superstep