Github user HeartSaVioR commented on the issue:
https://github.com/apache/storm/pull/2270
@Ethanlm
The performance looks great.
LocalOrShuffle would not work in that topology placement (it is only
effective when "a" worker has connected tasks), but we could assume it should
be fastest if such condition is met, so we can ignore that.
Great that it didn't show any outstanding performance hit even with no
network latency.
I'm still aware of concurrent hashmap lookup, but there're other
bottlenecks as well (acker, etc.) and more performance tuning might not be
accurate unless we resolve other bottleneck issue. So I'm OK to keep it as it
is for now.
+1
@roshannaik @govind-menon
Please let me know your thought about the performance test result or the
patch itself. I'll wait for several days (maybe next Monday in PST) and merge
if there's no outstanding comment.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
