[rabbitmq-discuss] ERLANG design flaws cause performance problems?

Michael Klishin michael.s.klishin at gmail.com
Tue Jun 4 21:40:57 BST 2013


2013/6/5 Guido Stepken <gstepken at googlemail.com>

>
> http://www.unlimitednovelty.com/2011/07/trouble-with-erlang-or-erlang-is-ghetto.html
>
> Are these resolved in the mean time?
>

A lot of that blog post is venting and is not backed by facts.

"Erlang sucks at managing memory" is
a meaningless statement unless it is backed by data and specifics about the
workload.

Oracle JVM with large heaps (> 8 GB in my experience) suffers from pauses
in seconds on old generation
collections. Can we say that "JVM sucks at managing memory"? I don't know.
For some workloads
this problem makes latency very unpredictable. For some apps it's not an
issue.

Erlang proceses do not share memory so its GC has significantly less work
to do and this
particular problem is not applicable to Erlang.

JIT is coming to the Erlang VM in one of the future releases [1]. Only time
will tell what kind of performance
gain for real world projects it may bring.

Be very suspicious of all "X is a ghetto" blog posts. They typically
provide next to no specific data
and tend to stir up the drama to make the author famous in the circles of
rockstar ninja developers.

1.
http://www.confreaks.com/videos/2039-erlangfactorysf2012-just-in-time-compiler-for-the-erlang-vm
-- 
MK

http://github.com/michaelklishin
http://twitter.com/michaelklishin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/attachments/20130605/a3d45db4/attachment.htm>


More information about the rabbitmq-discuss mailing list