Let It Crash
letitcrash.com · Company blog · on HN since 2012
The Akka team's blog on the Akka actor toolkit for the JVM, covering releases, performance benchmarks, and the actor model.
What HN saysHN commenters cite it for Akka performance data, such as a post showing actors scaling better than a traditional thread pool, and for release and course announcements.
20HN people shared it
8linked it in comments
13stories submitted
158story points
8 years agolast shared
2012people per year2026
Most-shared posts
- 150 million messages per second - on a single machine5 people · 4 points · first shared 2012 · 1 comments on HN
- 2Scalability of Fork Join Pool2 people · 38 points · first shared 2012 · 1 comments on HN
- 3Akka 2.4.0 Released2 people · 6 points · first shared 2015 · discussion on HN
- 4Carl Hewitt Explains the Essence of the Actor Model of Computation2 people · 5 points · first shared 2012 · discussion on HN
- 5Akka 2.2.0 released – distributed, concurrent applications on the JVM1 people · 91 points · first shared 2013 · 13 comments on HN
- 6Distributed (in-memory) graph processing with Akka1 people · 4 points · first shared 2012 · discussion on HN
- 7Where Akka came from1 people · 3 points · first shared 2013 · discussion on HN
- 8Akka: an event-driven middleware framework for distributed systems1 people · 3 points · first shared 2013 · discussion on HN
- 9When to use Typed Actors1 people · 2 points · first shared 2012 · discussion on HN
- 10Akka 2.4 M1 is released1 people · 1 points · first shared 2015 · discussion on HN
- 11Akka 2.2.0 successfully used with a cluster of 200 Nodes1 people · 1 points · first shared 2013 · discussion on HN
- 122 1 spotlight akka actors enter scala distribution1 people · first shared 2013
- 13New coursera course on principles of reactive1 people · first shared 2013
- 14Homepage1 people · first shared 2014
What HN people wrote when they linked it
jaimefjorge · 2013-10-16
Well written, good description and nice demo.
Would love to see more on how this is different to (or better than) Akka.
The programming model is actually close to Akka (with actor systems, supervision, receive method, message passing, etc).
The article states that Akka has no true lightweight threads. The guys behind Akka have put it running with 50M messages/second[1] and perfomance vs erlang seems to be good as well [2][3].
Perhaps a benchmark would be great.
Thanks for sharing.
[1] http://letitcrash.com/post/20397701710/50-million-messages-per-second-…
[2] http://uberblo.gs/2011/12/scala-akka-and-erlang-actor-benchmarks
[3] http://musings-of-an-erlang-priest.blogspot.pt/2012/07/i-only-trust-be… (discussing millions of messages is a good signal IMHO).
soc88 · 2012-04-03
- Yes, HotSpot is “the real” thing and includes the best Oracle has. There is a JVM implementation from Azul too, which people consider to be excellent.
- The case between Oracle and Google is not related to the code released by Sun/Oracle.
- You can write heavily optimized code in Java that is as fast or faster than C, without using escape hatches.
- There is no overhead for things you don't use (synchronization).
- The JVM has probably the fastest general-purpose implementation for concurrency primitives for high-level languages. See for example: http://letitcrash.com/post/20397701710/50-million-messages-per-second-…
- The lack of tail calls is very annoying. The rest just works.
-
zackangelo · 2015-04-03
Indeed, the Akka team published an blog post [1] a while back that demonstrates superior performance versus a traditional thread pool. The actor model seems to be a very favorable use case for F/J though.
[1] http://letitcrash.com/post/17607272336/scalability-of-fork-join-pool
based2 · 2014-02-20
tomku · 2013-08-27
One of the blog posts[1] I saw announcing it had a few more details than the course page, including a mention of RxJava.
[1]: http://letitcrash.com/post/59504788607/new-coursera-course-on-principl…
ken_railey · 2013-03-13
>One could adopt some kind of event-based system in scala or clojure, but Erlang and Go are alike in being the lone runtimes where running millions and millions of very small messaging processes is AOK no problem for the runtime
Is that really the case? Scala's Akka actor library seems like it would qualify just fine: http://letitcrash.com/post/20397701710/50-million-messages-per-second-…
fedesilva · 2013-01-04
I think that you will have to keep that sbt line there because akka-actor comes with the distribution but is a separate jar. So you will have to pull that dependency yourself in sbt.
"For Akka users this change does not have visible effects (you will continue to depend on the same artifacts)"
http://letitcrash.com/post/37716546018/2-1-spotlight-akka-actors-enter…
erichocean · 2012-09-03
Even more along these lines, Scala (another JVM-based language) has demonstrated 50 million message sends/per second with their Akka library.
http://letitcrash.com/post/20397701710/50-million-messages-per-second-…
I too like Node for small little tasks, tiny dev web servers, etc.
But the JVM is a far more mature platform for writing scalable, multi-core application servers, especially with modern languages like Scala and Clojure.