camel-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Cristiano Costantini <cristiano.costant...@gmail.com>
Subject Re: Kryo Data format
Date Wed, 05 Jun 2013 08:10:35 GMT
I lack of time but I would love to contribute;
I will check the link and hope to find an easy way to help ;-)
For sure, it is a good starting point that there is interest from the
community.

But first I will try to use it with camel as a
http://camel.apache.org/custom-dataformat.html

Regards,
Cristiano




2013/6/5 Claus Ibsen <claus.ibsen@gmail.com>

> Hi
>
> This seems cool. Sure like faster serialization libraries than what we
> have in the JDK.
> And its pure Java based, where as eg google protobuf is a bit more
> complicated to use.
>
> Keep on hacking on it. And as you may know we love contributions
> http://camel.apache.org/contributing.html
>
>
>
> On Tue, Jun 4, 2013 at 10:45 AM, Cristiano Costantini
> <cristiano.costantini@gmail.com> wrote:
> > Hello all,
> >
> > I was playing with the Kryo (https://code.google.com/p/kryo/)
> serialization
> > library, comparing it to Java default serialization and
> > Marshalling/Unmarshalling with JaxB.
> >
> > Taking aside Java Serialization, which would require me to implement
> > Serializable and I can't,
> > I like the fact that Kryo:
> > - has serialized my JaxB Annotated classes 5 time faster than JaxB
> > Marshalling
> > - it has proved to be compatible using only the default
> "FieldSerializer",
> > which don't require to implement Serializable on the classes.
> >
> >
> > I wonder if anyone has experience with Camel and Kryo and if anyone has
> > ever thought of a camel-kryo dataformat.
> > I now use JaxB to put classes in ActiveMQ, I'm worried with
> performances, I
> > think such data format would be able to provide a faster alternative non
> > intrusive alternative.
> >
> >
> >
> > Here the result of my comparison of Kryo vs JaxB (only
> > marshalling/serialization, dataset of 1000 random generated classes which
> > in XML are each one 480 nodes in average, with about 1,2 attributes per
> > node).
> >
> > Testing JaxB Marshalling
> > Total Time: 1.186 seconds
> > Total Size: 37689851 (string length)
> >
> > Testing Serialization with Kryo
> > Total Time: 0.216 seconds
> > Total Size: 3674796 (bytes)
> >
> > Checking Serialization with Kryo
> > Checked 1000 classes, 1000 instances, 1000 marshalled strings
> >
> > (in this last test I've verified the classes serialized/deserialized with
> > Kryo comparing them as marshalled XML string with .equals() )
>
>
>
> --
> Claus Ibsen
> -----------------
> www.camelone.org: The open source integration conference.
>
> Red Hat, Inc.
> FuseSource is now part of Red Hat
> Email: cibsen@redhat.com
> Web: http://fusesource.com
> Twitter: davsclaus
> Blog: http://davsclaus.com
> Author of Camel in Action: http://www.manning.com/ibsen
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message