Performance Improvements...

[James Carman <ja...@carmanconsulting.com>]

Have we given any thought to addressing Johnzon's standings on all the
benchmarks that are out there comparing JSON libraries?  I am writing a
JAX-RS microservice framework (I won't hawk my wares here :) and I would
really like to switch to using standard JSON-B/JSON-B APIs instead of the
proprietary GSON library.  However, GSON is currently the hands-down winner
for the microservice use case (handles smaller payloads better than anyone
else) for all of the benchmarks I've seen.  Have we already started
investigating how we can give Johnzon more horesepower?

Thanks,

James

Re: Performance Improvements...

[Romain Manni-Bucau <rm...@gmail.com>]

Know jsonp part got optimized a lot by Hendrik. The mapper - based on jsonp
and that jsonb reuses - can have some optim but always revealed fast enough
for "normal" payloads if used in a http or backend stack (ie if there is
any IO).

On raw perf we are a bit slower than jackson IIRC but not bad....and we
dont biase the bench like genson can have done misbehaving in prod at least.

If you find anything we can always hack it if it does worth it.


Le 24 juin 2017 13:56, "James Carman" <ja...@carmanconsulting.com> a écrit :

> Have we given any thought to addressing Johnzon's standings on all the
> benchmarks that are out there comparing JSON libraries?  I am writing a
> JAX-RS microservice framework (I won't hawk my wares here :) and I would
> really like to switch to using standard JSON-B/JSON-B APIs instead of the
> proprietary GSON library.  However, GSON is currently the hands-down winner
> for the microservice use case (handles smaller payloads better than anyone
> else) for all of the benchmarks I've seen.  Have we already started
> investigating how we can give Johnzon more horesepower?
>
> Thanks,
>
> James
>