juneau-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "John D. Ament" <johndam...@apache.org>
Subject Re: Does Hammock + Juneau make sense?
Date Thu, 30 Jun 2016 14:10:01 GMT
Hi James,

On Thu, Jun 30, 2016 at 9:51 AM James Bognar <james.bognar@salesforce.com>
wrote:

> Hi John,
>
> Would that be a merge of the microservice APIs?
>

Yes, I believe so.


>
> The Juneau Microservice API is the least mature part of the project.  It
> was developed less than a year ago as part of IBMs move to push Rational
> Team Concert into cloud deployments.  The microservices provided
> lightweight Jetty-based REST interfaces for communications within the
> various cloud components.
>


Similarly, Hammock provides Jetty integration, in addition to Undertow.  I
would like to add Tomcat to that list.  It uses JAX-RS for rest, but I
would like to expand that to support more tools, and wouldn't mind if we
built something.


>
> I'm not entirely sure the Microservice API is in it's final form.  As it
> developed, I noticed I was essentially reimplementing aspects of OSGi.  I'm
> now thinking that it makes sense to convert it into an actual OSGi
> Application.  I'd have to do some testing to see how "lightweight" an
> OSGi-based approach could be.
>

OSGi makes me cringe.  I would actually prefer we didn't look at OSGi short
term, but thats my own preference.


>
> I think we should consider the Microservice API to be "beta" and the
> architecture flexible.  So I'm totally open to discussions on changes to
> the API.
>

I'm glad to hear this.  Would it make sense for me to start thinking about
an SGA, or would it make sense for the IBM donation to start and see where
there's potential synergy?

John


>
> On Wed, Jun 29, 2016 at 6:42 PM, John D. Ament <johndament@apache.org>
> wrote:
>
> > Similar in spirit to Craig's recent email.  I'm wondering if it makes
> sense
> > to consider integrating, or even donation of, Hammock w/ Juneau.
> >
> > https://github.com/johnament/hammock
> >
> > Basically, I started Hammock > 2 years ago first conceptually as a way to
> > tie in the CDI model w/ a TCP server like Netty.  I was able to build out
> > some extensions w/ Resteasy and got a working prototype together.  I've
> > done some refinement, dropped Netty (not for any specific reason other
> than
> > HTTP all the things) and gotten it working for some basic use cases.  I
> > also ended up doing a JavaOne presentation on it.
> >
> > There was a recent discussion w/ a software vendor to fold it into their
> > portfolio.  Turns out the person I was speaking with wasn't actually in
> any
> > authority to do so, so while some public donations have occurred, no
> formal
> > convergence has happened.  So I still own the code.
> >
> > Now I'm wondering if it makes sense to somehow work Hammock and Juneau
> > together.  Hammock is basically microservices on the CDI stack.  It was
> > always intended for smaller deployments, SE based bootstrapping for a
> > container that launched some internal services (DB connections, servlet
> > containers, etc).  Is there some potential synergy here?
> >
> > John
> >
>
>
>
> --
> James Bognar
>

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