Concepting a new IPC (was Re: "DBus Embedded" - a clean break)

[Robin Bate Boerop]

On 31 January 2011 16:16, Alberto Mardegan <mardy at users.sourceforge.net> wrote:
>  I've been convinced that attempting to implement my ideas [0] as
> part of D-Bus makes more sense than starting a new project (which was
> my initial plan).

I'm glad that you have decided to contribute your implementation to
D-Bus, instead of creating a new project.

I have read your blog post. Your design is masterfully conceived. Well
done! It will solve or improve the performance-related problems with
D-Bus that you mentioned, along with many more that you didn't
mention, such as priority inversion and message size limits with
sockets.

> And for the far future, some yet to be refined ideas:
> - Make the proxy concept (such as DBusGProxy, GDBusProxy) part of the
> core client API, because it can be used to provide further
> optimizations to the p2p case

Arranging for the binding libraries to allocate their
already-serialized objects inside of the shared memory that is used to
transmit the object would be a great optimization. Save serialization
and save another object copy. Optimization for the far future,
perhaps.

> Also, what are your views on these proposed changes?

I am in favour of changes of the scope and magnitude that you suggest.
Having studied your blog post, I think that your design is an
excellent starting point for these changes. They are a great
improvement on the current D-Bus architecture. They will help D-Bus to
become an industrial-grade IPC system, instead of something that is
avoided by those who are performance-sensitive.

-- 
Robin Bate Boerop