[nanomsg] Re: Name service experiments
- From: Paul Colomiets <paul@xxxxxxxxxxxxxx>
- To: nanomsg@xxxxxxxxxxxxx
- Date: Wed, 11 Sep 2013 22:52:48 +0300
Hi Alex,
On Wed, Sep 11, 2013 at 8:20 PM, Alex Elsayed <eternaleye@xxxxxxxxx> wrote:
> On Wednesday, September 11, 2013 12:42:33 PM Paul Colomiets wrote:
> Well, what about annotating some points on the graph to be _roles_ rather than
> individual nodes? With REQ/REP for instance, due to its stateless nature, the
> 'endpoints' could describe a _class_ of nodes rather than individuals?
>
> So for instance, you could have A(REQ) -> B(device) -> C(REP), but A and C are
> annotated as roles. Then _any_ node can join as 'A' or 'C', get the
> appropriate configuration for that _role_, and life goes on.
>
Yes. In the current implementation there is role "client" that is
fallback when nothing is matched. So anybody can connect as a client.
The replier side is more complex, since for the scheme on the example
worker needs at least to communicate to the name service the data
center the worker is in.
On the IRC we are more or less agreed that it's better to have
arbitrary key=value pairs to match rules against rather than host and
appname pairs.
I'm yet only starting to experiment with rule-based topology, so stay tuned.
--
Paul
Other related posts:
