[nanomsg] Re: Loadbalancer with async backend
- From: Martin Sustrik <sustrik@xxxxxxxxxx>
- To: nanomsg@xxxxxxxxxxxxx
- Date: Fri, 10 Jan 2014 13:45:38 +0100
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 10/01/14 13:14, Laurent Alebarde wrote:
> Martin Sustrik <sustrik@...> writes:
>
>> Yes, you can. In fact nanomsg does that for you. How it works is
>> that the path back to the original requester is stored in the
>> message, but it's done under the hood so you don't see this
>> additional data associated with the message. You simply send and
>> receive messages and they get to the correct destinations
>> automagically.
>>
> Awsome ! It would be interresting to add in the API some plumbing
> functions to manipulate the channel ID stack. It would authorize
> one route for the request, and another one for the response.
That would break the req/rep model, no? The vary nature of the pattern
is that the reply is delivered to the original requester.
Or, if you meant that the reply can use different route to get to the
original requester, that's what IP is for... routing around the
failure and so on. Why duplicate the functionality on higher level?
> BTW, do you have a complete example code somewhere available with
> scemas please ?
What schemas?
Martin
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/
iQEcBAEBAgAGBQJSz+trAAoJENTpVjxCNN9Y+BYH/j1+e4QpIvMESibJOASDvPwb
e7QHdbnWZCBGMtKG3JaS/vxWeVx26c1Ii4rh1DLenCZFjFM3WK3ePWNBsYcxYFas
yGehPM17uh7LHu6vx3FtkclPRqcmJZWD8eXJnnfMvPi31j5Rn5P0+tDKSx8TV7Hw
JShCtURHkHHURXl25WJhr4MgnLSzdLmt3sqLMfRrSvGlSBCOPlt3KlBKu55zeviA
jv6Vupt70dGFCi8yU7SPB0wu9X1f8fMnCI0fpcf93FLr3nF58DYD/xGZy+FgqrGR
ZSkBgKUgLADozkvBhFuo0nIErTeLAWLf6Yjqeg3YkBElahlfXekcidOED6jgvQY=
=aahy
-----END PGP SIGNATURE-----
Other related posts:
