[nanomsg] Re: Port sharing, tcpmux, web-related roadmap et c.

  • From: Bent Cardan <bent@xxxxxxxxxxxxxxxxxxxx>
  • To: nanomsg@xxxxxxxxxxxxx
  • Date: Wed, 19 Nov 2014 23:43:45 -0500

Thank you, Martin. That's great, with just a descriptor allocation, looks
like there's nothing incurred after the connection. Could we even stop the
daemon after the connection is established, or would that process need to
continue during normal transport? -Bent

On Wed, Nov 19, 2014 at 11:30 PM, Martin Sustrik <sustrik@xxxxxxxxxx> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hi Bent,
>
>
> > What sort of I/O model, blocking or async (I would imagine
> > implementing select and poll) presents the lowest average cost for
> > binding communication between the tcpmux daemon and its
> > corresponding recv function in nanomsg?
> >
> > Would the tcpmux daemon issue some sort of signal or call to
> > nn_recv?
>
> First of all, tcpmux daemon is involved only in the connection
> establishment phase. After that it hands the connection to the
> application and gets out of the way. Therefore, performance-wise,
> tcpmux is equivalent to simple tcp transport.
>
> If you want to look at the code, the connection hand-over is done here:
>
> https://github.com/nanomsg/nanomsg/blob/tcpmux/src/devices/tcpmuxd.c#L282
>
> Martin
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.11 (GNU/Linux)
>
> iQEcBAEBAgAGBQJUbW5/AAoJENTpVjxCNN9Y2ZEH/2NbNpjdSrbXZgJxQOylLhRE
> x1Us3ZDQFvrKyzmWN9iErXKJwIfh/p9yTx7hrrDw3UVi+e2SxcBbIrQ3K44S5W2/
> gyvJ+vzTStn1K4CmUcIJNeFeEX6PqlYQLdANbTYN90ANZuZxeYF4X7zz4Z85yzny
> NksFEpwUWm0/qfG2zC/Bl2uaNtYGPTgwh810iPRR/4wtFpL+p3Ir81VaLMRhdG+u
> By1dj4e3cL1PUP3oAlm259nt2z8Stns3fY/P5hvpFy05KLzQZXly9BZ3+EWWF7kX
> 14aqH3T4iP9dLkEgiCh+iyUKczW//UAM9gOskXwYQIhcuZkUcke1MexMzOrX2vA=
> =+4oc
> -----END PGP SIGNATURE-----
>
>

Other related posts:

  1. Home
  2. nanomsg
  3. Archive
  4. 11-2014