is there any traffic on a silence tcp connection?` - TCP-IP

This is a discussion on is there any traffic on a silence tcp connection?` - TCP-IP ; supposing two ends, say a and b, established a tcp connection succesfully, from then on, both ends won't do any communication. my question is that is there any tranfic between the two ends after time t_a and before time t_b, ...

+ Reply to Thread
Results 1 to 14 of 14

Thread: is there any traffic on a silence tcp connection?`

  1. is there any traffic on a silence tcp connection?`

    supposing two ends, say a and b, established a tcp connection
    succesfully, from then on, both ends won't do any communication. my
    question is that is there any tranfic between the two ends after time
    t_a and before time t_b, where t_a is the time of connection
    established, and t_b is the time of ( for some reason ) the connection
    disconnected?

    my answer is no, but i still want to get confirmation from you.
    thanks.

    -
    woody


  2. Re: is there any traffic on a silence tcp connection?`

    "Steven Woody" writes:
    > supposing two ends, say a and b, established a tcp connection
    > succesfully, from then on, both ends won't do any communication. my
    > question is that is there any tranfic between the two ends after time
    > t_a and before time t_b, where t_a is the time of connection
    > established, and t_b is the time of ( for some reason ) the connection
    > disconnected?
    >
    > my answer is no, but i still want to get confirmation from you.
    > thanks.


    There might be. Google for "tcp keepalive."

    --
    James Carlson, KISS Network
    Sun Microsystems / 1 Network Drive 71.232W Vox +1 781 442 2084
    MS UBUR02-212 / Burlington MA 01803-2757 42.496N Fax +1 781 442 1677

  3. Re: is there any traffic on a silence tcp connection?`


    James Carlson wrote:
    > "Steven Woody" writes:
    > > supposing two ends, say a and b, established a tcp connection
    > > succesfully, from then on, both ends won't do any communication. my
    > > question is that is there any tranfic between the two ends after time
    > > t_a and before time t_b, where t_a is the time of connection
    > > established, and t_b is the time of ( for some reason ) the connection
    > > disconnected?
    > >
    > > my answer is no, but i still want to get confirmation from you.
    > > thanks.

    >
    > There might be. Google for "tcp keepalive."


    thanks. is the 'so_keepalive' option default enabled? if so, how long a
    detecting packet will be send down to the line?

    -
    woody


  4. Re: is there any traffic on a silence tcp connection?`

    On 03 Aug 2006, "Steven Woody" wrote:

    > James Carlson wrote:
    >> "Steven Woody" writes:
    >> > supposing two ends, say a and b, established a tcp connection
    >> > succesfully, from then on, both ends won't do any
    >> > communication. my question is that is there any tranfic between
    >> > the two ends after time t_a and before time t_b, where t_a is
    >> > the time of connection established, and t_b is the time of (
    >> > for some reason ) the connection disconnected?
    >> >
    >> > my answer is no, but i still want to get confirmation from you.
    >> > thanks.

    >>
    >> There might be. Google for "tcp keepalive."

    >
    > thanks. is the 'so_keepalive' option default enabled? if so, how
    > long a detecting packet will be send down to the line?


    No, it's disabled by default, at least in the OSs I'm familiar of
    (Windows, Linux, Solaris, BSD, eCos).

    The keepalive timeout is usually quite long. The default in many
    implementations is 2 hours and is a system wide setting, not a per
    connection setting. I'm not sure what the rationale for such a long
    timeout is, but presumably it's left to the higher layer protocols to
    define what alive means and to keep the connection "alive."

    Dave

    --
    D.a.v.i.d T.i.k.t.i.n
    t.i.k.t.i.n [at] a.d.v.a.n.c.e.d.r.e.l.a.y [dot] c.o.m

  5. Re: is there any traffic on a silence tcp connection?`

    David Tiktin wrote:
    > On 03 Aug 2006, "Steven Woody" wrote:
    >
    > > James Carlson wrote:
    > >> "Steven Woody" writes:
    > >> > supposing two ends, say a and b, established a tcp connection
    > >> > succesfully, from then on, both ends won't do any
    > >> > communication. my question is that is there any tranfic between
    > >> > the two ends after time t_a and before time t_b, where t_a is
    > >> > the time of connection established, and t_b is the time of (
    > >> > for some reason ) the connection disconnected?
    > >> >
    > >> > my answer is no, but i still want to get confirmation from you.
    > >> > thanks.
    > >>
    > >> There might be. Google for "tcp keepalive."

    > >
    > > thanks. is the 'so_keepalive' option default enabled? if so, how
    > > long a detecting packet will be send down to the line?

    >
    > No, it's disabled by default, at least in the OSs I'm familiar of
    > (Windows, Linux, Solaris, BSD, eCos).
    >
    > The keepalive timeout is usually quite long. The default in many
    > implementations is 2 hours and is a system wide setting, not a per
    > connection setting. I'm not sure what the rationale for such a long
    > timeout is, but presumably it's left to the higher layer protocols to
    > define what alive means and to keep the connection "alive."
    >
    > Dave
    >
    > --
    > D.a.v.i.d T.i.k.t.i.n
    > t.i.k.t.i.n [at] a.d.v.a.n.c.e.d.r.e.l.a.y [dot] c.o.m


    thanks!


  6. Re: is there any traffic on a silence tcp connection?`

    sorry, i still want to ask what api used to set so_keepalive option,
    server set it or client set it? my man page seems lack of enought
    detail on this.

    thank you.

    -
    woody


  7. Re: is there any traffic on a silence tcp connection?`

    In article <1154655131.436690.209270@b28g2000cwb.googlegroups. com>,
    Steven Woody wrote:

    >sorry, i still want to ask what api used to set so_keepalive option,
    >server set it or client set it? my man page seems lack of enought
    >detail on this.


    If you don't already know the answer and haven't spent the little
    effort neede to check the Gooble archive of this newsgroup, you need
    the short answer. It is:

    Don't use operating system keepalives. Modify your application
    programs to send messages to each other. These messages could be
    NOPs or job-completion notices sent at the rate your application
    needs to know to restart the TCP connection, but no faster than the
    connection time-out or 5 or 10 minutes.

    For old local discissions, see all except the first hit in
    http://groups.google.com/groups?q=ke...p-ip&scoring=d


    Vernon Schryver vjs@rhyolite.com

  8. Re: is there any traffic on a silence tcp connection?`

    In article <1154655131.436690.209270@b28g2000cwb.googlegroups. com>,
    "Steven Woody" wrote:

    > sorry, i still want to ask what api used to set so_keepalive option,
    > server set it or client set it? my man page seems lack of enought
    > detail on this.


    Either or both. It causes that system to send keepalives, and has no
    effect on the other system. So whichever one wants to be able to detect
    connection failures can send it.

    --
    Barry Margolin, barmar@alum.mit.edu
    Arlington, MA
    *** PLEASE post questions in newsgroups, not directly to me ***
    *** PLEASE don't copy me on replies, I'll read them in the group ***

  9. Re: is there any traffic on a silence tcp connection?`

    In article ,
    David Tiktin wrote:

    > I'm not sure what the rationale for such a long
    > timeout is, but presumably it's left to the higher layer protocols to
    > define what alive means and to keep the connection "alive."


    The rationale is that temporary network failures should not cause
    applications to fail. TCP/IP is intended to be tolerant of network
    problems; if you're not actively communicating when the network is down,
    why would you need to be notified about it?

    TCP keepalives are therefore intended to detect hard failures and
    prevent these connections from lingering forever. 2 hours seemed like a
    good starting point for "forever".

    --
    Barry Margolin, barmar@alum.mit.edu
    Arlington, MA
    *** PLEASE post questions in newsgroups, not directly to me ***
    *** PLEASE don't copy me on replies, I'll read them in the group ***

  10. Re: is there any traffic on a silence tcp connection?`

    On 3 Aug 2006 08:25:54 -0700, "Steven Woody"
    wrote:

    >thanks. is the 'so_keepalive' option default enabled? if so, how long a
    >detecting packet will be send down to the line?


    It isn't enabled by default. And if enabled the interval is normally
    two hours. But the interval can be changed globally in the kernel.

    Emil

    --
    Philosys Software GmbH System Software Phone: +49 89/321407-40
    Edisonstrasse 6 is our Fax: +49 89/321407-12
    85716 Unterschleissheim Speciality EMail: egn@philosys.de
    Germany WWW: www.philosys.de

  11. Re: is there any traffic on a silence tcp connection?`


    Emil Naepflein wrote:
    > On 3 Aug 2006 08:25:54 -0700, "Steven Woody"
    > wrote:
    >
    > >thanks. is the 'so_keepalive' option default enabled? if so, how long a
    > >detecting packet will be send down to the line?

    >
    > It isn't enabled by default. And if enabled the interval is normally
    > two hours. But the interval can be changed globally in the kernel.
    >
    > Emil


    do you mean that individule program can only turn on/off ( via socket
    api ) the feature but dont has a way to change the interval of the keep
    alive packet?

    >
    > --
    > Philosys Software GmbH System Software Phone: +49 89/321407-40
    > Edisonstrasse 6 is our Fax: +49 89/321407-12
    > 85716 Unterschleissheim Speciality EMail: egn@philosys.de
    > Germany WWW: www.philosys.de



  12. Re: is there any traffic on a silence tcp connection?`

    Steven Woody wrote:
    > do you mean that individule program can only turn on/off ( via
    > socket api ) the feature but dont has a way to change the interval
    > of the keep alive packet?


    Depends on the TCP stack. Some may have additional setsockopts to
    controll some of those things - if they do, they will _probably_ be
    documented in the "tcp" manpage on the system. However, relying on
    such things is definitely "non-portable" and brings us, once again ,
    to the suggestion that an application wanting "keepalive"
    functionality should implement it's own rather than rely on
    SO_KEEPALIVE.

    rick jones
    --
    denial, anger, bargaining, depression, acceptance, rebirth...
    where do you want to be today?
    these opinions are mine, all mine; HP might not want them anyway...
    feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...

  13. Re: is there any traffic on a silence tcp connection?`


    Rick Jones wrote:
    > Steven Woody wrote:
    > > do you mean that individule program can only turn on/off ( via
    > > socket api ) the feature but dont has a way to change the interval
    > > of the keep alive packet?

    >
    > Depends on the TCP stack. Some may have additional setsockopts to
    > controll some of those things - if they do, they will _probably_ be
    > documented in the "tcp" manpage on the system. However, relying on
    > such things is definitely "non-portable" and brings us, once again ,
    > to the suggestion that an application wanting "keepalive"
    > functionality should implement it's own rather than rely on
    > SO_KEEPALIVE.
    >
    > rick jones
    > --
    > denial, anger, bargaining, depression, acceptance, rebirth...
    > where do you want to be today?
    > these opinions are mine, all mine; HP might not want them anyway...
    > feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...


    thank you!


  14. Re: is there any traffic on a silence tcp connection?`

    On Fri, 04 Aug 2006 19:13:33 GMT, Rick Jones
    wrote:

    >However, relying on
    >such things is definitely "non-portable" and brings us, once again ,
    >to the suggestion that an application wanting "keepalive"
    >functionality should implement it's own rather than rely on
    >SO_KEEPALIVE.


    Agree fully!

    Emil

    --
    Philosys Software GmbH System Software Phone: +49 89/321407-40
    Edisonstrasse 6 is our Fax: +49 89/321407-12
    85716 Unterschleissheim Speciality EMail: egn@philosys.de
    Germany WWW: www.philosys.de

+ Reply to Thread