[SIPForum-discussion] SIP using TCP or UDP?

Fri Jan 19 21:51:09 UTC 2007

Hi Rommel,

Ok, I'll bite: "and use UDP unless explicitly stated and just fragment
the packet which the terminating UA would most probably reject", why is
it that most terminating UAs, evidently, would behave that way?...i.e.,
why don't they live over stacks that do packet reassembly?

Thanks,
BobN

________________________________

	From: discussion-bounces at sipforum.org
[mailto:discussion-bounces at sipforum.org] On Behalf Of Rommel Bajamundi
	Sent: Friday, January 19, 2007 4:16 PM
	To: 'Prasad MNR'; 'raman kumar'
	Cc: discussion at sipforum.org; rezaul.kabir at gmail.com
	Subject: Re: [SIPForum-discussion] SIP using TCP or UDP?

	I think RFC 3261 defines that the payload should be over 1300
bytes before falling back to TCP unless the MTU size is explicitly
known.  In an IMS implementation, I am fairly sure that you will be
pushing the edge of the 1300 byte limit, or the 1500 byte MTU size,
once you start adding in the codecs supported, long FQDNs, privacy,
multiple hops and elements adding themselves in the via.

	I would say listen to the standard, have it start out as UDP,
if you push the limit then fall back to TCP.  I have found that in
networks today TCP support for SIP is fairly spotty.  Support of
RFC3261 TCP fallback mechanisms even spottier, but it is the agreed
upon way of handling these situations.  Otherwise you're stuck with
RFC2543, and use UDP unless explicitly stated and just fragment the
packet which the terminating UA would most probably reject.

	Regards,

	Rommel

________________________________

	From: discussion-bounces at sipforum.org
[mailto:discussion-bounces at sipforum.org] On Behalf Of Prasad MNR
	Sent: Sunday, January 14, 2007 8:09 PM
	To: raman kumar
	Cc: discussion at sipforum.org; rezaul.kabir at gmail.com
	Subject: Re: [SIPForum-discussion] SIP using TCP or UDP?

	Hi,
	The following is a quote from the book "Understanding Sip" by
Alan B. Johnston.

	"UDP provides the simplest transport for user agents and
servers and allows them to operate without transport layer state.
however UDP offers no congestion control. A series of lost packets on a
heavily loaded IP link can cause retransmissions, which in turn produce
more lost packets and can push the link into congestion collapse. Also
UDP may only be used for SIP when the message (and its response) is
known to be less than the MTU size of the ip network. For simple SIP
message, this is not a problem. however for large messages containing
multiple message bodies and large header fields, this can be a problem.
In this case, TCP must be used, since SIP does not support
fragmentation at the SIP layer." 

	I think answers Madan's question and also explains why Raman
has suggested to use UDP if its a local network. I may be wrong at this
point of time as I have just started the journey of learning SIP.

	Regards, 
	Prasad MNR 

	On 1/15/07, raman kumar <ramank24 at gmail.com> wrote: 

		Razual has explained the right reson. To use the power
of UDP
		connectionaless protocol the
		timer mechanism is inbuilt in the sip 

		On 14/01/07, Rezaul Kabir <rezaulkabir at gmail.com>
wrote:
		> I have not heard any such restrictions on the size of
packet. But UDP
		> is used by default and if you closely look at RTC
3261 Section 
		> 13.3.1.4 you will notice that it talks about a timer
to retransmit
		> packets and similar example is available in the rfc
in several areas.
		> This behavior of retransmission clearly indicates
that it is desirable 
		> to use non-reliable transport such as UDP to send the
packets.
		> According to RFC3263 Section 4.1 is the Sip Uri does
not specify any
		> transport then UDP should be selected. Also by
RFC2543 UDP is
		> mandatory if no transport is defined and for good
interpretability.
		> TCP should only be used if one is thinking of using
SIPS/TLS OR the
		> Sip Uri specifies TCP as transport.
		>
		>
		> ---- 
		> Rezaul Kabir
		>
		>
		> On 1/13/07, madan kumar <mada2k at gmail.com> wrote:
		> > i far as i know udp is used when the packet size is
less than 1200 bytes 
		> and
		> > TCP is used when it exceeds 1200 bytes ..... do any
of u have a
		> explanation
		> > for this
		> >
		> >
		> > On 1/10/07, raman kumar < ramank24 at gmail.com
<mailto:ramank24 at gmail.com> > wrote:
		> > > when using the sip at the back-end use the UDP
		> > > When using in the local premises use TCP
		> > >
		> > >
		> > >
		> > > On 26/12/06, Ramachandran <
rgunasekaran at velankani.com> wrote:
		> > > > Dear all,
		> > > >
		> > > >
		> > > >
		> > > >  I am not very sure about the following, 
		> > > >
		> > > >
		> > > >
		> > > > What is ideal transport protocol for SIP with
IMS deployment?
		> > > >
		> > > > When to use SIP using UDP? 
		> > > >
		> > > >
		> > > >
		> > > > Thanks,
		> > > >
		> > > > Ram
		> > > >
		> > > >
		> > > >
		> > > >
		> > > >
		> > > _______________________________________________
		> > > This is the SIP Forum discussion mailing list
		> > > TO UNSUBSCRIBE, or edit your delivery options,
please visit 
		> > http://sipforum.org/mailman/listinfo/discussion
		> > > Post to the list at discussion at sipforum.org 
		> > >
		> >
		> >
		> > _______________________________________________
		> > This is the SIP Forum discussion mailing list
		> > TO UNSUBSCRIBE, or edit your delivery options,
please visit 
		> > http://sipforum.org/mailman/listinfo/discussion
		> > Post to the list at discussion at sipforum.org 
		> >
		> >
		> >
		>
		_______________________________________________
		This is the SIP Forum discussion mailing list
		TO UNSUBSCRIBE, or edit your delivery options, please
visit http://sipforum.org/mailman/listinfo/discussion
		Post to the list at discussion at sipforum.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sipforum.org/pipermail/discussion/attachments/20070119/3ab623f1/attachment-0002.html>