exactly, and you can configure in opensips cfg whatever sequence of protocols to be tried (by doing serial forking).
Regards, Bogdan Daniel Goepp wrote: > If NAPTR was the method being used, but in this case is not setup. So > the problem I'm trying to solve is what to do when the endpoints and > the other gateways are not being explicit via either URI or DNS. The > endpoint is not specifying transport in the RURI, and DNS SRV is setup > and exists with entries for both TCP and UDP, but no NAPTR. In this > case, it's really leaving it up to the proxy to decide what transport > to prefer. > > -dg > > > On Thu, Apr 15, 2010 at 8:44 AM, Bogdan-Andrei Iancu > <[email protected] <mailto:[email protected]>> wrote: > > Daniel, > > So, the transport to be used is chosen as follows: > A) if "transport" in URI -> use it > B) if no port in URI, try to use NAPTR > B1) if NAPTR -> use the advertised protocols as per weight and > priority in NAPTR records (in the given order) > B2) if no NAPTR -> it is up to the proxy to choose the protocol > (obeying the URI scheme, like sips). > > I guess you are referring to B2 case, which you can do in script, > using > a serial forking scenario (forcing different protos via $du ). > > Regards, > Bogdan > > Daniel Goepp wrote: > > I did, but I don't see anything that says it would be a violation of > > SIP to try a number of transports in sequence, and to determine that > > sequence by the proxy. And I do understand that this is not a > problem > > with OpenSIPS, just trying to find a way to accomplish something new > > perhaps. We are working with some other networks which use a > > different method of selecting transport. And we are trying to > > implement a similar method to select transport using OpenSIPS. Very > > similar to how route advancing works, but at the transport level. I > > am working now to do it at the scripting level, but just thought > this > > might be something to consider actually implementing as > functionality > > at the application level. Perhaps it is just too unique of a > problem > > and not worth it. The problem I believe is UDP is clearly the most > > commonly used transport, but as devices get more capable (for > example > > in our situation with large SDPs for video, and traversing multiple > > proxies adding record routes), packet sizes get larger, and TCP > > becomes a more reliable transport. Additionally many folks we work > > with would prefer that is a secure connection is available, then it > > should be used. So the defaults on their network proxies will > attempt > > in the order of TLS, TCP then UDP to place a call. > > > > I will continue my work to try to get this going, but thought I > would > > post for comments here to get thoughts on the matter, or > > recommendations on how this would be best implemented using > OpenSIPS. > > > > Thanks. > > > > -dg > > > > > > On Thu, Apr 15, 2010 at 1:59 AM, Bogdan-Andrei Iancu > > <[email protected] <mailto:[email protected]> > <mailto:[email protected] <mailto:[email protected]>>> > wrote: > > > > Hi Daniel, > > > > > > Have you read RFC3263 about Locating SIP Servers (with proto > > selection) ? > > > > http://www.ietf.org/rfc/rfc3263.txt > > > > Regards, > > Bogdan > > > > Daniel Goepp wrote: > > > I had a previous question regarding default transport, and the > > > response I got was that the RFC says the order should be UDP, > > TCP then > > > TLS. However after reading into this further (and some recent > > > experiences with other platforms) I'm wondering if I have > a new > > > feature request for OpenSIPS. >From the RFC 3263 - > Section 4.1 > > > Selecting a Transport Protocol, I read it as saying: > > > > > > 1. If specified in the RURI it SHOULD use this (but doesn't > > have to) > > > 2. Bunch of stuff on NAPTR (which we are not doing) > > > 3. The section related to what we are doing: > > > > > > -----clip----- > > > If no NAPTR records are found, the client constructs SRV > queries for > > > those transport protocols it supports, and does a query > for each. > > > Queries are done using the service identifier "_sip" for SIP > > URIs and > > > "_sips" for SIPS URIs. A particular transport is supported > if the > > > query is successful. The client MAY use any transport > protocol it > > > desires which is supported by the server. > > > > > > This is a change from RFC 2543. It specified that a client > would > > > lookup SRV records for all transports it supported, and > merge the > > > priority values across those records. Then, it would > choose the > > > most preferred record. > > > > > > If no SRV records are found, the client SHOULD use TCP for > a SIPS > > > URI, and UDP for a SIP URI. However, another transport > protocol, > > > such as TCP, MAY be used if the guidelines of SIP mandate > it for > > this > > > particular request. That is the case, for example, for > requests that > > > exceed the path MTU. > > > -----clip----- > > > > > > The way I read this is that OpenSIPS MAY select whatever > > transport it > > > likes, and if no DNS SRV is found, then it would default to > > using UDP > > > first for SIP. But in our set, DNS SRV does exist, and there > > are both > > > TCP and UDP records. We would like to decide the default > transport > > > order to use, starting with TLS then TCP then UDP. I think I > > can try > > > to write something in the script to do this, but I'm not sure > > yet. I > > > don't want to rewrite the RURI, I just want to specify the > > transport. > > > It would be great if there was a global variable defined > in the > > config > > > that was something like "transport_order=TLS,TCP,UDP" > > > > > > Thoughts? > > > > > > -dg > > > > > > ------------------------------------------------------------------------ > > > > > > _______________________________________________ > > > Users mailing list > > > [email protected] <mailto:[email protected]> > <mailto:[email protected] <mailto:[email protected]>> > > > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > > > > > > > > -- > > Bogdan-Andrei Iancu > > www.voice-system.ro <http://www.voice-system.ro> > <http://www.voice-system.ro> > > > > > > _______________________________________________ > > Users mailing list > > [email protected] <mailto:[email protected]> > <mailto:[email protected] <mailto:[email protected]>> > > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > > > > > > ------------------------------------------------------------------------ > > > > _______________________________________________ > > Users mailing list > > [email protected] <mailto:[email protected]> > > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > > > > -- > Bogdan-Andrei Iancu > www.voice-system.ro <http://www.voice-system.ro> > > > _______________________________________________ > Users mailing list > [email protected] <mailto:[email protected]> > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > > > ------------------------------------------------------------------------ > > _______________________________________________ > Users mailing list > [email protected] > http://lists.opensips.org/cgi-bin/mailman/listinfo/users > -- Bogdan-Andrei Iancu www.voice-system.ro _______________________________________________ Users mailing list [email protected] http://lists.opensips.org/cgi-bin/mailman/listinfo/users
