Re: AS concepts and Relay in SUA and M3UA
Ilie Glib <[email protected]>
| Newsgroups | gmane.ietf.sigtran |
|---|---|
| Message-ID | <[email protected]> |
Brian, well, why RFC does not crarify it? Ilie On 12/20/05, Brian F. G. Bidulock <[email protected]> wrote: > Ilie, > > It does not say that the SGP cannot have a different AS Table per > association. > > --brian > > Ilie Glib wrote: (Tue, 20 Dec 2005 14:51:12) > > Brian > > > > what you say (RC values must be unique per AS within IPSP - IPSP > > relation and SGP-ASP relation) MUST be clarified in both SUA and M3UA > > RFCs. > > > > Currently the RFCs do not define exactly the scope of uniqueness of RC values. > > > > In fact they do and for instance SUA RFC says in Chapter 1.2 Terminology > > > > Routing Context - An Application Server Process may be configured to > > process traffic within more than one Application Server. In this > > case, the Routing Context parameter is exchanged between the SGP and > > the ASP (or between two ASPs), identifying the relevant Application > > Server. From the perspective of an SGP/ASP, the Routing Context > > uniquely identifies the range of traffic associated with a particular > > Application Server, which the ASP is configured to receive. There is > > a 1:1 relationship between a Routing Context value and a Routing Key > > within an AS. Therefore the Routing Context can be viewed as an > > index into an AS Table containing the AS Routing Keys. > > > > The last sentence is especially interesting and in fact backs my > > statement that a Signalling Process shall have as many RCs as there > > are remote SUA peers (ASes) it serves. > > One could draw a similar conclusion from M3UA definitions. > > > > Ilie > > > > On 12/20/05, Brian F. G. Bidulock <[email protected]> wrote: > > > Ilie, > > > > > > No, RC assignment is no different than M3UA: it only needs to uniquely > > > identify a traffic flow within an association. When there is only one > > > traffic flow per association, RC can be trivial (i.e. 0). > > > > > > Nevertheless, the sender of ASP Active need not include RC, ASP Active > > > Ack has a mandatory RC value which both peers can populate every subsequent > > > message requiring an RC. > > > > > > --brian > > > > > > Ilie Glib wrote: (Tue, 20 Dec 2005 14:19:18) > > > > Brian > > > > > > > > According to SUA RFC the RC value shall be unique (at least) within a > > > > Signalling Process, thus one RC value per "signalling relation" is > > > > needed. Consequently a Signalling Process shall have as many RCs as > > > > there are remote SUA peers (ASes). So one value is certainly not > > > > enough. > > > > > > > > Ilie > > > > > > > > On 12/20/05, Brian F. G. Bidulock <[email protected]> wrote: > > > > > Ilie, > > > > > > > > > > There is little difference between RC=0 in every message and an > > > > > optional Routing Context parameter. > > > > > > > > > > --brian > > > > > > > > > > Ilie Glib wrote: (Tue, 20 Dec 2005 10:06:15) > > > > > > Hello Brian, > > > > > > > > > > > > FYI the RC is mandatory parameter in SUA traffic messages. > > > > > > > > > > > > Ilie > > > > > > > > > > > > On 12/20/05, Brian F. G. Bidulock <[email protected]> wrote: > > > > > > > Stanislav, > > > > > > > > > > > > > > Stanislav Ivanovich wrote: (Mon, 19 Dec 2005 12:20:21) > > > > > > > > > > > > > > > > Brian, > > > > > > ... > > > > > > > > > > > > > > You are never required to assign RC to traffic flows between any SUA entities. > > > > > > > When no RC is assigned, and the association handles only one traffic flow, > > > > > > > it is optional in all messages. Don't use it if you don't see a use for it. > > > > > > > > > > > > > > > > > > > -- > > > > > > Ilie > > > > > > > > > > -- > > > > > Brian F. G. Bidulock > > > > > [email protected] > > > > > http://www.openss7.org/ > > > > > > > > > > > > > > > > > -- > > > > Ilie > > > > > > -- > > > Brian F. G. Bidulock > > > [email protected] > > > http://www.openss7.org/ > > > > > > > > > -- > > Ilie > > -- > Brian F. G. Bidulock > [email protected] > http://www.openss7.org/ > -- Ilie