Re: Origin based routing and Sourse Addresses in Routing Key Definition
Stanislav Ivanovich <[email protected]>
| Newsgroups | gmane.ietf.sigtran |
|---|---|
| Message-ID | <[email protected]> |
Ilie,
Look at Brian's answer posted to you on 6th of December in http://www1.ietf.org/mail-archive/web/sigtran/current/msg05497.html
Here is the copy.
-------------------------------------------------------------------------------------------
Ilie,
An AS is a User (MTP-User, SCCP-User). In a peer-to-peer signalling
relation without relay or transfer, there are two Users, however,
there is only one signalling relation. Each end has an address.
the pair of addresses defines the signalling relation. A signalling
relation is an RK, and an RC.
As Tolga puts it well, at one IPSP, there is a 1:1:1 relationship
between an AS (local concept) and RC and an RK. Same at the other.
The RCs are the same and the RK's are symmetric (both define the
signalling relation from the view at only one end). If you know
one RK you know the other. Each end defines an AS (without relay)
and both form one signalling relation.
--brian
-------------------------------------------------------------------------------------------
Even more, regardless of originating addresss which are mandatory in SE model (according to Brian's and Tolga's answers) the most relay unfriedly principle still remians:
An applicatiion process (ASP or IPSP) must node SEND data within an application context (RC) for which it has not activated itself in.In aniother words an application at a process ASP/IPSP is dead (and is not allowed to send traffic) until the process activates itself for the AS.
In another words (as Tolga and Brian said a couple of weeks ago) an AS is not a sink but an entitiy which both receives and generates traffic.
/ Stanslav
Ilie Glib <[email protected]> wrote:
Hello Folks,
recently in "ASP Capabilities value 0x2 of interworking field" thread
On 12/20/05, Brian F. G. Bidulock wrote:
> Stanislav,
>
> Stanislav Ivanovich wrote: (Tue, 20 Dec 2005 07:44:05)
> >
> >
> > ....
> > In xxUA networks AS/RC concept in mandatory SE model both originating
> > and destination addresses have to be specified and only
> > these combinations are allowed. This is extremely unfriendly for
> > relay!
>
> No they don't. Any routing key is possible. Dynamic registration does not
> support all possible routing keys. Provisioned routing keys can take any
> form. Also note that SUA registration supports multiple and ranges of
> addresses in a dynamically registered routing key. Not even M3UA restricts
> its routing keys to a single originating or destination point code. You
> can have one destination HLR point code and 400 MSC originating point codes
> if you'd like. Or visa versa. Or 400 of each.
>
> >
IMO having Originating Addresses in the definition of the AS and RK
was a very nice idea. Most probably the reason for it was to off-load
the destination Signalling Processes and have finer granularity in
traffic routing and distribution via origin based routing (due to low
capacity Signalling Processes).
In case of SE-IPSP when dynamic RKM is in use, ASs and RKs shall
contain both source and destination addresses as part of RK
definition. As long as Relay is not used having source and destination
addresses as part of RK does not imply routing based on origin, where
depending on source address traffic can go to different IP
destinations although the SS7 destination address is the same.
As consequence of SE-IPSP and dynamic RKM, the definition of the AS
has drifted to the a concept of signalling relation with one or more
applications on one side of it and one or more applications on the
other side. As far as I understand there was no objective to enforce
origin based routing as a mandatory feature for relay. However, having
SE mandatory, one can draw the conclusion that origin based routing is
mandatory for Relay nodes.
Brian's statement above "Provisioned routing keys can take any form"
is fine with me, and hopefully with everybody on this list, since it
was not objected by anyone.
I interpret it as "origin based routing is optional in Relay nodes".
Please can people on the list make clear whether it is mandatory or
optional for SUA relay to have routing based on source address.
I was not able to fine any statement in the RFC that makes routing
based on origin mandatory, except the fact that SE is mandatory.
Thank you in advance
Ilie
_______________________________________________
Sigtran mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/sigtran
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
_______________________________________________
Sigtran mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/sigtran