RE: Recommendation for SUA modifications

Brian,

When an ASP sends a message to the SG (e.g., a CLDT) without a point
code in the source address, how does the SG know what to put in the OPC
of the MTP label?

If the SG can do this, it can also determine what the affected PC is of
the SCON coming from the same ASP.  

Regards,
Lincoln 

-----Original Message-----
From: [email protected] [mailto:[email protected]] On
Behalf Of Brian F. G. Bidulock
Sent: Thursday, October 13, 2005 5:14 AM
To: Tolga Asveren
Cc: [email protected]
Subject: Re: [Sigtran] Recommendation for SUA modifications

Tolga,

Tolga Asveren wrote:                          (Wed, 12 Oct 2005
14:34:27)
> > > "The ASP SHOULD NOT send Data or SSNM messages for the related 
> > > Routing Context(s) before receiving an ASP Active Ack message, or 
> > > it will risk message loss."
> > > I know it is not a "MUST NOT" but still it is nota good thing to 
> > > do unless one is aware of consequences and happy with them.
> >
> > That would be sending messages with RC 1 in them before receiving 
> > ASP Active Ack for RC 1.  That passage does not speak to the 
> > addressing of the messages.  Other passages do.  The SG provides 
> > uniform connectivity to all ASPs in an AS and can route any messagew

> > within a Network Appearance.
> [TOLGA]Here I still think different, i.e. I consider AS state to be 
> honored in both directions of message flow based on the above
passage-.

I know what I meant by that sentence when I wrote it.

One of the possible responses of the SG to an ASP Active request is an
ASP Inactive Ack, or management blocking.  My point was that if an ASP
blindly sends messages to the SG following ASP Active, without waiting
for an ASP Active Ack, that the SG may have to discard them in the case
that it responds with an ASP Inactive Ack or management blocking.

Therefore, if the ASP wishes to avoid risking "message loss", it will
wait for the ASP Active Ack.

If you recall, I was not successful on the argument of symmetrical flows
within a routing context.  RK is only used by an SG for routing messages
to the ASP, not visa versa.

If messages in M3UA could not be sent from ASP to SG with different
addresses than appear in the RK, then the SG would have to compare the
routing labels in DATA messages received from the ASP to police the
arrangement and do something (no procedure exists yet) in event of a
violation.  Currently, the SG can just process the message for M3UA,
because the addresses in it are fully specified and the message is self
contained.  The SG even knows the proper RC value to place in any SSNM
messages sent to the ASP in response to respond to the proper AS,
regardless of the addresses in the message.

In fact there is no procedure in either M3UA or SUA that relies on the
ASP applying a symmetric routing key to that of the SG.

In SUA placing a symmetric requirement on the ASP is not possible,
because not even addresses are symmetric (a transaction query often
includes GT, whereas the response does not).

All that the RC does is identify the AS to an SG and corresponds to a
routing key used by the SG to send messages to the ASP.  It is a false
assertion that missing parts of messages sent from ASP to SG can be
recreated based on RC.

--brian

--
Brian F. G. Bidulock
[email protected]
http://www.openss7.org/

_______________________________________________
Sigtran mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/sigtran

______________________________________________________________________
  This email message has been scanned by PineApp Mail-Secure and has
been found clean.