Re: Multiple SUA SGs, sending for SSP from SGs

Stanislav,

I think you are still way off base by thinking that the AS is so foreign
to SS7.  Think of an RK:AS:RC 3-tuple as identifying an MTP-SAP or
SCCP-SAP.

An SCCP-User at an SCCP-SAP does track the state of remote SCCP-SAPs
(Users) of which it is concerned (i.e., that it sends to), and those
are, in all but the most exceptional of cases, the SCCP-SAPs that send
to it.  Nevertheles, the RK:AS:RC simply identifies the SCCP-SAP and is
therefore, really just an SCCP-SAPI.  For many SS7 implementations, the
SCCP-SAPI is just Network Identifier, Point Code and Subsystem Number.
For SUA it can also be IP Address and be restricted further to a Global
Title range.

The same is true for M3UA.  An MTP-User at an MTP-SAP does track the
state of remote MTP-SAPs (Users) (again to which it sends), and those
are, again, in all but the most exceptional of cases, the MTP-SAPs that
send to it.  Nevertheless, the RK:AS:RC simply identifies the MTP-SAP
and is therefore, really just an MTP-SAPI.  Also, for may SS7
implementations, the MTP-SAPI is a Network Identifier, Point Code an and
SI value.  For some, it is just a Network Identifier and Point Code.
For others it is a Network Identifier.

Just because you can define any Routing Key under the Sun does not mean
that all of them will be useful or even rational.  Certainly the most
useful ones are those that correspond to SS7 MTP-SAPs and SCCP-SAPs.
Finer granularity allows distribution of applications over a broader
number of hosts than is otherwise possible, without allocation of unique
point codes.  Because some National point code numbering spaces are
exhausted, providing the capability of distributing applications without
requiring point code assignment is a necessary evil.

I really don't see any limitations with the RK:AS:RC approach to
MTP-SAPIs and SCCP-SAPIs.  In fact, a single RK:AS:RC can identify
multiple MTP-SAPs or SCCP-SAPs and, therefore, better identifies the
user behind the SAPs instead of just the SAPs themselves, as would be
accomplished by only addressing individual SAPs.

Now, MTP Transfer can be large in scope (i.e., entire National
signalling networks at a primary STP), and a RK based on NA alone would
permit this.  SCCP, OTOH, is not so large in scope.  SCCP Relay can be
as granular as the coupling of specific connection sections for protocol
class 2 and 3 relay.

--brian

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