🚀 go-pugleaf

(Randall Atkinson) writes:
>
>In article <2115@devnull.mpd.tandem.com>, 
>	nath@mpd.tandem.com (Narendra M. Nath (Pi-Net)) writes:
>>> 
>>> Does anyone out there know about any existing MIB for SS7.
>>> If so please send me a copy
>

Then	<mschenk@research.ptt.nl> Mike Schenk writes:
>
>%  You mean managing SS7 with SNMP? That does not seem very appropriate
>% and I have not heard about anyone doing it. However, managing SS7 with
>% CMIS/CMIP is something completely different.
>
>  Why the hell not use SNMP ?  It is more secure than CMIP and it
>WORKS properly and interoperably and it is deployed in 100 times more
>sites than CMIP/CMOT is.  The only thing "completely different" is
>that using CMIP make NO SENSE technically, while using SNMP makes a
>hell of a lot of sense.  
>
Well one reason, is that many implementations of SNMP are not transport
level indepenendent. That's not to say that this situation can't be remedied
but until the majority of implementations are TL independent, they can only
manage TCP/UDP/IP based networks. Not exactly a winning situation when one is
trying to manage multiprotocol based networks. Ala, SS7. Proxy agents facilitate
such management, however unless the organization that one is implementing
network management functionality for is one that has an enterprise-specific
MIB registered with the Internet Assigned Number Authority (IANA) then the
the use of polling is the only management mechanism that is available. If a
large network is being managed, then this scenario may not be feasible.

As far as SNMP being more secure than CMIP is concerned I think that such
statements are quite silly. CMIP has not even been completed as far a standard-
ization is concerened (to my knowledge) and for that matter I don't think that
security for SNMP has been, either.  Security features via kerberos and DES are
in the works but that is different than the actual act of having been walked
thru the Internet standardization process and the acheivement of final Internet
Standard status. For that matter the two standards address different aspects of
network managemtent and there is nothing which inherently prevents the two
protocols from being implemented and utilized in an internet scenario. Their
utilization is not mutually exclusive.

>  No doubt this is why BELLCORE has published a number of MIBs for use
>with SNMP.
The beauty of MIB utilization is that the description, implementation and
management  of managed objects can be done in a fashion which is independent
of an underlying platform. ASN.1/BER define a presentation level abstraction
upon which an application level protocol can be based. SNMP, being an appl-
ication level protocol fits into this category.  Moreover, the use of SMI and
MIBs provides an ability to define managed objects (amoung other things) which
can be used between both Internet based management systems and ISO based net
management systems. Therefore, BELLCOREs choice of implementing managed agents
via SMI/MIB was not simply for utilization via SNMP. For that matter, ASN.1/BER
are ISO based standards.

"ANS.1 is destined, for better or worse, to become the network programming
language of the 90s, just as the C programming language is largely seen as
having been the systems programming language of the 80s. So, the choice of ASN.1
is a good one."   The Simple Book by Marshall Rose  

Ergo BELLCOREs choice.
-- 
||--  Douglas R. Donahue           --  GE Aerospace                         --||
||--  ddonahue@atl.ge.com          ||  Advanced Technology Laboratory       --||
--||  609 866 6614                 ||  Route 38 : Bldg 145-2                ||--
--||  8 * 777 6614                 --  Moorestown, NJ 08057                 ||--