> CMIP is the "Common Management Information Protocol", best known as the
> SNMP-equivalent for SNA/APPN networks.  I'd say unless their Linux machine
> is going to be involved in SNA networking, it's irrelevant.

Actually CMIP has nothing specifically to do with SNA (the protocol) -- as
you say, it is an SNMP-workalike, but also works for any OSI derivable
protocol (commonly used in telco TDM and FR switches at layer 2, and for
applications monitoring in OSI networks, such as the few remaining
X.400-based mail systems or X.25 based PAD nets). It does also work for SNA,
though.

> > Do you know if they are currently supported in any way?  Is the
> > support different on Intel as opposed to zSeries?
> Perhaps the Linux-SNA code has some support for CMIP.

With ISODE installed, it does, but the support also extends to any supported
protocol stack on the same system as the CMIP agent.


> As for "TMN Framework", does that mean Tivoli?  Just guessing...

Nope. TMN = Telecommunications Network Management Framework. Its a
combination of protocols and best practices commonly used in telcos to
encourage/ensure interoperability of NMS and back-office systems between
carriers. It covers element management, trouble ticket exchange, upgrade
processes, and about 11,000 pages of other stuff. It's complicated.

Wrt to Dave Chase's earlier questions, Marshall Rose's ISODE (ISO
Development Environment) does provide support for most of the base OSI and
CMIP operations, and some components of TMN, but only on Intel and
PowerPC-based Linuxen, AFAIK.  It won't work on zSeries hardware with QDIO
interfaces without tunneling to a outboard Intel box because QDIO support is
currently IP-only.  LCS devices might work, but I don't have a easy way to
test it and ISODE is *huge* -- it takes 6-7 hours to compile on a fairly
large Intel box, and needs raw network access in a way that would be very
difficult to support for current zSeries hardware.  CIPs should work fine,
but don't have a way to test it.

Wrt to OSI, yes it's ugly, but it exists and wishing won't make it go away.
There are tools to deal with it, and it's no uglier than any other non-IP
protocol (even if it is screamingly less efficient).  It had some good
ideas, just lousy reality checking.

-- db



>
> Cheers,
> Vic Cross
>

Reply via email to