Conference azur::mcc

The only sample AM we provide with the toolkit is for DECnet Phase IV,
which uses NICE.  The purpose is to provide information on how to write
an AM, not to document a specific network protocol.

If your customer is using DNA CMIP, then they should probably use the
DNA5 AM rather than writing a new access module.

-- Erik

    Hi Erik,
    
    I questioned my customer on why she could not use the DNA5_AM and the
    following is what she sent back.  BTW, she is currently working other
    MCC issues with the engineering group around DECnis event sinking to an
    extensions node.  The company is Reuters.
    
    Walt
    
    
    
Hello,

What we would like to do is to define an entity "X" and define it as having
some child entities of the NODE entity and others which are application 
specific.  Also, we might want to implement the application specific 
entities as a private MIB extension accessible through the TCP/IP SNMP AM.   
We want to do this because CMIP/SNMP are expensive both in terms of network 
bandwidth and node resources.   These are both very sensitive issues to us 
and if we are only using a subset of the information obtained by an AM, why 
obtain all of it?  Also, we want to simplify the display so that the
operator only has 1 view of a node, not many to contend with.    We have
minimal experience concerning programming with DECnet OSI.  That is why I 
asked if you had a "sample" for a NODElike AM.  Is there a reason why there 
is only a sample provided for a NODE4like AM?  Any light that you can shed 
on this would be greatly appreciated.



                                        Thanks,


                                        Lisa DeMasi
        

Oh.  Apparently they are trying to get arround the difficulties in
managing multi-protocol entities by writing a combined AM.  Interesting
idea, but sounds like a lot of work.

As to the examples, as I said before, the toolkit is intended to document
how to write an AM, not how to program a certain protocol.  The fact that
the sample AM uses a "node4like AM" is completely coincidental.  We are
likely at some point to replace it with a more portable AM.  

I'll see if I can dig up any more information, but it is not likely
that we would be able to help someone write a new OSI AM; given our
current resources we are hard pressed to support our own.

-- Erik

    Thanks Erik,  any information you can provide would be greatly
    appreciated.
    
    Walt

    Hi Erik,
    
    RE: .3
    
    Can we give her a copy of (or atleast a partial copy) our OSI AM so she
    would have something to go on?
    
    I have also entered a note in the phasev conference but have received
    even less response from there.
    
    Thanks,
    Walt

Sorry it took so long to reply.  I had forwareded the original note
to someone else, who apparently mis-filed it.

Re-writing a DECnet/OSI AM is really not a good choice.
We have invested many years of engineering in development of our AM,
and it is probably the single largest module in the system.
In any case, we can not give away this source code as a training tool.
A better approach for the customer would probably be to write an FM to
provide the model they desire, and which would then call both the SNMP AM
and the DNA5 AM according to the services which it required.

If the customer needs further assistance in understanding this problem,
I would suggest that you contact Larry Wellington, who supervises
the NSM technical support and training group.  He might know of someone
who could assist in whatever development the customer chooses.

-- Erik

    Thanks Erik, I will persue it from that angle with the customer.
    
    Walt