> Yes, that was what I thought. But I received an answer from someone at
> IBM (regarding a DITTO problem between nodes) that indicated that a
> virtual CTC would work. I believe that person must have been mistaken,
> because I cannot see how! So, I approached the list for verification!

At one point in the past, Princeton University had a tool for extending
CTC connections over IP, written specifically for extending NJE
connections (ie, before RSCS had TCPNJE support). It worked very well
for that purpose. I did a little bit of tinkering with it at a previous
job and got it to support VTAM PU connections, but those mods are long
lost in the holocaust at the end of that employer's VM system, and now
that Princeton is mainframe-free, I suspect the source to that tool is
also lost. 

BCTCs are highly lock-step devices (ESCON CTCs are too, just different);
it's easy to stall one in a funny state you can't get out of without a
reset, or have to deal with missing interrupts. If you try to extend
them over a media that doesn't support guaranteed delivery in real time,
it gets *very* interesting. 

What are you trying to do with the ISLINK connection? SFS? If so, find a
copy of IPGATE, which will transport APPC over IP. We use it heavily for
sharing SFS pools between VM systems. Remote PERFKIT sessions appear to
work as well (if you actually understand enough about APPC to configure
the COMDIR entries correctly). 

Reply via email to