On Mar 11, 2022, at 10:40 AM, Trevor Warwick
<twarwick(a)gmail.com> wrote:
I'm a recent joiner to HECnet, and wondered if anyone else has played around with
Phase V software at all ?
Obviously HECnet itself is a Phase IV network, but you can connect Phase V VAX endnodes
using the Phase IV compatibility mode.
All my systems are simulated with simh, and getting an Ethernet connection to work was
very straightforward, there's just a bit more local configuration required than with
Phase IV.
However, I spent quite a few years at DEC working on VAX synchronous device drivers, so I
was interested to see whether I could get a WAN connection going. The biggest issue was
device support. DECnet-VAX Phase V doesn't support any of the older microcoded
interfaces (DMC11, DMV11, KMX11 etc), and the later devices (DSV11, DSB32, DST32, etc)
aren't supported by simh. The only intersection is actually the DUP/DPV11, where there
was a new VMS driver (SEDRIVER) for Phase V, that replaces the individual
protocol-specific drivers that were previously used with these devices.
I've spent quite a long time this week dredging stuff out of memory from 30 years
ago, and still failing to get a DPV11 link to come up, until I eventually realised that a)
there are some significant differences between the DUP and DPV, and b) simh does not
support the DPV properly (how is one supposed to know this?) ! So after changing my
simulated microVAX into an 8600, and the DPV into a DUP, I've managed to get a DDCMP
link to come up between this and a PyDecnet router.
There still seem to be some rough edges, as it won't always come up until PyDecnet is
bounced, and it's also a bit slow for interactive use, as if there's something
that's only working on a retransmission, or timer expiry. Anyway, it's been a fun
nostalgia exercise to get this far, I'll probably mess about with it some more. Let me
know if you have any questions...
Could you send me traces of the PyDECnet issues you're seeing? I have never seen a
VMS DDCMP, but I've run the PyDECnet DDCMP against several others and it all works
without the sort of troubles you mentioned. I'd like to understand what the problem
is and if PyDECnet either needs a bugfix, or a workaround for a VMS bug.
Some day I may do at least partial Phase V in PyDECnet. Not soon, it's a big job.
paul