[HECnet] Multinet alternatives ...

Bob, I can't speak to Tops-10 with much authority any more, but I believe that in order to get the Phase IV support in Tops-20, you have to be running version 5 or, more likely, version 6 or 7.? All of these require a KL class machine due to the requirement for extended addressing. When I last checked, SIMH was 'only' capable of emulating the KA, KI and KS PDP-10 processors, none of which will support more than a 256K virtual code space.? I believe MRC was able to fit version 5 into the KS, but those sources are not currently available (in fact, they may be lost). But, as he frequently said, the KL is only 'reasonable' processor as it has 23 bit virtual address space, which puts you into the megabyte range.? Later processors have a 30 bit address space, which gets you a gigabyte address space.? That's plenty of space unless you are doing multimedia. Thus, your only reasonable alternative is the KLH10 emulator which will do Phase IV over Ethernet and the CI.? It is superior to SIMH in certain respects as it has a multi-threaded model where devices can and do run concurrently to the processors.? However, certain things remain either unimplemented (such as the CI) or minimally implemented. This is relevant because the DECSYSTEM-20 architecture had a KL surrounded by up to four PDP-11's, one of which ran the DECnet code (branded a DN20).? These handled synchronous communications and did the DDCMP.? The problem is that the interface between the KL and the PDP-11, called a DTE for Data-Ten-to-Eleven, implementation is minimal.? This would need to be fleshed out both for KLH10 and SIMH and some sort of data path agreed to (one assumes UDP).? Then you could boot the DN20 code (I forget the internal name, maybe MCB) and have DDCMP from Tops-20. Sounds like fun, but an extensive amount of it... ??? ??? --T On 2/25/20 10:40 AM, Robert Armstrong wrote: