Looks like there's no DMC on the PDP-10, no DMC OR KDP on the MicroVAX,
and no KDP on the VAX780.
However, the PDP-11 has the DUP. Looks like I can use that as the
go-between.
MRC with the help of Stu Grosman did a phase 4 port to KS tops20. It
uses teh KMC/DUP, but has a limitation, as there was not enough memory
left, all buffers had to be in one page, the decnet MTU is not 576 but
376 or something like that. This does not matter unless you have two
interfaces and try to be a transit node. (it can be routein-iv).
It speaks DDCMP on the CYNC interface, so anything that speaks DDCMP
can talk to it, DUP,DQ, DMC DMR ...
-P
On Fri, 2 May 2014, Paul_Koning at Dell.com wrote:
On May 1, 2014, at 9:36 PM, Cory Smelosky <b4 at gewt.net> wrote:
On Thu, 1 May 2014, Cory Smelosky wrote:
Looks like there's no DMC on the PDP-10, no DMC OR KDP on the MicroVAX, and no KDP on the VAX780.
However, the PDP-11 has the DUP. Looks like I can use that as the go-between.
Wonderinf if this is an RSTS/E bug...or a simh bug.
Device XK0: does not interrupt - device disabled.
Probably a SIMH limitation. It s complaining about the KMC11. RSTS supports those only for use with the RJ2780 emulator; it doesn t use them with DECnet.
The message comes from the hardware scan code, where it looks around the bus looking for devices and makes them interrupt to learn what vector each uses. For the KMC11, it does that by loading a short program into it. That only works if the KMC emulation knows how to emulate a KMC well enough to run that program. If it emulates a KMC only as a KMC/DUP pair that speaks DDCMP, this won t work.
Ahhhh.
If you want a RSTS system to connect to your Phase III machine, you ll want to use a DMC (or DMR/DMP/DMV, they are all roughly interchangeable). In a sufficiently recent SIMH, the DMC emulation speaks real DDCMP so it should talk with a software DDCMP implementation, such as one that uses a DUP. Or, since it doesn t know sync from async, it will probably talk to a software DDCMP implementation that uses a terminal interface.
That I can do. Once I figure out the TOPS-20 DECnet configuration. ;)
paul
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects_______________________________________________
Simh mailing list
Simh at trailing-edge.comhttp://mailman.trailing-edge.com/mailman/listinfo/simh
BTW, this message went to the SIMH list as well since Cory s message was addressed to it, but it bounced ( not allowed to post to this list ). If it s useful for that list, please forward it.
paul
On May 2, 2014, at 10:36 AM, <Paul_Koning at Dell.com> <Paul_Koning at Dell.com> wrote:
On May 1, 2014, at 9:36 PM, Cory Smelosky <b4 at gewt.net> wrote:
On Thu, 1 May 2014, Cory Smelosky wrote:
Looks like there's no DMC on the PDP-10, no DMC OR KDP on the MicroVAX, and no KDP on the VAX780.
However, the PDP-11 has the DUP. Looks like I can use that as the go-between.
Wonderinf if this is an RSTS/E bug...or a simh bug.
Device XK0: does not interrupt - device disabled.
Probably a SIMH limitation. It s complaining about the KMC11. RSTS supports those only for use with the RJ2780 emulator; it doesn t use them with DECnet.
The message comes from the hardware scan code, where it looks around the bus looking for devices and makes them interrupt to learn what vector each uses. For the KMC11, it does that by loading a short program into it. That only works if the KMC emulation knows how to emulate a KMC well enough to run that program. If it emulates a KMC only as a KMC/DUP pair that speaks DDCMP, this won t work.
If you want a RSTS system to connect to your Phase III machine, you ll want to use a DMC (or DMR/DMP/DMV, they are all roughly interchangeable). In a sufficiently recent SIMH, the DMC emulation speaks real DDCMP so it should talk with a software DDCMP implementation, such as one that uses a DUP. Or, since it doesn t know sync from async, it will probably talk to a software DDCMP implementation that uses a terminal interface.
paul
On May 2, 2014, at 9:43 AM, Mark Benson <md.benson at gmail.com> wrote:
On 2 May 2014, at 14:33, Cory Smelosky <b4 at gewt.net> wrote:
On Fri, 2 May 2014, Brian Hechinger wrote:
Cory already covered ipv6 and cisco.
I've been thinking about setting Paul's python DECnet router up. I'll do that next week. I think it can do tunnel like things. Paul?
Can anyone direct me to this mystical Python creation? :)
svn://akdesign.dyndns.org/pydecnet/trunk/pydecnet
It needs Python 3. It also needs the Python 3 port of the daemon module, which you can find at svn://akdesign.dyndns.org/pydecnet/trunk/daemon. (I haven t heard from the owner of the daemon module, or I would have sent him the changes.)
Enjoy. There s some documentation there, and some sample config files. Please send questions, bug reports, and the like my way.
paul
On May 2, 2014, at 9:31 AM, Brian Hechinger <wonko at 4amlunch.net> wrote:
Cory already covered ipv6 and cisco.
I've been thinking about setting Paul's python DECnet router up. I'll do that next week. I think it can do tunnel like things. Paul?
It does Multinet and GRE. I discourage Multinet because it was designed by the clueless. GRE works well.
DECnet/Python does not itself do any tunnels. It will speak those protocols directly to the remote system. If you have to deal with NAT, the NAT gateway may be able to do the appropriate address mapping. If you need an actual tunnel, you ll have to set that up separately; for example, you could use IPSec if you want to.
DECnet/Python is work in progress, but at this point it should be good enough to do basic routing for Ethernet as well as GRE. It has SIMH 3.9 style data-only DMC emulation handling, but it doesn t yet do the newer full DDCMP support; that s on my to do list.
paul
El 02/05/2014, a les 14.23, Johnny Billquist <bqt at softjar.se> va escriure:
On 2014-05-02 14:10, Jordi Guillaumes i Pons wrote:
El 02/05/2014, a les 14.03, Johnny Billquist <bqt at softjar.se> va escriure:
Yay! Works. But damn, that was slow. :-)
Also, seems like FAL on Ultrix-32 have some bugs. :-)
Thanks. One more system checked against RSX with my new fix in place.
If you need to do more testing I can move the instance to a bigger host.
Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES
No worry. Slow speed is not a big problem. I'm just doing silly tests anyway.
If anyone have sources for DECnet for Ultrix, I can report that DECnet/Ultrix have an Y2K bug for file timestamps.
That is version V4.0, which is not Y2K compliant. 4.5 was patched for Y2K compatibility, but I have not been able to find the DECNET kit for it.
Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES
On May 1, 2014, at 9:36 PM, Cory Smelosky <b4 at gewt.net> wrote:
On Thu, 1 May 2014, Cory Smelosky wrote:
Looks like there's no DMC on the PDP-10, no DMC OR KDP on the MicroVAX, and no KDP on the VAX780.
However, the PDP-11 has the DUP. Looks like I can use that as the go-between.
Wonderinf if this is an RSTS/E bug...or a simh bug.
Device XK0: does not interrupt - device disabled.
Probably a SIMH limitation. It s complaining about the KMC11. RSTS supports those only for use with the RJ2780 emulator; it doesn t use them with DECnet.
The message comes from the hardware scan code, where it looks around the bus looking for devices and makes them interrupt to learn what vector each uses. For the KMC11, it does that by loading a short program into it. That only works if the KMC emulation knows how to emulate a KMC well enough to run that program. If it emulates a KMC only as a KMC/DUP pair that speaks DDCMP, this won t work.
If you want a RSTS system to connect to your Phase III machine, you ll want to use a DMC (or DMR/DMP/DMV, they are all roughly interchangeable). In a sufficiently recent SIMH, the DMC emulation speaks real DDCMP so it should talk with a software DDCMP implementation, such as one that uses a DUP. Or, since it doesn t know sync from async, it will probably talk to a software DDCMP implementation that uses a terminal interface.
paul
On 2014-05-02 16:12, Sridhar Ayengar wrote:
Johnny Billquist wrote:
Since there is no central point in the first place, the idea you seem to
be pursuing is meaningless. Besides, Update sits on unlimited traffic,
and with very high capacity links.
Ok. It was just a thought.
Well, you can still do it. I don't mind. It's just that it don't really improve things on a grand scale. But it could be useful for local people near you, or with special requirements that you could fulfill.
But from a pure traffic load point of view, there are no problems today. But since the whole thing is distributed anyway, it also means that you certainly can also contribute to the distribution.
So don't take me wrong. I did not intend to discourage you from chipping in.
Johnny
On 2014-05-02 16:10, Sridhar Ayengar wrote:
Johnny Billquist wrote:
How much traffic does all of Hecnet generate for a typical month,
roughly? Is it more than a couple hundred gigabytes?
No idea, but I seriously doubt it is anywhere near that.
But it is also almost impossible to measure. There are traffic going on
between different machines, local traffic within areas, local networks,
across the world, and everything in between, over various links.
I don't mind taking your word for it.
What do you all think? Do you think it might be a good idea to set up a
node on this colo to allow some (most?) people who have dynamic IP to
VPN into?
I will not stop you. People are free to connect more machines to HECnet, and if they route traffic, that is also ok. So just go ahead.
More access points in various places don't hurt.
Johnny
Johnny Billquist wrote:
Since there is no central point in the first place, the idea you seem to
be pursuing is meaningless. Besides, Update sits on unlimited traffic,
and with very high capacity links.
Ok. It was just a thought.
Peace... Sridhar
