On 2015-02-28 23:13, Clement T. Cole wrote:
I should have added when Stan and I wrote the original IP/TCP for VMS @ Tektronix did not have routing in it (nor mail support). We gave to CMU who enhanced it and I assume added routing. DEC did not support an IP stack until much later - Johnny probably remembers when it became available. I had stopped having to hacking on VMS when I left Tektronix in '81.
I don't know, actually.
However, it should be pointed out that routing is not the same as bridging...
Johnny
Sent from my iPad
On Feb 28, 2015, at 3:24 PM, Cory Smelosky <b4 at gewt.net> wrote:
On Sat, 28 Feb 2015, Johnny Billquist wrote:
Not sure what you mean with a star configuration. The first (proprietary) glass fiber repeaters were star designs, was that what you meant?
Not all ethernet segments have to be in one line. But since the maximum number of repeaters between any two nodes were two, you could (obviously) have repeaters in configurations that just made sure not more than two were involved in any given path, but there could be more than two totally.
The simplest such configuration would be a star.
How are you doing more than one line without bridges or repeaters? Have I misread?
(Not counting routing)
But my memory is fuzzy enough at this point that I should probably go read the docs instead of continuing to ramble here...
DEC also sold remote bridges and repeaters. A glass fiber trunc connected either two remote repeaters or bridges or one of each. I forgot how long a fiber segment could be, 2500 m IIRC. That gave you some room to plan on a large site. Two remote repeaters counted as one in the two repeater rule.
Expensive stuff though. A Lanbridge 100 was 30.000 guilders in 1988. A remote bridge was even more expensive.
Yeah.
But the ethernet was older than those devices. If my memory serves me right, the original repeater (from DEC) was the DEREP. Probably even more expensive back in the day. :-) And there were no bridges back then.
Hmmmm, when did VMS/BSD get software bridging capabilities?
Johnny
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
That is probably a story when the first thin wire repeaters were introduced. The h4000-aa wouldn't play nice when connected to a dempr or despr. The h4000-ba fixed that, or you could cut a capacitor (?) on the -aa and save money. A transceiver was 1200 guilders at the time.
Dec's thick wire repeaters (derep-ab and -rd) worked for me. One is still in use today...
Verzonden vanaf mijn BlackBerry 10-smartphone.
Origineel bericht
Van: Clement T. Cole
Verzonden: zaterdag 28 februari 2015 23:00
Aan: hecnet at Update.UU.SE
Beantwoorden: hecnet at Update.UU.SE
Onderwerp: Re: [HECnet] Maximum number of L2 routers?
Right. And many of the repeaters tended to not work well.
DEC had wholes name I forget - that was pretty much the only one that worked reliably. I have bad memories of repeater created issues where host A one seg1 just would not talk to host b on seg2 but c & d could talk to both and each other
Sent from my iPad
On Feb 28, 2015, at 2:28 PM, Johnny Billquist <bqt at softjar.se> wrote:
On 2015-02-28 21:21, Hans Vlems wrote:
Yeah of course you can have more than one L2 router in an area, didn't think of it :)
:-)
About the 10base5 segment, that was limited in length to 500 meters. Transceivers must be at least 2.5 meters apart, hence the 200 nodes. This was before a DELNI was invented. If my failing memory doesn't fail me that 200 node limit remained the same with Delni's.
But you were allowed to bridge several segments (using repeaters). I think the maximum distance, including hops, between any two nodes had to be within 1000m, but you could also do stars... You were not allowed to have more than two repeaters I also seem to remember.
Johnny
Verzonden vanaf mijn BlackBerry 10-smartphone.
Origineel bericht
Van: Johnny Billquist
Verzonden: zaterdag 28 februari 2015 20:55
Aan: hecnet at Update.UU.SE
Beantwoorden: hecnet at Update.UU.SE
Onderwerp: Re: [HECnet] Maximum number of L2 routers?
On 2015-02-28 20:51, Hans Vlems wrote:
64:-)?
You can definitely have more than one L2 router per area... :-)
I don't know for sure, as you probably have guessed...
There is an ncp executor parameter called maximum broadcast routers, default value 32 iirc.
Is there an architectural limit, depends on what you mean by ethernet segment. A 10base5 segment was limited to 200 nodes. 64 L2 routers might put a hefty broadcast load on it. On an extended ethernet LAN there is likely no technical limit.
Really? I have some vague memory of some limit of an ethernet segment,
but I can't recall any details now. Why 200?
Johnny
Verzonden vanaf mijn BlackBerry 10-smartphone.
Origineel bericht
Van: Jordi Guillaumes i Pons
Verzonden: zaterdag 28 februari 2015 20:37
Aan: HECnet Mailing List
Beantwoorden: hecnet at Update.UU.SE
Onderwerp: [HECnet] Maximum number of L2 routers?
Hello, list,
I'm trying to remember what is the maximum number of area routers allowed in a DECNET Phase IV ethernet segment. Anyone of you have that information at hand?
On other news, the old macbook I was using as home server has died, after years of service beyond the call of duty. I'm setting up my stuff using several smallish ARM computers. To be specific, now I'm running a cubietruck and and Odroid-C1 (and a raspberry Pi as router/firewall). I'm having trouble with the net connectivity, so some yo-yo disconnects from area 7 should be expected. Not so hard as last sunday, but I'm still doing quite a lot of reboots.
BTW, the SG1 multinet link seems to be down again (unless it is a problem in my side, which is completely possible).
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
On Sat, 28 Feb 2015, Robert Jarratt wrote:
It might be worth just checking that it isn't the disk that is the problem.
Does the power keep cycling with the disk removed? What about if you put a
different disk in?
I've had the main drive spin down before randomly and several other issues indicating a PSU issue.
Regards
Rob
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
It might be worth just checking that it isn't the disk that is the problem.
Does the power keep cycling with the disk removed? What about if you put a
different disk in?
Regards
Rob
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On
Behalf Of Cory Smelosky
Sent: 28 February 2015 21:41
To: dectec at dectec.info
Cc: hecnet at Update.UU.SE
Subject: [HECnet] VAXStation 4000/60 PSU info
Afternoon all,
Anyone have schematics for the 4000/60 PSU? Mine's acting funny again.
THinking I should probably just recap it.
Harddrive kept spinning down, then back up...then right back down.
Powered the entire thing off, attempted to turn back on to no avail. Let
sit for
several minutes and it powered right back up. Drive seems to remain
spinning
now.
Any other specific components I should test and replace while in there for
the
recap?
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
I should have added when Stan and I wrote the original IP/TCP for VMS @ Tektronix did not have routing in it (nor mail support). We gave to CMU who enhanced it and I assume added routing. DEC did not support an IP stack until much later - Johnny probably remembers when it became available. I had stopped having to hacking on VMS when I left Tektronix in '81.
Sent from my iPad
On Feb 28, 2015, at 3:24 PM, Cory Smelosky <b4 at gewt.net> wrote:
On Sat, 28 Feb 2015, Johnny Billquist wrote:
Not sure what you mean with a star configuration. The first (proprietary) glass fiber repeaters were star designs, was that what you meant?
Not all ethernet segments have to be in one line. But since the maximum number of repeaters between any two nodes were two, you could (obviously) have repeaters in configurations that just made sure not more than two were involved in any given path, but there could be more than two totally.
The simplest such configuration would be a star.
How are you doing more than one line without bridges or repeaters? Have I misread?
(Not counting routing)
But my memory is fuzzy enough at this point that I should probably go read the docs instead of continuing to ramble here...
DEC also sold remote bridges and repeaters. A glass fiber trunc connected either two remote repeaters or bridges or one of each. I forgot how long a fiber segment could be, 2500 m IIRC. That gave you some room to plan on a large site. Two remote repeaters counted as one in the two repeater rule.
Expensive stuff though. A Lanbridge 100 was 30.000 guilders in 1988. A remote bridge was even more expensive.
Yeah.
But the ethernet was older than those devices. If my memory serves me right, the original repeater (from DEC) was the DEREP. Probably even more expensive back in the day. :-) And there were no bridges back then.
Hmmmm, when did VMS/BSD get software bridging capabilities?
Johnny
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
Unix certainly could bridge and route with 4.1a after Sam wrote the routed stuff which he will tell you was ripped off from a Xerox routing protocol. My memory is hazy here. I think he started that work with BBN stack which a couple of us ran on 4.1 before joy created sockets. I have memories of debugging that with Eric cooper and Sam because they were in Evans and I ran the link in Cory in the CAD group at the time as one of our 3 780s was on both the Cory LAN and the other end of a link to Evans [3meg Ethernet not even 3com or interlan he in those days]. My memory is that the other end of that Evans link was the system KIM which was on the Evans internal LAN and the link between the buildings.
Sent from my iPad
On Feb 28, 2015, at 3:24 PM, Cory Smelosky <b4 at gewt.net> wrote:
On Sat, 28 Feb 2015, Johnny Billquist wrote:
Not sure what you mean with a star configuration. The first (proprietary) glass fiber repeaters were star designs, was that what you meant?
Not all ethernet segments have to be in one line. But since the maximum number of repeaters between any two nodes were two, you could (obviously) have repeaters in configurations that just made sure not more than two were involved in any given path, but there could be more than two totally.
The simplest such configuration would be a star.
How are you doing more than one line without bridges or repeaters? Have I misread?
(Not counting routing)
But my memory is fuzzy enough at this point that I should probably go read the docs instead of continuing to ramble here...
DEC also sold remote bridges and repeaters. A glass fiber trunc connected either two remote repeaters or bridges or one of each. I forgot how long a fiber segment could be, 2500 m IIRC. That gave you some room to plan on a large site. Two remote repeaters counted as one in the two repeater rule.
Expensive stuff though. A Lanbridge 100 was 30.000 guilders in 1988. A remote bridge was even more expensive.
Yeah.
But the ethernet was older than those devices. If my memory serves me right, the original repeater (from DEC) was the DEREP. Probably even more expensive back in the day. :-) And there were no bridges back then.
Hmmmm, when did VMS/BSD get software bridging capabilities?
Johnny
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
Right. And many of the repeaters tended to not work well.
DEC had wholes name I forget - that was pretty much the only one that worked reliably. I have bad memories of repeater created issues where host A one seg1 just would not talk to host b on seg2 but c & d could talk to both and each other
Sent from my iPad
On Feb 28, 2015, at 2:28 PM, Johnny Billquist <bqt at softjar.se> wrote:
On 2015-02-28 21:21, Hans Vlems wrote:
Yeah of course you can have more than one L2 router in an area, didn't think of it :)
:-)
About the 10base5 segment, that was limited in length to 500 meters. Transceivers must be at least 2.5 meters apart, hence the 200 nodes. This was before a DELNI was invented. If my failing memory doesn't fail me that 200 node limit remained the same with Delni's.
But you were allowed to bridge several segments (using repeaters). I think the maximum distance, including hops, between any two nodes had to be within 1000m, but you could also do stars... You were not allowed to have more than two repeaters I also seem to remember.
Johnny
Verzonden vanaf mijn BlackBerry 10-smartphone.
Origineel bericht
Van: Johnny Billquist
Verzonden: zaterdag 28 februari 2015 20:55
Aan: hecnet at Update.UU.SE
Beantwoorden: hecnet at Update.UU.SE
Onderwerp: Re: [HECnet] Maximum number of L2 routers?
On 2015-02-28 20:51, Hans Vlems wrote:
64:-)?
You can definitely have more than one L2 router per area... :-)
I don't know for sure, as you probably have guessed...
There is an ncp executor parameter called maximum broadcast routers, default value 32 iirc.
Is there an architectural limit, depends on what you mean by ethernet segment. A 10base5 segment was limited to 200 nodes. 64 L2 routers might put a hefty broadcast load on it. On an extended ethernet LAN there is likely no technical limit.
Really? I have some vague memory of some limit of an ethernet segment,
but I can't recall any details now. Why 200?
Johnny
Verzonden vanaf mijn BlackBerry 10-smartphone.
Origineel bericht
Van: Jordi Guillaumes i Pons
Verzonden: zaterdag 28 februari 2015 20:37
Aan: HECnet Mailing List
Beantwoorden: hecnet at Update.UU.SE
Onderwerp: [HECnet] Maximum number of L2 routers?
Hello, list,
I'm trying to remember what is the maximum number of area routers allowed in a DECNET Phase IV ethernet segment. Anyone of you have that information at hand?
On other news, the old macbook I was using as home server has died, after years of service beyond the call of duty. I'm setting up my stuff using several smallish ARM computers. To be specific, now I'm running a cubietruck and and Odroid-C1 (and a raspberry Pi as router/firewall). I'm having trouble with the net connectivity, so some yo-yo disconnects from area 7 should be expected. Not so hard as last sunday, but I'm still doing quite a lot of reboots.
BTW, the SG1 multinet link seems to be down again (unless it is a problem in my side, which is completely possible).
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
On Sat, 28 Feb 2015, Johnny Billquist wrote:
How are you doing more than one line without bridges or repeaters? Have
I misread?
(Not counting routing)
I was referring to repeaters. They are totally invisible to the ethernet, so I do not consider to cables connected by a repeater to be two different ethernet segments.
Ahhhh.
But if others do consider that separate segments, then we atleast understand each other. :-)
Yeah. I understand you now.
But the reality of it is that on your ethernet connection, you can see more than 199 other machines, even without switches or bridges.
The 500m limit of the ethernet cable is just because of signal quality. The design as such put an upper limit on how far apart two hosts can be, in time, for the collision detection to work properly. The is a hard limit.
The 2.5 meter between transcievers is also more of an effect of physics. :-)
Physics always get in the way!
Hmmmm, when did VMS/BSD get software bridging capabilities?
No idea. Can VMS do bridging now?
Not sure, that's why I was asking. ;)
Johnny
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
Afternoon all,
Anyone have schematics for the 4000/60 PSU? Mine's acting funny again. THinking I should probably just recap it.
Harddrive kept spinning down, then back up...then right back down. Powered the entire thing off, attempted to turn back on to no avail. Let sit for several minutes and it powered right back up. Drive seems to remain spinning now.
Any other specific components I should test and replace while in there for the recap?
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
Bridges and repeaters, DEB ETs and DEREPs, are proprietary dedicated devices.
Verzonden vanaf mijn BlackBerry 10-smartphone.
Origineel bericht
Van: Cory Smelosky
Verzonden: zaterdag 28 februari 2015 22:24
Aan: hecnet at Update.UU.SE
Beantwoorden: hecnet at Update.UU.SE
Onderwerp: Re: [HECnet] Maximum number of L2 routers?
On Sat, 28 Feb 2015, Johnny Billquist wrote:
Not sure what you mean with a star configuration. The first (proprietary)
glass fiber repeaters were star designs, was that what you meant?
Not all ethernet segments have to be in one line. But since the maximum
number of repeaters between any two nodes were two, you could (obviously)
have repeaters in configurations that just made sure not more than two were
involved in any given path, but there could be more than two totally.
The simplest such configuration would be a star.
How are you doing more than one line without bridges or repeaters? Have I
misread?
(Not counting routing)
But my memory is fuzzy enough at this point that I should probably go read
the docs instead of continuing to ramble here...
DEC also sold remote bridges and repeaters. A glass fiber trunc connected
either two remote repeaters or bridges or one of each. I forgot how long a
fiber segment could be, 2500 m IIRC. That gave you some room to plan on a
large site. Two remote repeaters counted as one in the two repeater rule.
Expensive stuff though. A Lanbridge 100 was 30.000 guilders in 1988. A
remote bridge was even more expensive.
Yeah.
But the ethernet was older than those devices. If my memory serves me right,
the original repeater (from DEC) was the DEREP. Probably even more expensive
back in the day. :-) And there were no bridges back then.
Hmmmm, when did VMS/BSD get software bridging capabilities?
Johnny
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
