12 Aug
2018
12 Aug
'18
8:20 p.m.
Hi,
This is the sort of information I'm after (see attachment). A basic area connectivity
map.
X is just the same area connectivity
Y means there's direct evidence of area adjacency
R means there's indirect evidence, ie. a->b implies b->a but I can't see the
area list from b
? means 'maybe' somewhere here if only I could see from another node
So area 1 (row) is adjacent to itself (X) and areas 2,3,5,7,8,14,19,28,30,39,42,59 and 62
(Y)
Keith
PS. I sent the map as a screen image of my putty terminal because of non-fixed-width
characters etc..
Keith
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On Behalf Of
Johnny Billquist
Sent: 12 August 2018 14:12
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] My RSTS/E system...
Hi.
On 2018-08-12 14:16, Keith Halewood wrote:
Hi,
Regarding the mapping, you may have noticed the odd NML (I think) originating from here.
I've been parsing the output of areas, circuit/node/area adjacency and then the
script, such as it is, hops around and attempts the 'next stage' of discovery via
tell <node of interest> etc..
NML is properly called, as a protocol, NICE. Network Information and Control Exchange, or
something like that.
Accessing information through that does not normally generate any logging information in
RSX, and does not bother me. But as I mentioned below, Cisco routers, for example, does
not implement this protocol, so you can't explore things beyond those points easily.
It has its obvious limitations - a snapshot of
currently connected, lowest cost adjacency (nodes and areas) but of course if a
'tell' is rebuffed by a node, then I can only infer.
By exploring adjacent nodes and known circuits, you'd get a more complete picture, as
it will also show you connections which are not the lowest cost paths.
However, as I said, it will only give a partial picture anyway. I also found out the other
day that it don't seem like RSTS/E implements it either, or maybe it depends on
version and maybe it's optional.
Not sure how much Tops-10 supports either. And it might be even more complex/limited with
odd implementations, like the Linux hack.
Johnny
Regards,
Keith
-----Original Message-----
From: owner-hecnet at Update.UU.SE [mailto:owner-hecnet at Update.UU.SE] On
Behalf Of Johnny Billquist
Sent: 12 August 2018 10:25
To: hecnet at Update.UU.SE
Subject: Re: [HECnet] My RSTS/E system...
On 2018-08-12 04:55, Mark Abene wrote:
--
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
Honestly mapping the whole network seems a little
redundant, since
Johnny already does that and the information is reasonably up to date
on his web site.
Well, that's not really true. I do not map, or try to keep some kind of map of
HECnet. I have my nodename database, which is entirely voluntary to register nodes in, and
I know that not all nodes on HECnet are in there.
In addition, this does not give much clue about how things are connected, and for some
machines, nothing more than the owner, name and number is registered.
That said, I do not think it is realistic to automatically try to map out HECnet anyway,
so I think that any such attempts should be thought about a second time, and probably be
abandoned.
There are several problems with trying any kind of mapping. The first is that there are
no proper protocol that is suitable. NICE would be the obvious first choice, but for
example Cisco routers do not provide NICE at all. And so, you can not traverse and follow
connections that way.
There are no other way of just exploring what connections exist, so then we instead have
to resort to an exhaustive search of all node numbers, using all kind of protocols a
machine *might* answer to, which will trigger all kind of logging events and possibly
security systems getting activated, which is not good.
Some kind of voluntary system is the only acceptable option I think. If people want to,
that could be a way. Possibly using FAL to explore and find other nodes, but it would have
to be based on lists of machines to probe, which would need to be manually kept up to
date.
Similar things have been attempted in the past, with mixed to poor results, I would say.
Johnny
-Mark
On Sat, Aug 11, 2018 at 7:41 AM, Supratim Sanyal <supratim at riseup.net
<mailto:supratim at riseup.net>> wrote:
Hi Mark,
I have DCL script generating this web page once a week - perhaps you
see this pinging your system? Or maybe I was looking for a guest
account many months ago for some reason I can?t remember.
http://sanyalnet-openvms-vax.freeddns.org:82/falserver/hecnet-status.
html
<http://sanyalnet-openvms-vax.freeddns.org:82/falserver/hecnet-status.
html>
Supratim Sanyal
Germantown, MD
On Aug 11, 2018, at 12:15 AM, Mark Abene <phiber at phiber.com
<mailto:phiber at phiber.com>> wrote:
OK, so who is "SANYAL" at node
1.550? If you'd like an account on
my RSTS/E system, all you need do is ask. I'm one of those sys
admins who actually reads his event logs.? :)
Regards,
Mark