Subj : One entry point, multiple destination points.
To   : Eric Renfro
From : mark lewis
Date : Sat Sep 19 2015 02:40 pm


19 Sep 15 13:40, you wrote to me:

ER>>> Here's the situation. One of my downlinks wants to setup their IRex
ER>>> and FastEcho on thheir system to receive mail, one from me, one from
ER>>> another link.

ML>> should be a pretty standard setup...

ER>>> They run multiple BBS systems at the same time using some telnet2bbs
ER>>> link tool of some kind, I'm not entirely sure about that, but their
ER>>> BBS's support includes JAM message bases so..

ML>> ok... hummm... are they sharing the same JAM bases between those
ML>> BBSes?? do they all use the same FTN address? if not then they will
ML>> need a FE setup for each BBS instance... even if they are all on the
ML>> same machine... that means either copying packets destined for the
ML>> others to their inbounds or using some mailer to send to them as
ML>> normal... i have 4 or 5 systems here all using their own mailers for
ML>> this but they all also have their own FTN addresses as required...

ER> No sharing or JAM message bases, all completely independant of each
ER> other.

so effectively completely separate installations...

ER> Ideally, not sharing the same FTN address, but if that is possible,
ER> then it could be viable, though I'd think a point node would be better
ER> personally. Have one BinkD server act as the primary node for the
ER> whole list of internal BBS's, and then serve out to each point from
ER> there acting as the gateway. That reduces my redundant load, and puts
ER> them in proper control as a point-node should be, since he only runs a
ER> mailer on one system and one mailer, IRex, and all on Windows, which I
ER> haven't touched much for years. :)

you could do that but some folks tend to frown on points being bbses with
multiple users... ignorance at its best since nothing prevents it other than
opinion...

ER> So, here's the question of all questions then. How do you determine
ER> which PKT is for which address, or can you easily? If that's
ER> reasonably plausible then, hmmm.. I know sort of the concept, but I've
ER> never done this setup myself, always had many different systems to
ER> work with, or these days, multiple systems and multiple virtual
ER> machines. :)

if you're talking about my manually of packets, it depends on the mailer
format... for example, with fastecho and frontdoor, fastecho calculates a
bundle name used for each remote system... i forget the formula, though... on
my main system, i have a script that uses the 4DOS "describe" command to place
the destination system's address and system operator name in the file's
description as would be seen on a bbs... but with raw pkts, this cannot be done
so easily... pkts contain the destination address in their header so something
would have to peek into the header to see which system it is destined to... in
a BSO environment, there's the XXXXYYYY.pkt format which might be used but if
one's tosser uses different pkt names all the time, then one might analyze the
?lo files to see what pkts are included in the XXXXYYYY.?lo file... XXXX being
the hex net with leading zeros and YYYY being the hex node with leading
zeros...

but with all of this, it is much easier to simply use different mailers for
each individual system and let the sending mailer deal with which system mail
is destined for... they're built for this task ;)  so, like i have here, this
particular system has three separate instllations on it... there are three
separate mailers each running on their own port... i have another system here
that has another three separate systems running on it... they also have their
own individual ports and they're different than the ones on this system and
certainly different that the ports used on my main system... yeah, seven
complete and separate systems behind one IP... four bbses and three points...

for example only:
main system: all services on standard ports
system 1: telnet on 40023, binkd on 24555
system 2: telnet on 41023, binkd on 24556
system 3: telnet on 42023, binkd on 24557
system 4: telnet on 43023, binkd on 24558
system 5: telnet on 44023, binkd on 24559
system 6: telnet on 45023, binkd on 24560
system 7: telnet on 46023, binkd on 24561

with the above, say that system 1 is running SBBS with all services... all
would be using non-standard ports in the 40xxx range... web on 40080, ftp on
40020 and 40021, ssh on 40022, rlogin on 40513.

if system 2 is also running SBBS, then all of their services would be on
non-standard ports in the 41xxx range just like system 1's...

the ugly part is getting the domain stuff right and being able to pass to the
proper ports as well as ensuring that the firewall is port forwarding properly
AND that any connection tracking helpers the firewall may use are also
configured to recognise the additional ports (eg: iptables and tracking ftp
connections for established connections)...

once everything is set up, then you can easily send to their different systems
as easily as they can have their main system be a hub for their internal
systems... whether they are flying full node addresses or point addresses...

ER>>> Can anyone help me come up with a reasonably logical idea of how to
ER>>> set up this link to receive mail from me and internally relay it to
ER>>> several point point nodes for their multiple BBS systems they have
ER>>> running?

ML>> other than the above ""hints"" for multiple BBSes behind one IP, the
ML>> setup should be no different than any other distribution hub system...
ML>> the specific points being that each system has its own FTN address
ML>> (full node or points) and that there is a proper entry in the nodelist
ML>> pointing to each domain and/or IP address...

ER> Hmmm. Definitely a bit more difficult than I imagined. I keep wanting to
ER> utilize his linux system, but he's not so keen on that. But, with it, I
ER> could make it handle most everything fairly well, up until it has to still
ER> handle mail tossing to the various Windows system's BBS systems, then that
ER> situation comes right back to a system with multiple point nodes and BBSs
ER> to toss to. heh.

yep and the same thing even if the linux box was also used... one might run
virtual machines for each separate system but the same port assignment problem
is in place and running virtual machines means a virtual network and all that
additional setup mishmash rigamarole... yeah, i've got that here... one system
with a node and two points is running windows (vista) with no VMs and this
system is linux with numerous VMs and two of the separate systems here are
bbses in their own VMs complete with their own firewalls also running in VMs...

ER> Is it unreasonable to assign multiple nodelist entries to someone
ER> running multiple BBS's even if they're behind a single IP/mailer, and
ER> just have my side route all their IP's to their respective mailers?
ER> Coming from an NC perspective on that question. hehe

no, it is not unreasonable to assign multiple node numbers to one IP fronting
multiple systems... but don't think of one IP fronting one mailer unless that
mailer is going to be hubbing the mail for all the other systems... then you've
got to look at netmail routing, too, whereas with individual node numbers, the
routing is a cinch and there's not a lot to be done because it is no different
than any other routing...

is that all confusing enough for ya? ;)

)\/(ark

... That wasn't a bit nice.
---
* Origin:  (1:3634/12.73)

You are viewing proxied material from cvs.synchro.net. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.