My 'intranet' : FAQ

Index

Step 1 : Hardware
Step 2 : RISC OS
Step 3 : Windows 3.11
Step 4 : Actually doing it
Connecting to...
Miscellany
  • FAQ

 

1. So how does calling Argonet work?

As far as I can work out, it is Deep Magic™.
More helpfully, it does appear that different 'devices' can have different IP addresses assigned. In this way, I can run a system with the following:
  • ppp0: PPP serial connection to Argonet - 62.188.193.201 (session only)
  • eh0 : Network interface card drive - 192.168.0.1 or 'alyson' (always)
  • lo0 : Local loopback - 127.0.0.1 or 'loopback' (always); alias 'localhost'
A deeper explanation of Ethernet and dynamic IP is probably required if we are to make sense of it all. However, I would prefer not to have my brains splattered over my bedroom, so as far as we go here... It Works.

 

2. Why is 10-baseT more reliable than 10-base2?

10-base2 is much less reliable than 10-baseT because it is string of connections from one computer to the next, all rigged up with BNC twist-lock connectors. It is terminated. A break in the cable can kill the network (by varying degrees, it depends where/how the cable was broken). Losing a terminator may not be noticed, or it may send your network into heart attack.

On the other hand, 10-baseT uses a 'hub' to interconnect all of the machines. There is a direct link between each machine and the hub. This is pretty hairy for home installations, and shocking for new installations in older buildings; but new buildings where networking is envisaged tend to be 'flood wired' if the contract can afford it. That means, network wiring is laid on in the very infrastructure of the building so adding a network and moving machines around is not that expensive.
If there is a cabling fault, it will be obvious to the administrator where the fault lies. S/he needs simply look for a "Port OK" lamp that is not lit when it should be. And if a station is freaking out, the control on the hub usually given odd names like "Segment partitioning" can be used to temporarily disable a port. The rest of the network carries on happily.

 

3. So what was up with ShareFS?

To be honest, I'm not sure. RISC OS <-> Windows data transfer was fine. Communicating between two RISC OS machines with sockets was fine. ShareFS also worked right up until the second time you connect to the internet. Then all sorts of weird crap happened, the Freeway module picked up on the Internet IP address, and the Resolver just simply ceased to function.
I'm not sure exactly why, but repeated testing with ShareFS and IP addresses shows that it is totally and absolutely ShareFS and IP address 10.0.0.x that is at fault.
It has been suggested that x.0.0.x is for special purposes (consider broadcasts, or localhost on 127.0.0.1), but I do not know enough about the various layers involved, and how my i-cubed driver interprets them, to say for sure.
What I can say is after changing to 192.168.0.x, all the problems have gone and it all works totally smoothly. Yay!

 


Copyright © 2002 Richard Murray