Dynamic FDDI (was RE: [rescue] octane question)

[Dave McGuire]

On January 17, Loomis, Rip wrote:
> Hmm...someone edu-ma-cate me on something I've always
> wondered:
> 
> Can FDDI (or any token-passing network method) support
> dynamic config of newly-added nodes (ala DHCP?)  Not
> that it's an absolute requirement--but my laptop surely
> does love it.  And if FDDI can't support it, then could
> someone explain why DHCP is not an asset?

  We're talking about different layers here...you can run DHCP on FDDI.
Nothing wrong with that, man.  FDDI has MAC addresses, MAC-supported
broadcast addresses, MAC-supported multicast addresses, etcetera just
like ethernet.  In that regard (but in few others) it works in a very
similar way.

> (I think that the answer is that dual-connected FDDI is
> perfect for high-reliability[0] networks whose configs
> rarely change and not necessarily suited for networks
> with transient devices--but I would love a real answer.)

  FDDI is much better suited to networks with transient devices than
ethernet is.  FDDI has a real station management subprotocol (SMT)
that handles ring insertions and removals gracefully.  With ethernet,
you just plug in, and if the electrical disturbance caused by your
plugging-in screws up a packet, it winds up being treated like a
collision...because the damn thing isn't smart enough to tell the
difference!

      -Dave

-- 
Dave McGuire
St. Petersburg, FL         "Less talk.  More synthohol." --Lt. Worf