At 11:38 AM 1/26/02 -0800, Matt Schalit wrote:
>Schuyler Erle wrote:
[...]
>As far as the "roaming ip problem", can't several people's last 24
>bits of their MAC address be identical?
Yes, this is an interesting-looking project, and I too am glad to see it
with a presence on the LEAF list.
I saw this "roaming ip problem" in the NoCat stuff too and wondered about
it. Of course the last 24 bits can be identical, since otherwise the
Mac-address standard wouldn't need the first 24 bits. Schuyler can, of
course, give us the "real" answer, but here's my guess.
NoCat is using a trick here that basically plays the odds --
while NoCat is small, the chances of duplicate half-MAC addresses
are low, even globally
if (when?) NoCat gets large, the individual WiFi LANs are still
NAT'd, so there is only a conflict if 2 roamers with matching half-MACs are
on the same WiFi LAN at the same time. Not impossible, but sufficiently rare
that I see why NoCat might choose to shrug off the problem.
in the long run, I suppose IPv6 (remember IPv6?) will solve the
problem, since it uses addresses that are longer than MAC addresses.
A global solution under Ipv4 (I think) would be to match an IP address to
the user in the authentication database. It could even be a hash-table based
on the half-MAC address. But given the odds, this solution may be more
trouble than it is worth.
--
------------------------------------"Never tell me the odds!"---
Ray Olszewski -- Han Solo
Palo Alto, CA [EMAIL PROTECTED]
----------------------------------------------------------------
_______________________________________________
Leaf-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/leaf-devel
