On Fri, 13 Oct 2000, Gerhard Sittig wrote:
On Fri, Oct 13, 2000 at 15:30 +0000, Geordon VanTassle wrote:
Oct 12 22:44:05 moat kernel: martian source 8e3ffea9 for fffffea9, dev eth1 Oct 12 22:44:05 moat kernel: ll header: ff ff ff ff ff ff 00 10 a4 aa da e2 08 00 Oct 12 22:44:05 moat kernel: Packet log: input REJECT eth1 PROTO=17 169.254.63.142:137 169.254.255.255:137 L=96 S=0x00 I=11008 F=0x0000 T=128 (#69)
This is some braindead MS system with a failed DHCP attempt or some other "auto config" (i.e. not knowingly configured) network stuff.
Somebody at MS had the bright idea to say "if we don't have / don't get an address, we simply dice us one." That's when they use the IPs you can see above. Watch the "wide" netmask and the broadcast for "the other unconfigured machines"! Seems MS wants to cope with ignorant environments and even run when not setup correctly ...
Sadly I couldn't find any reference to where those IPs come from. RFC1918 addresses are explicitely reserved at the registrars. Those MS fallbacks seem not to be. Do they collide in some future time with some other "customer"? Or are they registered and reserved and I'm just too blind to see?
Hmmm, I saw that failed DHCP attempts resulted in bogus IP addresses, never noticed that it was always the same (range of) ip adresses. Thanks for pointing that out. A whois lookup tells us that 169.254.0.0/16 are IANA addresses reserved for use with "Link Local Networks." I haven't found any more information about the purpose of that range, maybe someone else on this list knows. Stefan PS: hoping micro$oft will clean up it's act is like hoping your boss will quadruple your wages tomorrow. It's a very pleasant hope that many people entertain, but just not very realistic.