Your IP Address is: 38.107.191.82
RE: [moonv6] /120 prefix length at UNH
From: Bound, Jim (jim.bound@hp.com)
Date: 10/16/03
- Next message: Alain Durand: "Re: [moonv6] /120 prefix length at UNH"
- Previous message: Bound, Jim: "RE: [moonv6] /120 prefix length at UNH"
- Maybe in reply to: Alain Durand: "[moonv6] /120 prefix length at UNH"
- Next in thread: Alain Durand: "Re: [moonv6] /120 prefix length at UNH"
- Reply: Alain Durand: "Re: [moonv6] /120 prefix length at UNH"
moonv6 post from "Bound, Jim" <jim.bound@hp.com>
> At each end of each point to point link, there would be an
> interface. If /120 prefixes are use to number point to point
> links, those
> interfaces would be configured
> with addresses that use only 8 bits for interface ID. This is
> where the
> violation of 3513 is.
OK I agree with this concern. I would argue that is a bad implementation? It took the routing prefix and applied it to prefix configuration for the interface. Two different notions of prefix use. I relayed that in last mail to Ed.
Ok. I will support what you say that by using prefixes that are 001 (still not clear that not 000 is correct to say) could provide a problem to nodes on a ptp-pt link that will rely on the prefix length of the route to configure their interfaces.
But I still think that is a bad implementation. But at least we see where each of us was coming from and that is very valuable to me anyway. We did this a lot faster here than we would have in the IETF as a note :--)
I am now concerned that some implementations did use just 64 bits for prefix search in implementations now????
Should we define a test to just see if implementations can process 000 prefix to see if they can find other than 001 and have to check all 128bits? Not now but later? I think we have a moral Internet architecture responsiblity to do this is my .25 cents.
thanks
/jim
This archive was generated by hypermail 2.1.7 : 12/01/06 EST