On Fri, 18 Oct 2002, Nick 'Sharkey' Moore wrote: > On Thu, Oct 17, 2002 at 12:54:34PM +0300, Pekka Savola wrote: > > > > Optimistic DAD is a useful optimization because DAD is far more > > likely to succeed than fail, by a factor of at least 10,000,000,000 > > to one[SOTO]. This makes it worth a little disruption in the failure > > case to provide faster handovers in the successful case, as long as > > the disruption is recoverable. > > > > ==> this is totally, and completely wrong. [SOTO] only provide analysis > > in *some* cases, in particular autoconfigured vs privacy addresses. > > I see what you mean. I need to make it clear that [SOTO] and I are > referring to strongly random addresses ... this is a requirement > for my Optimistic DAD draft anyway for exactly this reason.
Yep. > Do you this it is fair to say: > > DAD is far more likely to succeed than fail FOR RANDOMLY > AUTOCONFIGURED ADDRESSES, by a factor of at least 10,000,000,000 > to one[SOTO]. This sounds rather good, using the lower case, of course. What's included in "randomly autoconfigured addresses" is a bit blurry, though. Is it RFC3041 only? Does it include EUI64-based schemes, etc? > Do you think it would be advisable for me to add: > > Optimistic DAD MUST NOT be used for manually configured addresses > > ... because as you point out, manually configured addresses are > far more likely to fail and being a MobileIP type I'm mainly > interested in configuring CoAs anyway. That is sounds good to me. I'd also add a line of reasoning behind that, like: Optimistic DAD MUST NOT be used for manually configured addresses, as those are much, much more likely to have collisions. > > For manually assigned addresses, I believe the ratio is closer > > to 1:10 or 1:100 (unmeasurable, of course). > > I'd be really interested to know if anyone has any indicative > figures on MAC address collision: it is inevitable that somewhere > out there there are two adaptors with the same MAC address due > to human frailties, but how many? I don't have figures but Francis Dupont reported this happening once to him. It should also be noted that people *do* change MAC addresses manually (possibly leading to clashes), e.g. in internet exchange points per policy and in e.g. access lines (e.g. xDSL) if ISP expects to see only one MAC address. But these aren't usually a shared medium or config'd addresses are used, so this may not be that big a problem. Vendors who ship e.g. 4-port Ethernet adapters (e.g some Sun products) with the same mac address in each port are problematic of course (and this isn't the first time -- switches usually don't like it:-). If you connect two ports to the same segment.. -- Pekka Savola "Tell me of difficulties surmounted, Netcore Oy not those you stumble over and fall" Systems. Networks. Security. -- Robert Jordan: A Crown of Swords -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
