Nick, Doing review of the draft now. It is good. No opinion yet. Don't get me wrong I think we need to solve this I just want to be careful how, because we got that ND stuff a runnin now and test deployed. But many of us have products now in 2nd and 3rd release. But O-DAD appears to be add on so that's cool. Just don't want to burn an interface without just cause so being cool as product engineer :--)
On the theoretical discussion. I think the entire layer 3 effort to all agree and standardize (Charlie and others I know clearly are testing) is far far from baked as standard. I do refuse to listen to folks who tell me we are close when I am on the mail list or reading the archive and 5 very very bright people all correct do agree, but none of them will compromise with each other technically. Then I say from 10 years in the IETF working and implementing that this is simply going to take awhile. On the ignore the standard diatribe. I simply believe the people who support my day job are now wanting technology that requires mobile ipv6, ipv6, ipsec/TLS, BGP+, TCP-Multicast-Fixes et al, SCTP, and lots of good technology created here, are simply tired of waiting for us in this community. If they have enough money and opportunity then they can attract the top vendors. I do know of such giants that are ready to discuss lets get vendor x,y,z and contract gurus a,b,c together and make this work. When we are done we can build specifications for standards bodies. That will happen soon. At that point they are not even listening to us here. L2 vs L3. This battle rages. I am positive we require layer 3 handover simply because I believe mobility requires IP to be totally successful and I would argue IPv6 and IPv4 is a dead end street. But per the baked discussion above we are far from agreement on hmipv6, LMM, and PKI (mostly in the market) and without these tid-bits I just don't see L3 happening first. Here is my vision of large $$$$$$ deployment of Mobility: 1. First there is none. Users accept to reconnect first but using IPv6. 2. Second. We get full products with MIPv6 and we still use L2 for handover. Similar to 3G but more 3G+ (I can't say 4G cause I still believe its marketing hype but maybe not I have stopped following it). 3. At some point in the future we finally get to L3 handovers. 4. I believe AAA will be as important as Ipsec and will play a big role in all three cases above to make it work and charge money for it :--) I think #3 best case for true deployment into large enough test beds that it is ready for real deployment (I roam in L.A. from home to airport and stay connected) and will scale across town via the routers, is best case 5 years away. I usually send first draft comments to just the authors and I hope to next week as I do have technical input for you. Regards, /jim -----Original Message----- From: Nick 'Sharkey' Moore [mailto:sharkey@;zoic.org] Sent: Thursday, October 17, 2002 2:45 AM To: Bound, Jim Cc: Alper E. YEGIN; [EMAIL PROTECTED]; Pekka Savola; [EMAIL PROTECTED] Subject: Re: Optimistic DAD draft ... On Thu, Oct 17, 2002 at 02:34:59AM -0400, Bound, Jim wrote: > From: Nick 'Sharkey' Moore [mailto:sharkey@;zoic.org] > > > Bound, Jim wrote: > > > > > > None of us working on this are even clear layer 3 handover will > > > ever work? Not sure if that matters does it? Are we talking > > > about the future? > > > We're pretty clear on this: we've tested it. > > Have you tested hmipv6 to the scale that mobility operator working > with ISPs can cover the entire metropolis of L.A.? To clarify: we've been doing some work to identify and isolate the causes of L3 handover delays, and they each seem solvable by one means or another. For example, hmipv6 offers one possible way of eliminating the RTT delay caused by sending BUs. In its current form, it probably isn't scalable to that level, but sadly I don't have a megapolis to test it with -- I'm a research engineer not a product engineer :-) Actually, JinHyeock reminded me that I'd forgotten to mention the Movement Detection Delay ... we've been tackling that with inter-layer signalling to prompt router solicitations. > No one has deployed layer 3 in production environment I am aware of? > Nor would I trust it yet simply because of this mail discussion. Ummm ... I hope it isn't my Optimism which has caused your concerns ... although if you've got technical comments or concerns re: my draft I'd love to hear them. It is, after all, a work in progress. The Optimistic DAD thing comes about because I thought the FH stuff was getting a bit theoretical and I felt that a set of small, easily implemented patches, one per cause of L3 handover delay, would be a good way to move forward in the short term. cheers, -----Nick -------------------------------------------------------------------- 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] --------------------------------------------------------------------
