On Mon, Mar 4, 2013 at 11:05 AM, Noel Chiappa <[email protected]> wrote: > I hear you, but Rome wasn't built in a day. Before the RIRs (for addresses), > and the registrar/registry system for DNS names, there were simpler things; > the current systems weren't brought forth full-grown, like Athena from Zeus' > brow.
Yes, and look how much more accessible the domain name system is now. If you want LISP to flourish, then you should be thinking about how to make it more accessible. The community of researchers, developers, and users LISP has today is not sufficient to solve some of its technical problems relating to making it possible to scale up. Increasing the community of users and businesses vested in LISP's potential success will increase its chances of fulfilling its goals -- like making it able to scale up to become an alternative to today's DFZ technologies. > lacks: some of which we may know about, others we probably don't. But any we > don't know about I personally would very much like to hear about. (Although > that's probably a new thread.) > > But just out of curiousity, what's the issue with negative cache entries? I > haven't ever thought about them, so you may have seen something worth hearing > about. I believe that you and I have discussed this issue on this WG list before. See my posts from ~2011 on negative cache entries. It is TRIVIAL to reduce a serious problem with negative cache entries by 90%. Maybe my suggestion, that makes the problem order-of-magnitude better, isn't even the best fix. Someone else could think of a superior solution. As it stands, the LISP community has refused to even acknowledge the problem exists or is worth solving. Back to the EID block issue, imagine I'm godaddy.com and I have the opportunity to sell numbers the way I sell letters now. If I think I can make money doing it, I will. And I'll make it as easy as possible for end-users. I'll build a scaled-up mapping service for my customers. I might even fund some fellowships for researchers like yourself to work on LISP. I'll definitely be vested in its commercial success, because I'll see it as a way to make money doing something I'm already good at -- essentially selling virtual goods in a user-friendly way. Maybe it will be learned that for-profit companies aren't good stewards of IP addresses. On the other hand, some folks have pretty serious complaints about the RIRs now, and they charge a lot of money for their virtual goods. They are also extremely difficult to work with, so the barrier-to-entry is artificially high. Given that this is a limited-time experimental block, the potential rewards from selecting one or more private, for-profit companies to register EID blocks seem to greatly outweigh the potential hazards. Besides that, the hazards are a good thing. If the worst happens (whatever that is) then valuable things will be learned. If it doesn't, LISP will have more businesses supporting it and more end-users who want it to work for them. Dino's assertion that EID block registration must be performed at no more than the cost of performing the service is myopic. The service should be provided with someone who has an interest in LISP's success, and who has experience at performing registration services, including authenticating companies and people. If you allow domain registrars to make some money by providing EID registration then they should certainly want it to be commercially-successful. On Mon, Mar 4, 2013 at 12:54 PM, Arturo Servin <[email protected]> wrote: > I though that LISPs users would be ISPs. If LISP succeeds in its goals then ISPs will be AMONG the users of LISP. There should also be branch offices, power-users, anyone who wishes to be multi-homed, mobile, or simply to avoid ever renumbering, ought to be interested in LISP. That's why it needs a superior way to scale and the reason the map caching system exists (even though it has challenges to overcome.) That's also why the community needs to develop experience with low-barrier-to-entry EID allocations to small users. > And honestly, I do not how a business would have a profit allocating > EIDs. How do you think godaddy profits by allocating letters? They do it by being efficient and user-friendly. That is exactly what you should want. -- Jeff S Wheeler <[email protected]> Sr Network Operator / Innovative Network Concepts _______________________________________________ lisp mailing list [email protected] https://www.ietf.org/mailman/listinfo/lisp
