One question --- why can't one of the/additional root server ip addresses be a multicast address instead of all of them being unicast ones. ... then however that particular multicast address is served - DHT or some other structure... makes resolution extensible doesnt it?
In that case, the existing part of the DNS tree, needs to keep the translation of name to ip address that they are SURE are correct. So when a server hosting moves / or is in some way more agile ..from one part of the DNS to another, well instead of manually changing the ip address, its record is changed to map to the root multicast address. . In my dreams, it at least sounds cleaner than just dyndns.com endings to such names?... On 6/17/07, EdPimentl <[EMAIL PROTECTED]> wrote:
Really bad idea. -E On 6/17/07, David Barrett <[EMAIL PROTECTED]> wrote: > > Ah, heh, I was actually thinking along exactly the opposite lines. > > Rather than using a DHT to replace DNS, I was thinking of just using DNS > to > replace the DHT: > > Normally, every node would register itself in some DHT such that it can > be > located. I'm suggesting that every node just register itself using > dynamic > DNS for location via the internet (or using mDNS for location via a > disconnected LAN), and then there's no need for the DHT (at least, not > for > resolving peer names to IPs). > > Though now that I think of it... you could also do some really crazy > shit by > having a domain name like: > > <contenthash>.crazyshit.com > > Then, using standard dynamic DNS protocols, if you have a piece of > content > you just add yourself to the round-robin IP pool associated with that > domain > name. To get a list of all clients with the content, just do a standard > DNS > lookup on the domain. All the standard DNS infrastructure for > replicated > caching kicks in as normal. > > Anyway, the more I learn about DNS, the more I like it. It's one of the > few > systems that is truly decentralized at not only a technical > infrastructure > level, but also the legal/political level. > > -david > > > -----Original Message----- > > From: [EMAIL PROTECTED] [mailto:p2p-hackers- > > [EMAIL PROTECTED] On Behalf Of Alen Peacock > > Sent: Sunday, June 17, 2007 4:24 PM > > To: theory and practice of decentralized computer networks > > Subject: Re: [p2p-hackers] p2p, mDNS and scaling > > > > On 6/17/07, David Barrett <[EMAIL PROTECTED]> wrote: > > > > > > Indeed, I think mDNS + dynamic DNS is the recipe for a really > > interesting, > > > completely decentralized naming system that works seamlessly both on > and > > off > > > the internet: > > > > Some previous work on Semantic Free Referencing (SFR) and DNS might > > be of interest: "Untangling the Web from DNS" > > http://nms.csail.mit.edu/papers/sfr-nsdi04.html > > > > I'm not sure if this line of research ended up leading anywhere in > > the end, but thought the reference (if you hadn't already seen it) > > might be useful. > > > > Alen > > _______________________________________________ > > p2p-hackers mailing list > > [email protected] > > http://lists.zooko.com/mailman/listinfo/p2p-hackers > > _______________________________________________ > p2p-hackers mailing list > [email protected] > http://lists.zooko.com/mailman/listinfo/p2p-hackers > _______________________________________________ p2p-hackers mailing list [email protected] http://lists.zooko.com/mailman/listinfo/p2p-hackers
_______________________________________________ p2p-hackers mailing list [email protected] http://lists.zooko.com/mailman/listinfo/p2p-hackers
