Re: Reexamining premises (was Re: UN plans to take over our job!)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



In message <433D8628.8030902@xxxxxxxxxxxxxxxx>, Michael Mealling writes:
>> <entire discussion by smart people deleted for brevity>
>
>Might I suggest all participants in this discussion figure out what you
>really want to use DNS for if you were to assume it didn't exist in the
>first place. Imagine going back in time to 1986 and explaining to
>everyone at the IETF the way things would develop and then, after
>they've stopped laughing, imagine what kind of system would have
>resulted. My personal suspicion is that two things would be very different:
>
>There wouldn't be one monolithic namespace/protocol/system. At least two
>systems would exist: one for hiding IP network layer topology from apps
>and another for describing and naming services for end users.
>
>The system that faced the users would be inherently trademark friendly
>and wouln't be hierarchical. The output of such a system wouldn't be an
>IP address but instead a complex record that described a compound object
>called a 'service'. It might be what people today call "peer to peer"
>(although I have yet to find a good definition of what that means) but
>that might not be an issue since the names wouldn't be hierarchical.
>
>What I find humorous is that this community's default position seems to
>be to attempt to play politics with those who are professionals at it
>rather than solving the problems with technology which is what you'd
>think we're good at....

There are several crucial attributes that are hard to replicate that 
way.  One is uniqueness: whenever I do a query for a name, I get back 
exactly one answer, and it's the same answer everyone else should get.  
This is the problem with "alternate" roots -- depending on where you 
are, you can get a different answer.  It's also what differentiates it 
from a search engine -- my applications don't know how to make choices.

Beyond that, the mapping should be under control of the appropriate 
party.  I don't want the moral equivalent to "Google-bombing" to be 
able to divert, say, my incoming mail.

Finally, you need locality: people within an organization must be able 
to create their own names.

It may be that some of these requiremets are fundamentally at odds with 
the notion of full decentralization. 

		--Steven M. Bellovin, http://www.cs.columbia.edu/~smb



_______________________________________________

Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf

[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]