i agree with bob here - mapping of identifiers to addresses is really
outside the basic routing of packets. In fact, one could go so far as to say
that there are other ways to identify end-points besides topologically
meaningful addresses although that method has the advantage of compression.

vint



On Sat, May 29, 2010 at 11:45 AM, Bob Frankston <
[email protected]> wrote:

> This gets to my point that the DNS should not be a base level mechanism but
> an application outside the network. The premise of the DNS is that the
> mapping of intent to destination should necessarily be done by translating
> the intent into a name and then looking up the name.
>
> This makes no more sense than presuming that everyone wants a reliable
> circuit. The decision on reliable delivery should be outside the network
> with best efforts being the more fundamental protocol because best efforts
> reflects the underlying reality.
>
> In the same way mapping of intent should not be forced into the
> reductionist
> mold of a single name. I recently ordered another copy of Co-Pilot for my
> iPad and it had to bring down the entire 2GB map set over the network. If,
> instead, it had a protocol for hailing for the content it might've found it
> on an adjacent computer. This is what torrents aim for and that approach
> should be treated as alternative to the DNS with equal standing.
>
> Though this isn't the only problem with the DNS it does go to the heart of
> how we think about what the Internet is and what it can be. This doesn't
> mean that the Internet is or should be content delivery network as some
> people think. It's just that by moving the intent mapping outside the
> network we put content delivery on equal footing with other ways of and
> reasons for exchanging bits.
>
> -----Original Message-----
> From: [email protected]
> [mailto:nnsquad-bounces+nnsquad <nnsquad-bounces%2Bnnsquad>=
> [email protected]] On Behalf Of
> Lauren Weinstein
> Sent: Saturday, May 29, 2010 11:09
> To: [email protected]
> Subject: [ NNSquad ] DNS vs. CDN
>
>
> DNS vs. CDN
>
> http://bit.ly/cbU2T3  (Sajal Kayan)
>
> Unfortunately, not only poor performance but other (often undesirable, from
> the standpoint of many users) manipulations that some ISPs perform on their
> DNS services are part of what's driven the move to services like OpenDNS
> and
> Google DNS in the first place.
>
> But there are clearly other mechanisms that could be employed to better
> establish users' locations for CDN (Content Distribution
> Networks) homing purposes, and the example of DNS vs. CDN is simply another
> indication of DNS' rapidly increasing obsolescence vis-a-vis the Internet
> of
> the 21st century.
>
> --Lauren--
> NNSquad Moderator
>
>

Reply via email to