If DNS was invented today?

I was wondering, if DNS was made in 2013; how different would it be?

who cares

I do?

FrozenVoid would if he's still alive

It'd be distributed. Same with DHCP and ARP.

Every request would be JSON!

It would have web 2.0 social networking with holistic synergy

# i7z --help

i7z Tool Supports the following functions:
Append to a log file:  ./i7z --write a [OR] ./i7z -w a
Replacement instead of Append:  ./i7z --write l [OR] ./i7z -w l
Default log file name is cpu_freq_log.txt (single socket) or cpu_freq_log_dual_%d.txt (dual socket)
Specifying a different log file: ./i7z --logfile filename [OR] ./i7z -l filename
Specifying a particular socket to print: ./i7z --socket0 X
In order to print to a second socket use: ./i7z --socket1 X
To turn the ncurses GUI off use: ./i7z --nogui
Example: To print for two sockets and also change the log file ./i7z --socket0 0 --socket1 1 -logfile /tmp/logfilei7z -w l
Quitting i7z

# i7z --baka-baka
Segmentation fault (core dumped)
LINUX QUALITY!!

embedded buttcoins in every packet

>>6
in fact, the whole thing would be defined in terms of javascript, like a proper standard

You would need an API key to query it.

It's a damn good thing it wasn't invented in 2013.

>>null
nice dubs bro

(1) The protocol would be more TLV-oriented to allow easier extension.
(2) The RFC would use RFC 2119-style language.
(3) The master file syntax description would be an XML schema.
(4) It might support some sort of cookie authentication.
(5) There would be even more butthurt about the TLDs.
(6) Dan Kaminsky would be part of the working group and at the point of consensus every coauthor would be required to sign an agreement about a yearly bitching quota.
(7) It would support STARTTLS and (D)TLS in general.
(8) Due to the additional round-trips required for TLS, the RFC would recommend to disable encryption by default. ;_;

>>14
The master file syntax description would be an XML schema.
The IETF likes XML now? I would expect something more ASN.1-ish.

It might support some sort of cookie authentication.
Why?

There would be even more butthurt about the TLDs.
Is that even possible?

Due to the additional round-trips required for TLS, the RFC would recommend to disable encryption by default. ;_;
I don't think that should be an issue. Authentication is more important than confidentiality in this case.

awesome thanks

it would be written in javashit

>>15
The IETF likes XML now? I would expect something more ASN.1-ish.
True.

Why?
If you make cookies mandatory (with reasonably long validity periods the additional initial round-trip should be acceptable), that should prevent IP spoofing and thereby solve the whole issue of DNS amplification attacks or am I missing something?

Is that even possible?
Consider the following: lobbying on TV, demonstrations, murder, war, holocaust, nukes. There is always an additional level of butthurt. Anyway, I like the fact that many basic issues were decided on a long time ago when nobody cared (except for IP's address size of course).

I don't think that should be an issue. Authentication is more important than confidentiality in this case.
True.