[e2e] Time for a new Internet Protocol

[Jeroen Massar]

David P. Reed wrote:
> A motivation for TCP and then IP, TCP/IP, UDP/IP, RTP/IP, etc. was that
> network vendors had too much control over what could happen inside their
> networks.
[..]

All nicely put, but can you define exactly what you currently can't do
on your network, and what you want to be able to do with it?

ISP's I use don't filter my traffic, except on the origin address which
is quite logical.

Some locations do have some restrictions, eg tcd.ie filters almost
everything, but hey, that is their network and their policy, so not
something I can complain about, the network is intended to be used for
different purposes as set out by their policy, not for all kinds of
fancy things that I want to be doing. Although when nicely asked I am
pretty sure that some holes can be created under the disguise of it
being research of course. Also circumventing it is of course quite easy
if I really wanted to, as long as you can send a packet out and get one
back you are done.

> Therefore, it is time to do what is possible: construct a new overlay
> network [...]

Like http://www.isi.edu/xbone/ ?

And of course don't forget about IPv6, which for a lot of end users
today is still a tunneled mesh network which is mostly unfiltered.
At the moment RH0 is getting filtered in a lot of places merely because
of the DoS property and the argument 'be good to your neighbors'.

> One model is TOR, another is Joost.   Both of these services overlay
> rich functions on top of the Internet, while integrating servers and
> clients into a full Internet on top of today's Internets.

Joost is very nice, especially when the p2p was not enabled yet. Now
with the p2p network enabled it sucks up your connectivity. When you
have only a small amount of upstream bandwidth (remember that home users
have silly 1mbit/128kbit asymmetric links and not cool symmetric 1GE
links to their campus offices), then you will have video for a couple of
moments, but the second that you are going to provide back to the swarm
of viewers, which you do as it is P2P based (Joost works akin to the so
called VidTorrent project by some MIT folks), your upstream will be
filled up, and your downstream gets stuck, as those Joost ack packet's
won't come trough.

Working congestion control would be very helpful in these cases, the app
could of course limit the amount of data it is sending up too, but that
takes the whole P2P idea out and folks would then simply limit the
amount of traffic going up and the whole P2P net would fail...

Greets,
 Jeroen

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 311 bytes
Desc: OpenPGP digital signature
Url : http://mailman.postel.org/pipermail/end2end-interest/attachments/20070515/1fb6559b/signature.bin