Hi Gwren and Jason, you know TOR, right ?
http://freehaven.net/tor/tor-design.pdf
It is good, to read it some times again.
Do you think a network layer for all 2. generation p2p will be out soon
and then we do not need an amalgated p2p with proxy-mixture?
So then we should push ants a little bit, so that it has more users and more
media.
IMHo the strategy is to have SHA1 support as written and, like in ARES agrowing
column in the gui, so that users see a graphically growing line from 0-100 % for
each chunk (as well as it is in the upload tab with a number shown).
Ok, we have already taked about short time, sha1 and no gui enhancements at the
moment.
Hope you are not to busy these days.
Last thing: I read Jasons MUTE-end-to-end-homepage comment. I think ants is much
more developed with a lot of features and end to end makes the try, to open up a
packet, much more difficult.
But technically there is the open question: If I want to open op a channel to
transfer the key, then this channel could as well be cracked and man in the
middle could use my key, to then get the packerts in the second route. Because
the routs may be different, certain nodes are excluded, but who knows, that a
node is not fishing for keys, and another node then trying to attack in the
middle with these information.
Here Jason is right or I have not understood, how the key exchange could be
safe. Others i asked in personal mails haven ´t agreed on you. It is jsut one
meta foor over the other, on top, you are transferring a key unsecure. Of course
it is alwys one affort more to find the next channel. So it is the safest way,
but theoreticall yit is not impossible. See WASTE, to exchange public keys, you
have to choose a different WAY than over WASTE itself, so users use email or
telephone or IM.
IM could not be used for Ants, because then you stick to the public key the IP.
But hey ? what would be, if you use a Key generating mechanism like in WASTE,
and then transfer it over Jabber to at least 3 Online friends. These
"Instiututions" then could work as trusted third parties in the routing way.
though you reveal those IP adresses then may be.
Another way would be, if a initializing Peer is a Buddy, which has your
WASTE-Like Key.
Then you could use him as a Proxy shield forever. So the buddy finds out, if the
data packet was spoofed and the key channel is secure, and then forwards it to
me.
So think on the trusted peers, if they have manually entered the WASTE-Key like
Key of me in ants, then they can verify everting for me. Right ? The key then
could be transferrred over Jabber and only the architecture is changed, that
Every body needs Buddy friend he trust to exchange keys and then to proxy
eachother.
If the buddy is not online, then other buddies of me are allowed by the (just
before ) online buddy to give them a copy of a WASTE-Like KEy signed to my ID.
So one buddy in my buddylist gets My Waste key and coudl work a a trusted
instance.
Then My Buddylist organizes, that my trusted buddy is allowed to give my ID with
the WASTE-Like-key to all Other being online buddys of my list.
So after this, all buddies of my Messengerlist know my Key and coudl work as a
trsuted isntance and proxy me. If the "Best Buddy" (which I manually gave the
key) is offline, then the others of my buddylist could work in this function.
Then I have a proxy shield of all my buddies, forwarding packets to me,
Shielding my IP and IP ! and making and aprooving secured channels, which find
out, if the packet is wrong, because the secured Channel was made manually with
a trusted BEST BUDDY.
Then we have a buddy layer network, which is proxing my ID (not IP!, but as
well, so it is meta-meta-proxied!) and the last hop is secure.
A little bit complcated, but it works simply spoken like a WAST REMOTE CHANNEL
TO A BUDDY.
So each action I do in ants now,
is in the new model done remotely from me within a BEST Buddy, which then uses
ants.
e.q. Ants at home and a remote WASTE network at the office, which allows me to
use ants at home, Then a secure channel is made from work to home, which always
allows to identify wrong packets, because the WASTE-Remote-Channel key is
transfered over a trusted chat (you kow the jabber nic of a friend over telefone
or over other messengers).
So... if now home is me and WORK is a buddy, then this remote-ants model within
a buddylist could work as a totally trusted and secure network without any
problems.
Just use one buddy as a shield-proxy.
Is this design somehow clear?
So not using ants to transfer packets, and mute to transfer the key in a hybrid
model is one idea. but to use the buddies as a trusted shield would evenw ork
better,
You only need one buddy, then your WASTE-like-key could be broadcasted on the
group of your buddies in the messenger list like in waste automatically.
So again:
A is the searcher
B is the buddy
A and B use a WASTE secure channel
then B can work as a proxy for A and can downlaod everything,
A and B then will know, if any packet has been modified.
As well the ID and IP of A is secure, if a is just a Freind of Z and Z is the
real originator.
Ants1 network - > A -> B -> Ants2 network.
A and B have a trusted channel by WASTE key approvation.
then a man in the middle attack has to be in two clouds of the ants network, and
this would be impossible.
As well the WASTE-REMOTE Secure channel to BEST Buddy proxying my IP would work!
:-)