So, as most people know, Fred has most of the code about making
connections between nodes pretty abstracted out. Right now, the only
connection classes are for direct TCP connections between nodes, but
I've been wondering about other types of connections and
addresses. Some examples that might be interesting:
- AIM addresses ("aim/MrBadPigdog"). Messages between nodes go
over the AIM network, formatted or encapsulated as AIM
messages. (Recent moves by AOL to close the AIM servers may
make this more of a hassle than it would be already.)
- ICQ addresses ("icq/14171206"). Similar to AIM.
- SMTP/POP ("[EMAIL PROTECTED]"). This would
probably be unbearably slow, but it could be interesting. A
node would send messages as mail, and "listen" on a
particular POP box.
- Jabber ("[EMAIL PROTECTED]").
Another instant messaging system, with the advantage of
being highly decentralized. I think Jabber even has an
extension mechanism so you could add a special FNP message
type to the Jabber protocol.
- HTTP ("http/http://www.pigdog.org/cgi_bin/freenetnode";).
Nodes can present a Web interface, and messages can be sent
by POSTing the FNP message from one node to another. The
response can come back as the contents of the "query
result."
- SMB ("smb///MYCOMPUTER/Freenet"). This is Windows protocol
used to make printer and file share connections. Could be
interesting on a LAN, say.
- JXTA ("jxta/IDontKnowJXTASyntaxYet"). Despite Mr. Langley's
high-profile dismissal, it seems to me that JXTA would be as
good as any of these other protocols for sending and
receiving FNP messages.
None of these have particularly good anonymity protection built in,
but then again neither does TCP. Anonymity of node addresses is not a
goal, either. For the particularly paranoid, there's even some
anonymity protection in some of these protocols. For example, the
SMTP/POP scheme -could- use anonymizing remailers, or even chains of
anonymizing remailers. HTTP stuff *maybe* could go through SafeWeb or
ZeroKnowledge (although I don't know enough about those services to
say).
One cool part about all this is a given FNP message could travel over
different protocols in its lifetime, making it that much more
difficult to trace. Another is that different connection protocols
provide a crude level of steganography, AND might help people who are
behind draconian firewalls. Lastly, nodes could be implemented as
small "plugins" for software people already have and use a lot, like
AIM. So the acceptance level goes up, in a way, possibly.
The bad parts, of course, are that first of all none of these
tunnelled protocols is going to be nearly as fast as a direct TCP
connection. Some might be hard to fit into the current framework of
Fred (like HTTP). And, of course, adding N more addressing options is
just going to make Fred N times more confusing for novice users.
Anyways, I just wanted to throw this out there for comment.
~Mr. Bad
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Mr. Bad <[EMAIL PROTECTED]> | Pigdog Journal | http://pigdog.org/
freenet:MSK@SSK@u1AntQcZ81Y4c2tJKd1M87cZvPoQAge/pigdog+journal//
"Statements like this give the impression that this article was
written by a madman in a drug induced rage" -- Ben Franklin
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
_______________________________________________
freenet-tech mailing list
[EMAIL PROTECTED]
http://lists.freenetproject.org/mailman/listinfo/tech
