[mnet-devel] Ent status

[Joris Bontje]

[forwarded to mailinglist as requested by zooko]

<mids> zooko: can you give me a status on ent?
I have been reading cvs commits and p2p-hackers logs
where is it at the moment? are there special problems with it? learned lessons?

<zooko> Hm.
<zooko> ent current status: dormant
current problems: a combination of theoretical and practical

theoretical: I don't know how to make a DHT network be both self-healing and
efficient in the way that I wanted.

practical: the ent unit tests showed a failure to integrate a new node into an
existing ent network.  I'm pretty sure that this was just a race condition, but
I never debugged it because:

practical 2: the ent unit tests were way too ugly and complicated, since they
presented an entire simulated network in Python code.

organizational: I wanted to switch to Twisted before refactoring the unit tests
in the hopes that Twisted would simplify the simulated-network code, and I
wanted to do other stuff before switching to Twisted, and then I got distracted
with v0.6.2 debugging ....  :-)

lessons learned: maybe I should work on "advance the state of the art
of DHT design" in a separate step from "integrate a DHT with Mnet"...  I'd like
to hear more about self-healing Kademlia.  I know Bryan Ford just published a
white paper about his "Unmanaged IP" research, which includes the issue of
self-healing DHTs.
-- 
PGP Key: http://pgp.mit.edu:11371/pks/lookup?op=get&search=0xF19326A9
Key fingerprint = 730D 9B3A F406 F28A 957D 6397 31E8 6D4C F193 26A9


-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
_______________________________________________
mnet-devel mailing list
mnet-devel at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mnet-devel