JXTA is a standard (and accompanying reference implementation in Java) for P2P nodes to connect, form groups and ‘advertise services’. I’m currently looking at JXTA for the core networking technology in my project. Several things have struck me about JXTA’s design as overly presumptuous and are deal breakers for me.

All peers belong to the “World Peer Group”. I presume this means that all peers can communicate with each other because they belong to this group. Furthermore all peers are encouraged to rout traffic for each other. As a peer I would not be inclined to rout traffic unless it directly benefited me or other peers in my group. The last thing I want to be doing is routing traffic for some file sharing service when I’m trying to commit to changes to the group. My traffic comes first, preferable exclusively. The only thing I would agree to route for other peers would be discovery messages. Using the network at large to help people find each other is a fair exchange in value for traffic.

I’m aiming for a secure network. It seams peers can join my group without authenticating with one of the group members first. I want to have a custom authentication protocol with public key encryption.

There dost seam to be clear answers to these questions on the JXTA website either. I guess its time to dig into the protocol spec to find out.

My housing situation is getting stranger by the day. Now the usurpers, who wanted to rent the place out from under me and were putting pressure on the apartment complex, have disbanded. The poor guy that’s in there now wants to share the place with me. Unfortunately for him I’ve already promised to make other arrangements. I guess this just means more options for me.

I read all about the current copyright situation today Linux Journal. It was a very good article explaining why we live in very strange times as far as copyright is concerned. This lead on to some interesting stuff on telecoms as simple pervayors of bits because the network now consists of inteligent nodes (computers) instead of dumb telephones. All worth reading.

I’m in discussions with my profs over my thesis topic. They think the idea is interesting and that there is defiantly enough ‘research’ to constitute a thesis. Right now we are looking at how original the idea is. Groove Networks makes something similar. Their solution is aimed at business users and provides very structured co-ordination tools. On my project I’m theorizing that too much structure will actually be a hindrance to small projects because people work very informally. They do have problems arranging meeting times and such but not so badly that they need a calendar application to do so. A Wiki should be more that sufficient to act as message board, whiteboard, notepad and calendar.

Come to think of it I might not call it a Wiki because this term seams to elicit nothing but confusion from uninitiated users. The user interface will be a big push in the project. It has to be easy to use for people to adopt it. I’m also toying with the idea of implementing the CVS protocol locally so that tools like eclipse can talk to my software without modification. Not much is set in stone at the moment, other than the P2P nature of the software. I will have to evaluate the benefits of each potential feature and my available time.

Feature set? To Be Determined…
Finished? When Its Done.

I was in the middle of writing a water cooling beginners guide with a bunch of links to suppliers and so on. Seams
that Overclockers.com has read my mind and published a very good beginners
guide with all the same advice and links that it took me 3 days to find on the net. You can read their
guide here. They also
have link to some of their own articles on reservoir construction and using Radio Shack relays to switch on your pump.
Two great ways to save some money if you build a water cooled rig.

I’m currently considering water cooling my next box. This should give me about 10 degrees cooler temps on the CPU and
also eliminate the tow major sources of noise in the computer: the CPU fan and the Graphics card. Water cooling is at a
stage where it shouldn’t be too hard to get it right. There is lots of information out there now. also the cost isn’t
that bad, between $100 and $200. You could save that on your next CPU purchase and do an overclock.