Skip to main content

An Interview with a P2P/JXTA Brazlillian

Posted by turbogeek on February 16, 2006 at 12:48 AM PST

Edward Ribeiro is a Student at the at University of Brasilia doing some interesting work on JXTA for his Masters dissertation. Daniel Brookshier sat down via email for a chat about what he is doing with JXTA.

Name: Edward Ribeiro

JXTA ID: edwardribeiro

University of Brasilia, Brasilia

Daniel Brookshier (Q): Tell us a bit about yourself.

Edward Ribeiro (A): I am a second year MSc. student at University of Brasilia (Brazil). I am currently working in a government institution in Brazil.

Q: How long have you been working with JXTA?

A: I have been working with JXTA since the beginning of 2005. Nevertheless, I have played with it during short periods of time since its creation.

Q: What are you using JXTA for and in particular why use Peer-to-Peer as a

solution to your problem?


A: My masters dissertation is centered on distributed systems applied to computational biology problems and it is heavily based on JXTA. We are trying to develop a P2P "cluster-like" system to process DNA sequences on BLAST. We think the P2P model can be scalable and flexible solution to distributed programs execution and, of course, we were highly inspired by the good example of Protein@HOME.

We have just finishing our first prototype using JXTA. It works well, but there are some open questions/bugs related to JXTA that need to be solved before deploying it. My advisor was very happy with the little demo I showed.

Q: What attracted you to choose the JXTA technology and platform?

A: It is the most mature P2P solution in the market. I could have written a customized P2PP2P solution, but this was not the point in the dissertation work. I needed a mature platform to implement further ideas on it.

Q: What open source JXTA projects are you working on?

A: I could say that I am particularly interested in the 'platform' project. I keep posting on a regular basis and solving doubts as much as possible. I hope to be deep involved in this project after defending my thesis. I am particularly interested in the jxta-c project, but I didn't have any free time to spend on it.

Q: Is your company using JXTA and if please describe it if you can?

A: There's a consortium of universities paying a close attention at my project. They want to use the distributed infrastructure that I am building. Therefore, JXTA can potentially be used to integrate all these universities in a comp. biology network.

Q: In either your company or in your open source projects, do you need help and if so what skills or experience are you looking for?

A: My project is a great JXTA testbed and I would be more the grateful if you could provide me some personal advice. It needs to be secure, encouraged.

Q: What JXTA platform(s) are you using (J2SE, J2ME, C)?

A: J2SE

Q: This is for us to do a better job, what do you think needs to be added to the JXTA.org web site to help you?

A: At first, I think that data transmission (pipes, jxtasockets, jxtabidipipes) should have the highest rank in a priority list, because jxtasockets and jxtabidipipes are not 100% reliable or efficient as plain sockets. I could certainly live with some degree of inefficiency, but unreliability makes me too upset. A related point in this area would be the addition of QoS (Quality of Service) to JXTA.

Finally, an updated version of JXTA inner workings. Something like this (http://wiki.java.net/bin/view/Jxta/JXTAImplementationUseCases#Publish_an...).

Q: One more question: What is your most burning question about JXTA or P2P development? (I'll reply if I have the answer)

A: Could you suggest a good presence design for JXTA? I have implemented my own, but it is not quite reliable as I would like to.:-)

Q: I would be happy to! Try https:commons.jxta.org where we have some good code for a simple presence system.

Thank you for your time!

You are welcome.:-)