FOSS and IRC and Blogging, OH MY!
So, today was my first real experience with chatting in IRC. As a class, my Object Oriented Programming/Design classmates and I joined an IRC channel to edit some Wikis and get some information flowing about FOSS projects that we may want to work on. For the most part it was a bunch of silliness but there were some really cool ideas thrown out. One of my classmates suggested a project called irrlicht that seems pretty neat. Another couple of ideas that were tossed around were GIMP and Inkscape. The one that people seemed most interested in towards the end of class was a suggestion by someone outside of the class that happened to be in the channel; a project working with Eucalyptus.
Cloud computing seems to be the way of the future and it would be AWESOME to get in on a project involving the cloud before I even graduate. We’ll have to see what the class decides on in the next week. Whatever we decide to do this class promises to be an amazing experience for all of us involved.
Nimbus (http://scienceclouds.org/) is another tool like Eucalyptus, except it is 100% free and open source where Eucalyptus has free (somewhat intentionally broken) and a paid version. They are cool, the only problem is to develop it you need some hardware to test it, like a cloud environment.
I’m not sure what kind of scope you have for this class, but if you want a smaller project you can check out Brubeck, a Python web framework my friend is working on (http://brubeck.io/) that uses Mongrel2 (http://mongrel2.org/) from Zed Shaw as a frontend. It has been used as a project to help people in Hacker School (http://www.hackerschool.com/). If you’d like to contribute, just email the mailing list and James will be happy to offer some suggestions.
Thanks so much for the tip! I’m not entirely sure what the scope of the project is either but I will make sure to bring these suggestions to the professor. Thanks again!
Malcolm,
Thanks for the info. I have some relatively high-level contacts in the Eucalyptus project, so I’ve asked them to respond to your comments. I’m especially interested in their response to your comment about not being able to develop without the appropriate hardware.
There are some definite advantages to working in this project, as long as we can address the issue you raised.
I hope things are going well for you. Drop us a line and let us know what you are up to.
I think it’s great this class will involve some blogging, hopefully I can offer some insight form my work experience.
Most of my Eucalyptus experience comes second-hand. At my previous job we setup a free academic cloud infrastructure called DIAG (http://diagcomputing.org/). We decided to go with the Nimbus stack. Our IT department looked at Eucalyptus and they weren’t too fond of it. They found the free version to be buggy with the company stating the bugs were fixed in the commercial version. We also talked with the IT department for the, now defunct, Magellan cloud at Argonne National Labs, which was a massive setup (1,000+ physical machines, I believe) and they were very disappointed with Eucalyptus. They found it unstable at the scales they wanted to run. That being said, even the Nimbus stack we used was full of bugs and a challenging piece of software to install and maintain. The upside to Nimbus for us was that it is completely free and the developers are active on the mailing list so issues were addressed quickly. Another option is Openstack, although I don’t have any knowledge about that one.
My point about the hardware is just that one cannot test a piece of software designed for a distributed environment without having a distributed environment to test on. Changes may work fine when one host is up but have completely undefined results when 10 hosts are up. Perhaps Eucalyptus can be adequately developed on VMs in which case only a few physical machines would be necessary for testing.
Where is the most appropriate place to catch up? I don’t want to hijack Matt’s blog. The short version is I’m moving to Sweden in a few weeks for a new job developing Erlang. I’d be happy to post something if the CS department has a “Where Are They Now” section, although I can’t promise it to be a spellbinding story.