I spoke with John Mark and was directed here. :)
Jay Vyas on the big data team at Red Hat forwarded your thread of emails about GSoC to me. I was an undergrad researcher for 6 years at UConn while Jay was getting his PhD, and now, I'm a PhD student in Computer Science & Engineering at Notre Dame. My research interests can be summed as using math, physics, distributed computing, machine learning, and algorithms to support scientific work across multiple fields -- I have a relatively broad background.
I'm more interested in systems projects, e.g., integrating systems, that will get me deep into the concepts rather than peripheral projects. I like distributed systems, file systems, big data, etc.
Looking through your thread and list of ideas from last year, I found / thought of a couple interesting ideas:
* Cinder back up to Gluster
* RESTful / JSON interface to Gluster like WebHDFS. This would make it easier for other projects to use GlusterFS regardless of language Another GSoC group is looking at adding HDFS support to Disco, a MapReduce system in Erlang / Python. WebHDFS makes this easy. The key part is that you also need data locality information for scheduling. John Mark mentioned language bindings for Python, etc. -- I could build on top of those.
I'm very much open to other projects as well.
My knowledge of C/C++ is so-so -- I'm more of a Python / Java person with an interest in functional languages (e.g., Erlang). This may limit the projects that I could work on.
I'm very interested, however we're on a short deadline and there is a lot to do. The deadline is Friday. The Fedora application (not including the proposal) is a little more involved than other projects, but it's doable. To meet the deadline, I would need help crafting a proposal, especially from someone who knows the ideas, what the community is interested in, and can comment on the feasibility of the technical side.
What are everyone's thoughts?
Thanks!
RJ