Matteo's Wasps' Nest

Everyone needs a project for the summer, and I am no exception.

Apart from working on my main research topic (don’t ask me about it, I’m just making this up ), I’ll try to use my free time to write a series of posts on algorithms for MapReduce / Hadoop (MR/H). This will not be your usual “basic” stuff like how to naively implement SSSP or PageRank on MR/H. What I would like to do is to start from scientific articles published in conferences/journal (here is a good bibliography) and describe the techniques they used to design (and when possible, analyze) algorithms on MR/H. I’ll also try to present the weaknesses and the strengths of each paper and the open problems that arise from them.

I’m especially interested to see whether it would be possible to rigorously re-describe those algorithms in terms of a common computational model for MR/H, namely the one presented in [KSV2010], with slight modifications if/when needed.

In other words, it will be a one-man-reading-group, but my reports will be available on this blog.

I’m going to publish a tentative schedule of the covered papers soon.