Meeting Notes 2015.03.03

Source: https://docs.google.com/document/d/16RxM0qe4ZBKsQCU74O-620jFjFq-ky246xRQPdvEVy4/edit?usp=sharing

Progress done this week:

1. Setup collaborations (codebase, …)

2. Understand the differences between pipeline of 2-way joins vs. n-way joins

3. Understand the 1-bucket 2-way partitioning:

• division equation: |R| x n = |S| x m and n x m = #number of machines

• computation cost: area = a x b

• communication cost: a + b

Expected progress next week:

1. Install, run, and test Squall in local mode

2. Understand existing libraries in Squall: join, partition, query plan

3. Understand the hybercube partitioning in Section 5.1 of the paper X. Zhang, L. Chen, and M. Wang, “Efficient multi-way theta-join processing using mapreduce,” in VLDB, 2012.

4. Plan how to implement brute force partitioning for hybercube.