Due to the rapid development of massively parallel data processing systems such as MapReduce and Spark, there have been revived interests in designing algorithms in a massively parallel computational model. Computing multi-way joins, as one of the central algorithmic problems in databases, has received much attention recently. This paper surveys some of the recent algorithms, as well as lower bounds. We focus on multi-round algorithms, while referring readers to [27] for single-round algorithms.
展开▼