Скачать с ютуб Connected Components in MapReduce and Beyond; Sergei Vassilvitskii в хорошем качестве

Connected Components in MapReduce and Beyond; Sergei Vassilvitskii

Computing connected components of a graph lies at the core of many data mining algorithms, and is a fundamen-tal subroutine in graph clustering. This problem is well studied, yet many of the algorithms with good theoretical guarantees perform poorly in practice, especially when faced with graphs with billions of edges. We design improved al-gorithms based on traditional MapReduce architecture for large scale data analysis. We also explore the effect of aug-menting MapReduce with a distributed hash table (DHT) service. These are the fastest algorithms that easily scale to graphs with hundreds of billions of edges.