Buffered Streaming Graph Partitioning
share
Partitioning graphs into blocks of roughly equal size is a widely used tool when processing large graphs. Currently there is a gap in the space of available partitioning algorithms. On the one hand, there are streaming algorithms that have been adopted to partition massive graph data on small machines. In the streaming model, vertices arrive one at a time including their neighborhood and then have to be assigned directly to a block. These algorithms can partition huge graphs quickly with little memory, but they produce partitions with low quality. On the other hand, there are offline (shared-memory) multilevel algorithms that produce partitions with high quality but also need a machine with enough memory to partition a network. In this work, we make a first step to close this gap by presenting an algorithm that computes high-quality partitions of huge graphs using a single machine with little memory. First, we extend the streaming model to a more reasonable approach in practice: the buffered streaming model. In this model, a PE can store a batch of nodes (including their neighborhood) before making assignment decisions. When our algorithm receives a batch of nodes, we build a model graph that represents the nodes of the batch and the already present partition structure. This model enables us to apply multilevel algorithms and in turn compute high-quality solutions of huge graphs on cheap machines. To partition the model, we develop a multilevel algorithm that optimizes an objective function that has previously shown to be effective for the streaming setting. Surprisingly, this also removes the dependency on the number of blocks from the running time. Overall, our algorithm computes on average 55 than Fennel using a very small batch size. In addition, our algorithm is significantly faster than one of the main one-pass partitioning algorithms for larger amounts of blocks.
READ FULL TEXT