k-Connectivity in the Semi-Streaming Model

Abstract

We present the first semi-streaming algorithms to determine k-connectivity of an undirected graph with k being any constant. The semi-streaming model for graph algorithms was introduced by Muthukrishnan in 2003 and turns out to be useful when dealing with massive graphs streamed in from an external storage device. Our two semi-streaming algorithms each compute a sparse subgraph of an input graph G and can use this subgraph in a postprocessing step to decide k-connectivity of G. To this end the first algorithm reads the input stream only once and uses time O(k^2*n) to process each input edge. The second algorithm reads the input k+1 times and needs time O(k+alpha(n)) per input edge. Using its constructed subgraph the second algorithm can also generate all l-separators of the input graph for all l<k.