If we are running a graph algorithm on multiple machines, we would ideally divide the graph up into clusters with a minimum number of edges between the clusters.
Figuring out the minimum cut of a graph is NP-complete, so finding a good division between two machines is challenging, nevermind n-machines, or the more difficult generalization of requiring balanced clusters.
If we are running a graph algorithm on multiple machines, we would ideally divide the graph up into clusters with a minimum number of edges between the clusters.
Figuring out the minimum cut of a graph is NP-complete, so finding a good division between two machines is challenging, nevermind n-machines, or the more difficult generalization of requiring balanced clusters.