I was curious about the Redis sets used for social graph storage and using intersection of sets to find nodes in common. Would anyone have an idea about the complexity of this as both the number of nodes in the graph and the number of edges each node has?