As the size of your input graphs grows, the problem size increases exponentially, which can lead to very long run times as well as memory issues when trying to solve the sub graph isomorphism problem...

As the size of your input graphs grows, the problem size increases exponentially, which can lead to very long run times as well as memory issues when trying to solve the sub graph isomorphism problem (which is NP-complete)? Therefore, it is possible that some input datasets will cause g Span and other frequent sub graph mining programs to run out of memory long before they arrive at an answer.