FastGraph Pattern Matching

Duetorapid growth oftheInternet technology and newscientific/technological advances, thenumberofapplications thatmodeldataasgraphsincreases, because graphshave highexpressive powertomodelcomplicated structures. The dominance ofgraphsinreal-world applications asksfornew graphdatamanagement sothatuserscanaccess graphdata effectively andefficiently. Inthis paper, westudy agraphpattern matching problemoveralarge datagraph. Theproblem isto findallpatterns inalarge datagraphthatmatchauser-given graphpattern. We propose anewtwo-step R-join (reachability join) algorithm withfilter stepandfetch stepbasedonacluster- basedjoin-index withgraphcodes. We consider thefilter step asanR-semijoin, andpropose anewoptimization approach by interleaving R-joins withR-semijoins. We conducted extensive performance studies, andconfirm theefficiency ofourproposed newapproaches. I.INTRODUCTION