In this thesis the intersection of several large nondeterministic finite automata (NFA's) as well as minimization of a large deterministic finite automaton (DFA) in map-reduce are studied. We have derived a lower bound on replication rate for computing NFA intersections and provided three concrete algorithms for the problem. Our investigation of the replication rate for each of all three algorithms shows where each algorithm could be applied through detailed experiments on large datasets of finite automata. Denoting n the number of states in DFA A, we propose an algorithm to minimize A in n map-reduce rounds in the worst-case. Our experiments, however, indicate that the number of rounds, in practice, is much smaller than n for all DFA's we examined. In other words, this algorithm converges in d iterations by computing the equivalence classes of each state, where d is the diameter of the input DFA.