Performance Improvement of State Space Exploration by Regular & Diffrential Hashing Functions

1 . I n t r o d u c t i o n Hashing method is a well suited method to achieve state-space exploration for verification of distributed systems. Hashing method is used as searching method to accelerate the retrieval process of a particular state among large state space under exploration. This method enables searching, insertion and suppression operations to be done on average at a constant cost in number of comparisons. But the usual computation process of the hashing value has the first following drawback : its complexity is at least proportional to the key length, and unfortunately, the state descriptor, from which the keys are based, is in general very large (several hundreds of bytes [Doldi 92, Holzmann 91, Wolper 93]), if accurate modelling is considered. Moreover, some of the recent works on state-space exploration made an intensive use of hashing functions (bitstate method [Holzmann 88], multihash method [Wolper 93]) : two hashing function calls in Holzmann's Spin validation environment for each newly created state, Wolper recommends 20 calls for each newly created state to achieve large coverage of the state-space. These two methods reduce the amount of space needed to store the explored state-space but, as Wolper writes, due to the intensive function calls, they have the second following drawbacks : "computing 20 hash functions is quite expensive and will substantially slow down the search". In previous work on improvement of state-space exploration we have introduced differential hashing functions [Cousin 93]. These hashing functions use differential