Propagation of ESCL Cardinality Constraints with Respect to CEP Queries

Semantic optimization of database queries, i.e. the use of metadata (constraints) for query optimization, is well investigated and has led to signi cant performance gains. The algorithm semantically rewriting database queries presented in [2] is applicable to CEP queries with some adaptions. One of these adaptions concerns the treatment of views and the constraints for them. To reduce the number of constraints which must be manually speci ed by the user, the algorithm semantically rewriting database queries presented in [2] leads views back to their respective base relations so that constraints have to be de ned for base relations only. However, if a CEP engine involves automatic garbage collection [1] of events and states, the relations saving these events and states may become incomplete in the meantime such that CEP views cannot be led back to them without lost of tuples in some cases. Therefore, an approach opposite to the one in [2] is introduced in this work: ESCL [5] cardinality constraints are automatically propagated from base events (states) to the derived events (states) with respect to the queries deriving them, i.e. only the ESCL cardinality constraints for base events and states must be manually speci ed by the user. The propagation of the other kinds of ESCL constraints is to be investigated. The approach presented in this report is independent from an event query language. To illustrate it, CEP queries are expressed in StreamLog [5], a rst order logic language extended with temporal aspects as required for CEP.