A large class of real-time database problems has very high transaction loads and moderate deadlines. Traditional approaches have not been designed to handle such problems. A model based on the use of encapsulated events and rule objects has been developed. The model describes an active, real-time, object-oriented, memory-resident database environment (REACT). A system based on the model has been designed and implemented. A concurrency control algorithm was developed that uses the extra information available from the object-oriented and active features of REACT to pre-process the database and speed up concurrency control. Analysis was done for both single and multiple processor systems. For multiprocessor analysis a simulator was developed to verify the performance of REACT on a multiprocessor system. Examples of all the features needed for an actual system are given along with examples of how REACT can be used to solve real-world control and monitoring problems. Algorithms have been developed to allow users to test that the properties termination, confluence, and observable determinism hold for a target REACT database.