Computation of simulation data streams deriving from stored simulation results In general simulators are computer programs simulating the behavior of different artifacts based on parameters characterizing the simulated artifact. Traditional simulators are programs running some mathematical models to produce results stored in regular files. The results can also be stored in regular databases after proper reformatting. In the context of data streams it can be noticed that simulation methods also can be applied on live streams, rather than being run in batch over stored files. Such a simulator produces a simulation stream by implementing a simulation model. As a simple example, one could simulate the allowed temperature fluctuations in a building over time by a mathematical model knowing properties of the building. Often data produced by the simulation program will be stored in a repository and converted to a stream by reading from the repository. For example, the simulated temperature fluctuations over the year may be stored in a disk file and replayed as a live stream. A stream validator can compare a derived stream computed by a simulation program with the actual raw data (e.g. temperature) readings. The stream validator produces a validation stream. A validation stream can be a derived stream where the degree of deviations from the expected simulated result is computed. A validation stream can also be a stream of notification events signaling the need for action by some agent. When significant deviations are detected by the stream validator it may indicate either faulty behavior in the monitored artifact or that the simulation model is invalid. The streamed validation program with some certainty verifies that the simulation model is correct if no deviations between observed and simulated behavior are detected In the user survey the following scenarios related to stored data streams have been identified:
