Having seen that there are indeed attempts to deploy CORBA in real-time systems, we are encouraged to investigate possible applications in data acquisition systems for high energy physics. However, requirements for DAQ systems differ from real-time systems. For real-time systems the most important issue is not speed, but that data arrive in-time[KK90]. The behaviour must be predictable in the presence of competing tasks. In the case of a distributed system, the middleware must allow scaling to larger sets of processing elements. Other important issues are deterministic memory footprint and efficient processor usage in order to fulfil the first requirement. As an example, we may imagine a control system for a jet aircraft: Sensors deliver data about the aircraft to a real-time system that processes the information and displays results to the pilot. It has to issue warnings in critical situations and react autonomously[DBB$^$88] if the required reaction time is smaller than that of the pilot. Still, in such a hard real-time system the overall performance requirements are astonishingly low: event rates are in the range from 10 Hz to a 100 Hz. The size of the system is limited, mostly consisting of a main computer and a backup system. The amounts of data are in the range of some hundreds of sensors[HLS97]. Although a challenging area for providing deterministically behaving middleware it differs from the requirements of a scientific data acquisition system. A data acquisition system for recording physics processes has to scale to some thousands of physical entities. Object invocation rates will be in the range of 100 kHz. Memory footprint to be not only predictable, but be small as well. Timeliness and jitter is not so much an issue as good usage of the underlying network capabilities. Still, the real-time efforts can serve as a first attempt to apply object technology in high energy physics DAQ systems.