Adapting to Varying Demands in Resource Constrained Real-Time Devices

In-home entertainment systems are becoming popular because they allow a variety of consumer electronic (CE) devices to be connected, using a wireless network, to form a system capable of handling multimedia content. Using such a system provide the end-users the possibility of transparently streaming multimedia content between devices of varying capabilities. This is possible because the system adapts the multimedia content to match the capabilities of the receiving device. What looks simple to the end-user, is actually a very complex system that manages all existing multimedia streams and resources. It must manage all the varying resource demands, on all the constrained devices, in such a way that the resulting quality (video or audio playback) is acceptable to all the end-users of the system. In this thesis we investigate two different, but still related, issues within the in-home entertainment network. First, we look at how we can adapt to the the capabilities of the nodes, which contains processors of varying capabilities and also operating systems which also provide different capabilities. Secondly, we have to adapt to the varying capabilities of the wireless network when it is used for video streaming in the presence of other network traffic. For nodes, we present two scheduling methods that are extensions to the off-line scheduling paradigm. The first method aims at improving the handling of soft aperiodic tasks in an off-line scheduled system, which are normally handled in the background resulting in long response times. The method creates