Figure 4 Priority inheritance prevents priority reversal Ensuring availability For many systems, resource availability is critical. If a subsystem experiences a lack of CPU cycles, this can have very serious consequences. For example, an interrupted connection to a cardiac monitor could cause the central monitoring system to mistakenly believe that an alarm condition has occurred and dispatch first responders, or the system could fail to send an alarm message when a patient is truly in critical condition. Based on past experience, the way to solve this problem is to update the hardware or redesign the software. While sending newly designed software to a networked medical device is a viable option, this not only makes it costly, but may also invalidate the device's existing product certification. PartitioningPartitioning technology puts the problem of resource scarcity to rest by enforcing CPU budgeting and preventing processes from monopolizing CPU cycles. Two types of partitioning are available: fixed and adaptive. With fixed partitioning, the system designer specifies the allocated CPU time for each task. No task within a partition can consume more than the proportion of CPU time that has been allocated at the time of partitioning, in order to keep processes in other partitions readily available, and thus ensure that all critical processes are readily available. Unfortunately, with fixed partitioning, a process within another partition cannot be used even if cycles allocated elsewhere are unused because it cannot use more CPU cycles than those pre-allocated for the partition in which it is located. While fixed partitioning prevents the system from experiencing a scarcity of resources, it wastes CPU cycles and weakens the system's ability to cope with peak demand. Similar to fixed partitioning, adaptive partitioning also prevents resource scarcity. However, adaptive partitioning uses a dynamic scheduling algorithm to reallocate idle CPU cycles in one partition to another partition that requires more processing time. Adaptive partitioning performs resource allocation when processes within multiple partitions compete for CPU cycles. In this way, designers can fully rely on system resource guarantees and do not have to search for a cure for the reduced processing power caused by fixed partitions. Monitoring, stopping, or restarting processes Preventing chain failures of processes throughout the system and self-healing features are critical to a trustworthy operating system. Devices that need to ensure system availability can use hardware-oriented high-availability solutions and software monitoring programs. A monitoring program monitors the system and performs a multistage recovery or smooth shutdown as required. Depending on how it is executed, when a failure occurs it should either: terminate and then restart the failed process without rebooting the system, or terminate the failed process and related processes, initialize the hardware to a "safe" state, and then restart the terminated process in a coordinated manner, or, if the failure is severe or safety critical, it can shut down or shut down the entire system in a controlled manner. Or, in the case of a serious or safety-threatening failure, to shut down or reset the entire system in a controlled manner and alert the system operator. In all cases, the monitoring program must be able to monitor itself and respond flexibly to internal faults. If the monitoring program stops unexpectedly, it must replicate its state through a mirror process. Finally, software monitoring programs can monitor system events that are invisible to traditional hardware monitoring programs. For example, a hardware monitor can ensure that a driver is serving the hardware, but it is difficult to detect whether other programs are talking to the driver correctly. A software monitoring program can compensate for this by taking action before there is a problem with the driver itself. Conclusion The trend of aging populations and tightening health budgets in developed countries are contributing to a shift in healthcare delivery from the conventional hospital model to telemedicine and home-based self-help models. These changes and the latest technologies are driving an expanding market for telemedicine devices at a pace that would have been unimaginable a decade ago. By carefully analyzing the performance characteristics of their chosen operating system, device manufacturers are able to reduce costs and increase the chances that their products will succeed in the marketplace by successfully passing industry certifications.