Running device drivers as unprivileged user-level code, encapsulated into their own process, has often been proposed as a technique for increasing system robustness. However, in the past, systems based on user-level d...Running device drivers as unprivileged user-level code, encapsulated into their own process, has often been proposed as a technique for increasing system robustness. However, in the past, systems based on user-level drivers have generally exhibited poor I/O performance. Consequently, user-level device drivers have never caught on to any significant degree. In this paper we demonstrate that it is possible to build systems which employ user-level device drivers, without significant performance degradation, even for high-bandwidth devices such as Gigabit Ethernet.展开更多
This contribution discusses the concept of Reliability-Adaptive Systems （RAS） to multi-system operation. A fleet of independently operating systems and a single maintenance unit axe considered. It is the objective i...This contribution discusses the concept of Reliability-Adaptive Systems （RAS） to multi-system operation. A fleet of independently operating systems and a single maintenance unit axe considered. It is the objective in this paper to increase overall performance or workload respectively by avoiding delay due to busy maintenance units. This is achieved by concerted and coordinated derating of individual system performance, which increases reliability. Quantification is carried out by way of a convolution-based approach. The approach is tailored to fleets of ships, aeroplanes, spacecraft, and vehicles （trains, trams, buses, cars, trucks, etc.） - Finally, the effectiveness of derating is validated using different criteria. The RAS concept makes sense if average system output loss due to lowered performance level （yielding longer time to failure） is smaller than average loss due to waiting for maintenance in a non-adaptive case.展开更多
文摘Running device drivers as unprivileged user-level code, encapsulated into their own process, has often been proposed as a technique for increasing system robustness. However, in the past, systems based on user-level drivers have generally exhibited poor I/O performance. Consequently, user-level device drivers have never caught on to any significant degree. In this paper we demonstrate that it is possible to build systems which employ user-level device drivers, without significant performance degradation, even for high-bandwidth devices such as Gigabit Ethernet.
文摘This contribution discusses the concept of Reliability-Adaptive Systems （RAS） to multi-system operation. A fleet of independently operating systems and a single maintenance unit axe considered. It is the objective in this paper to increase overall performance or workload respectively by avoiding delay due to busy maintenance units. This is achieved by concerted and coordinated derating of individual system performance, which increases reliability. Quantification is carried out by way of a convolution-based approach. The approach is tailored to fleets of ships, aeroplanes, spacecraft, and vehicles （trains, trams, buses, cars, trucks, etc.） - Finally, the effectiveness of derating is validated using different criteria. The RAS concept makes sense if average system output loss due to lowered performance level （yielding longer time to failure） is smaller than average loss due to waiting for maintenance in a non-adaptive case.
