Committee ChairPeterson, Larry L.
MetadataShow full item record
PublisherThe University of Arizona.
RightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
AbstractThe advent of high-speed networks may soon increase the network bandwidth available to workstation class computers by two orders of magnitude. Combined with the dramatic increase in microprocessor speed, these technological advances make possible new kinds of applications, such as multimedia and parallel computing on networks of workstations. At the same time, the operating system, in its role as mediator and multiplexor of computing resources, is threatening to become a bottleneck. The underlying cause is that main memory performance has not kept up with the growth of CPU and I/O speed, thus opening a bandwidth gap between CPU and main memory, while closing the old gap between main memory and I/O. Current operating systems fail to properly take into account the performance characteristics of the memory subsystem. The trend towards server-based operating systems exacerbates this problem, since a modular OS structure tends to increase pressure on the memory system. This dissertation is concerned with the I/O bottleneck in operating systems, with particular focus on high-speed networking. We start by identifying the causes of this bottleneck, which are rooted in a mismatch of operating system behavior with the performance characteristics of modern computer hardware. Then, traditional approaches to supporting I/O in operating systems are re-evaluated in light of current hardware performance tradeoffs. This re-evaluation gives rise to a set of novel techniques that eliminate the I/O bottleneck.
Degree ProgramComputer Science