Newscast Computing
摘要:
Monitoring large computer networks often involves aggregation of various sorts of data that are distributed across network components. Finding extreme values, counting discrete observations or computing an average or a sum of some parameter values are typical examples of such "background" activities that provide input to monitoring systems. Another aspect of network management is fast and reliable information dissemination, like propagation of alarm signals. We present a novel approach to information aggregation and dissemination. It is based on a concept of a highly distributed, anonymous, democratic and non-deterministic form of collab- orative information processing: newscast computing. The main properties of this approach are scalability, robustness, adaptivity, and speed. The underlying protocol is very simple and can be implemented and run on huge networks of small computing devices, such as mobile phones, PDA's, sensors, etc. The usefulness of the newscast approach is illustrated by two algorithms for finding the maximum and the average of values that are distributed along the nodes of a network. The algorithms are formally proven to converge exponentially fast and numerous simulation experiments provide additional insights into their behavior. Finally, we demonstrate their applicability to several network monitoring tasks: finding the size of a network, counting nodes that left or joined the network, system load estimation, and fast propagation of alarms.
展开
被引量:
年份:
2003

























通过文献互助平台发起求助,成功后即可免费获取论文全文。
相似文献
参考文献
引证文献
辅助模式
引用
文献可以批量引用啦~
欢迎点我试用!