Research challenges of autonomic computing
Proceedings of the 27th international conference on Software engineering
A survey on wireless multimedia sensor networks
Computer Networks: The International Journal of Computer and Telecommunications Networking
Computer Networks, Fifth Edition: A Systems Approach
Computer Networks, Fifth Edition: A Systems Approach
Towards Autonomic Network Management: an Analysis of Current and Future Research Directions
IEEE Communications Surveys & Tutorials
Hi-index | 0.00 |
Ensuring the aspired outcome quality of network-based applications has ever implied an appropriate prognosis of the performance behavior the underlying communication structures will exhibit prior to potential optimization steps. However, the reliable forecast of correlative metrics is still a challenging task, especially in terms of wireless network systems that reject centralized or manual administration. In the context of our prospective self-management concept for autonomous analysis and control of volatile performance characteristics in Wireless Sensor Networks (WSNs), we relate a common network property known as packet transfer delay to various adjustable parameters peculiar to such radio network technologies. Our hands-on experiments reveal the reproducible influence of packet size, backoff period, and number of active neighbor nodes on the medium access procedure and involved performance indicators. By means of a closed formulation of permutable weighted drivers, we investigate the average-case predictability of inter-node end-to-end delays for arbitrary configurations of given network parameters. We validate our prediction method against basic multi-hop networking scenarios while verifying its practicability on a typical resource-scarce WSN platform. Leveraging measurement-driven inspection and conditional modeling of network attributes based on regression analysis, yield field test results substantiate the high precision of our approach to the estimation of performance-related WSN properties as the basis for application-aware performance optimization subject to projected complements.