Thesis
Rate-based failure detection for critical-infrastructure sensor networks
Washington State University
Master of Science (MS), Washington State University
2014
Handle:
https://hdl.handle.net/2376/102058
Abstract
Reliable fault tolerant communication is a vital component in advanced power grids (smart grids) that need to properly and correctly function amidst malfunctioning equipment, natural disasters, and cyber attacks. GridStat is a managed publish subscribe middleware communication infrastructure specifically designed to meet the high availability, low latency reliable communication requirements of these critical infrastructures. In order to maintain its services in face of malfunctions and attacks on the IT infrastructure it is built on, Gridstat's capabilities needs to be augmented using an adaptation service that detects and reacts to these events. Traditional work in failure detection has shown by that using heartbeat messages between components, it is possible to effectively identify components that crash and cease to operate. However, this technique will fail to identify Quality of Service (QoS) failures. By exploiting GridStat's rate-based semantics and complete knowledge of sensor flows at all network nodes, a rate-based failure detector was developed. These are more robust and capable of identifying these QoS failures. Additionally, these strategies have been compared with heartbeat failure detectors by using the DETER testbed, and have show that rate-based failure detectors are a potential solution for identifying and locating QoS failures in a managed publish subscribe system.
Metrics
1 File views/ downloads
15 Record Views
Details
- Title
- Rate-based failure detection for critical-infrastructure sensor networks
- Creators
- Brett Emery Trabun Johnson
- Contributors
- David Bakken (Degree Supervisor)
- Awarding Institution
- Washington State University
- Academic Unit
- Electrical Engineering and Computer Science, School of
- Theses and Dissertations
- Master of Science (MS), Washington State University
- Publisher
- Washington State University; [Pullman, Washington] :
- Identifiers
- 99900525021001842
- Language
- English
- Resource Type
- Thesis