Project Proposed: This project, acquiring server and networking equipment for activities in the area of secure and dependable computing, supports research in enterprise distributed systems, wireless mobile networks, and sensor networks. Projects include - Byzantine fault tolerance for long-running, nondeterministic systems, - Performability in wireless mobile networks, and - Hybrid emulation of wireless sensor networks. Novel methods are sought for providing Byzantine fault tolerance to long-running, nondeterministic distributed systems. Performance in highly stressed mobile wireless network is investigated in terms of high node mobility and strong interference, as is understanding of nodes? boundary behavior under extreme conditions seeking novel methods to survive the stress via cooperation. Lastly, a novel framework for hybrid emulation of wireless sensor networks is proposed to enable affordable experimentation at scale. The instrumentation consists of a cluster of servers connected by a managed high-speed switch forming a dynamically-configurable distributed computing testbed, servers to join the PlanetLab Consortium, and networking equipment to connect the testbed with an existing wireless network testbed. Broader Impacts: This project should have broad impact on security and dependability education. The tools and methodologies produced will be used by collaborators both in academia and industry (e.g., factory automation and highway work-zone safety). The infrastructure and outcomes will be integrated with the education curriculum; courses on fault tolerant computing and performance analysis will be developed. Moreover, the team collaborates with high schools providing research experience for teachers.
The research projects supported by the funded infrastructure are investigating novel algorithms and engineering approaches that will make a variety of computing systems more secure and dependable, including enterprise systems, wireless mobile systems, and sensor systems. More specifically, the principal investigators have targeted the following three research areas: (1) novel methods for building dependable distributed systems, (2) performability in highly stressed mobile wireless networks, and (3) a novel framework for hybrid emulation of wireless sensor networks. All anticipated goals and objectives for this project have been accomplished. The research outcomes in the first research area include: The design and implementation of a Byzantine fault tolerance framework that is capable of handling various replica nondeterministic behaviors and proactively recover itself to support long-running mission critical applications. The design and implementation of a Byzantine fault tolerance framework that enables concurrent execution of client’s requests for applications that use software-transactional memory. The design and implementation of a Byzantine fault tolerance framework for Web services business transactions and activities that use complicated system models. The framework significantly reduces the runtime overhead by exploiting the application semantics. The design and implementation of a low latency fault tolerance system for generic distributed applications. The objective of the work is to minimize the end-to-end latency for fault tolerant soft-real time systems by making the primary responding to a client’s request immediately. The research outcomes in the second research area include: The design and implementation of behavior-based mobility prediction scheme to eliminate the scanning overhead incurred in wireless networks. This is achieved by considering not only location information but also multifaceted user behavior based on other factors such as group, time-of-day, and duration. The design and evaluation of realistic mobility model called Clustered Mobility Model (CMM), in which nodes do not move randomly but move for a reason: Nodes tend to move towards a certain waypoint where more nodes are already present. Some subareas are highly populated with nodes but others are sparse. The design and implementation of frame forwarding mechanism in mobile multihop networks, which is useful to relieve the stress from ever increasing traffic demand. The key idea is to forward a frame over multiple hops with a single channel contention. The research outcomes in the third research area include: The design and implementation of a novel, extensible sensor node platform that supports multiple microcontrollers. This support for multiple microcontrollers, in addition to the scheduler software that can properly manage them, allows for sophisticated experimentation. The design and evaluation of a simulation platform for self-stabilizing systems. Sensor networks are frequently deployed in unknown states, and expected to stabilize to desired states. The simulation platform that we have designed and developed allows application designers to see the impact of algorithms before they are deployed in the network.
