Not available / text

Not available / text

A sensor is a battery-operated small computer with an antenna and a sensing board that can sense magnetism, sound, heat, etc. Sensors in a network communicate and cooperate with other sensors to perform given tasks. A sensor network is exposed to various dynamic factors and faults, such as topology changes, energy saving features, unreliable communication, and hardware/software failures. Thus, protocols in this sensor network should be able to adapt to dynamic factors and recover from faults. In this dissertation, we focus on designing and analyzing a class of sensor network protocols, called self-stabilizing protocols. A self-stabilizing protocol is guaranteed to return to a state where it performs its intended function correctly, when some dynamic factors or faults corrupt the state of the protocol arbitrarily. Therefore, in order to make a sensor network resilient to dynamic factors and faults, each protocol in the sensor network should be self-stabilizing. We first develop a state-based model that can be used to formally specify sensor network protocols. This model accommodates several unique characteristics of sensor networks, such as unavoidable local broadcast, probabilistic message transmission, asymmetric communication, message collision, and timeout actions and randomization steps. Second, we present analysis methods for verifying and analyzing the correctness and self-stabilization properties of sensor network protocols specified in this model. Third, using the state-based model and analysis methods, we design three self-stabilizing sensor network protocols, prove their self-stabilization properties, and estimate their performance. These three self-stabilizing protocols are a sentry-sleeper protocol that elects a sentry from a group of sensors at the beginning of each time period, a logical grid routing protocol that builds a routing tree whose root is the base station, and a family of flood sequencing protocols that distinguish between fresh and redundant flood messages using sequence numbers. / text

M. Tech. Electrical Engineering. / A virtual Mobile Voice over IP (MVoIP) service can be implemented by a Mobile VoIP Operator (MVoIPO) in conjunction with a Mobile Network Operator (MNO). MVoIPOs do not operate their own mobile network infrastructure. Instead, they use the MNO's packet-based cellular network. However, the coexistence between the MVoIPO and the MNO raises two related problems: first, how to handle interconnection settlements, and second, how to (inter)connect the two operators to make such settlements. This dissertation uses a game-theoretic modelling approach to show that it is mutually beneficial economically if the MNO allows the MVoIPO to operate on its network. Further, a Service Level Agreement (SLA)-based Transparent Settlement Agreement (TSA) model is proposed to solve the first problem. The TSA model algorithm calculates the MVoIPO's throughput distribution at the edge of a UMTS Core Network (CN). This facilitates the determination of levels of conformance to the pre-set throughput thresholds and, subsequently, the issuing of compensation to the MVoIPO by the MNO after generating an economically acceptable volume of traffic. Further, possible network architecture to solve the second problem is suggested, by combining the TSA model algorithm, the UMTS CN, the IP Multimedia Subsystem (IMS), and the Online Charging System (OCS)

Not available / text

Not available / text

published_or_final_version / abstract / Electrical and Electronic Engineering / Doctoral / Doctor of Philosophy

published_or_final_version / abstract / toc / Electrical and Electronic Engineering / Doctoral / Doctor of Philosophy

published_or_final_version / abstract / toc / Computer Science and Information Systems / Master / Master of Philosophy

published_or_final_version / Electrical and Electronic Engineering / Master / Master of Philosophy