검색 상세

Dynamic Notification Aggregation CoAP Based on Markov Chain

  • NGUYEN TUNG ANH (대학원 컴퓨터정보학과 IoT(Internet of Things))

초록/요약

Nowadays, wireless sensor networks (WSNs) have greatly influenced many aspects of society. WSNs can be defined as a network of spatially dispersed devices and sensors that are capable of exchanging information collected from a manipulated area through a wireless link. Devices and sensors are nodes acting as tiny computers. They are embedded devices in charge of detecting physical or environmental conditions, producing output in the form of an electrical signal that is delivered to an operator for further processing. WSNs are able to cover a great number of spatially distributed nodes, from just a few to several hundred, or even thousands. Size and cost constraints on sensors and embedded devices generate obstacles for WSNs, for example, needing low energy consumption, and having limited memory and restricted computational ability and communications bandwidth. Furthermore, as a demand of development, WSNs must be compatible with the Internet. Thus, Constrained Application Protocol (CoAP) and Observing Resources in CoAP (CoAP/Observe), which follow the client/server paradigm, were proposed by the Internet Engineering Task Force (IETF) in the context of low-power and low-rate networks for tackling the aforementioned problems. CoAP is a representational state transfer application protocol for constrained nodes and constrained (e.g, low-power, lossy) networks. Although CoAP is designated under the client/server paradigm, a node can be a server or a client, depending on its function. CoAP creates favorable conditions to communicate between nodes and with the wider Internet using identical protocols. Thanks to compatibility with Hypertext Transfer Protocol (HTTP), CoAP provides a way to exchange information not only between nodes in constrained networks but also between constrained network and general nodes on the Internet. CoAP/Observe extends the original CoAP protocol and relies on the design pattern from the Gang of Four (GOF). CoAP/Observe is a model of the best-effort approach, which tries to keep the same state in both client and server with low power consumption. To reduce the power consumption in nodes, states are transmitted to the client when resources meet particular conditions, called CoAP Conditional Observe. This method also makes the sending time between two states a random variable, which generates diverse approaches to mathematically model for the reduction of power consumption. Due to increasing amounts of sensor overtime, many responses will be transmitted over WSNs. Although other CoAP Conditional Observe methods have made great contributions to the diminished energy of nodes in a constrained network, they are still not resilient. In this thesis, an approach to reducing energy consumption in constrained devices based on CoAP Conditional Observe is proposed. The proposed method relies on a Markov process with its probability transition matrix for predicting packets to combine in order to save energy. It is plausible that there are many methods to forecast opportunities for aggregation of packets. But the Markov process was chosen in this research because its approach requires low computational power and energy, which is appropriate to constrained nodes and WSNs. In this thesis, a mathematics model and an algorithm are proposed. For evaluating the efficiency of the proposed method, CoAP and CoAP Conditional Observe are used as baselines where energy consumption and waiting time are considered.

more

목차

I. INTRODUCTION 1
II. RELATED WORK 4
II.1. Constrained Application Protocol 4
II.1.1. Constrained Application Protocol Introduction 4
II.1.2. Messaging Model 5
II.1.3. Request/response Model 8
II.1.4. Proxying 11
II.2. Observing Resource in Constrained Application Protocol 12
III. PROPOSED MODEL 15
III.1. Aggregation 15
III.1.1. Scheduling 18
III.1.2. Maximum likelihood estimation 22
III.1.3. Algorithm 24
IV. SIMULATION AND RESULT 28
IV.1. Measurement of power consumption 28
IV.2. Waiting time in proxy 42
V. CONCLUSION 47
REFERENCES 48
ACKNOWLEDGEMENT 50

more
반출 Meta View 목록