Energy-efficient Intelligent Routing Theory And Technology Research For Cloud Computing

With the increasing number of IT devices in cloud computing, the scale of data centers is becoming larger as well. The issue for energy consumption of cloud environment has become increasingly prominent, and energy consumption has become a major obstacle to sustainable development of the Internet and telecommunication industries. In the traditional design of network system, network reliability is improved through redundant links and devices in response to the sudden fault failure. However, it doesn’t take the network energy efficiency into consideration. Hence, with the expansion of network size and the continuous update of network devices, some problems such as high energy consumption, low efficiency and high waste become a major challenge to achieve energy saving.In cloud computing, each virtual machine for resource scheduling is random and sudden, as opposed to the traditional Internet, the network traffic of cloud computing has greater volatility and time variability. In such a highly dynamic change of network traffic, how to reduce energy consumption and improve the efficiency of communication is a major problem with which cloud computing network is faced. This thesis studies the energy-efficient intelligent routing technology for cloud computing, to improve network performance and network efficiency. In the traditional backbone network, the path selection of IP flows is mainly based on various routing protocols, and these algorithms actually select the link bandwidth with maximum resource or the minimum hops for communication, and such way of choices often run counter for the purpose of energy saving. Therefore, just relying on the current routing protocol cannot achieve the energy savings of cloud computing network. Based on a flexible separation of cloud computing network, this thesis carries out a research of energy-efficient mechanism for cloud computing from three kinds of energy-efficient and intelligent routing algorithm.In this thesis, we use two methods to achieve energy-efficient networks. The one is energy-efficient routing strategy and the other is sleeping strategy. This thesis puts forward three routing algorithms based on intelligent routing technology. The first one is energy-efficient routing algorithm based on Niche Genetic Algorithm (NGAERA); the second one is routing algorithm based on Energy efficiency and Cognitive theory (ECRA); the third one is energy-efficient routing algorithm based on time series forecasting theory and sleeping strategy (ARGMERA). NGAERA algorithm gets the initial population by random Depth First Search algorithm and fitness values, and then uses the Niche Genetic Algorithm to make genetic manipulation, and obtain the global optimal solution of minimum energy consumption for network. However, the time complexity of this algorithm is higher. Therefore, this thesis puts forward ECRA algorithm, on the basis of energy efficiency priority thought, it makes the classical Dijkstra routing algorithm which belongs to OSPF combine with cognitive theory, to achieve the weights adaptive minimum energy consumption routing algorithm, which can change link weight dynamically according to network energy consumption, and ensure that the energy consumption of network is minimum. Finally, due to the highly dynamic change characteristics of cloud network traffic, ARGMERA routing algorithm is proposed. This algorithm is divided into two parts, in the first part, the time series forecasting theory is used to build traffic perception model; in the second part, ECRA algorithm is used to establish sleeping strategy and energy-efficient routing strategy, to achieve a meaningful prototype of energy-efficient routing technology.