Research On Load Balancing For Future Internet Resources

The rapid development makes the internet nowadays face huge challenges on aspects of scalability, security, mobility, controllability and so on. And these challenges have become the choke points for internet development. This dissertation, which is based on national973projects--"SOFIA:Service-Oriented Future Internet Architecture and mechanism studies", will do the research about "load balancing for future internet resources" in hopes of providing reference to design of the new generation of Internet architecture, as well as laying foundation for the researches of a new generation Internet resources management. Future Internet will make many kinds of resources work together to provide services to Internet users and inevitably face new challenges in terms of resource management. We need a new thinking and perspective to change the understanding of this issue. Therefore carrying out the researches on load balance for future internet can ensure the reasonable work of resources as well as the work efficiency and resource utilization of the Internet.In the process of research, since the "future Internet" is a new thing, there is not a unified and clear understanding in the academic world. So in order to study the issue of load balancing for future internet resources, it should come in the first priority that understanding about what is the Future Internet. And on this basis, this dissertation does further study of Internet resource innovation management model and framework. Then internet-level and data-center-level load balancing allocation strategies, which are self-organizational and self-adaptive, are further studied based on the framework mentioned above. The main research contents and conclusion of this dissertation are:(1) From the angle of the service-providing and network management, I analyze the main research contents of "Future Internet" and characterize its main features. Since the current researches on " Future Internet " mostly focus on network construction, network technology and other technical perspectives, so we need to study from the angles of service and management, as the future internet is service-oriented and self-management. In this part, the paper extracts the main research contents of future internet and depicts the main features of future internet in service providing perspective, after analyzing the projects about "future internet" lunched by countries all over the world. And do the following research around its " awareness" feature.(2) This part of paper puts forward a future Internet resource management philosophy, service-oriented network resource management models and framework, and load balancing models for future internet resource based on "awareness" and "situation awareness", which will alter network resources management from passive to active. By studying the concept of situational awareness and its mature theory in military, medical, and network security, this dissertation proposes the component requirements and functional requirements of future Internet resources situational awareness model, based on witch this dissertation presents service-oriented network resource management situational awareness model and its framework. This may lay foundation for network resource management platform in the future. Finally, this dissertation will bring out the future Internet architecture facing load balancing and put forward the following researches based on it.(3) Load balancing allocation strategy and algorithm based on game theory are designed in perspective of network resource. This part of the manuscript analyzes the review of research on "load balancing" algorithms to abstract the future of the Internet resource management issue into a hierarchical resource allocation model by combining the research content of (1) and (2). In the model, I take into account the resource requester’s selfishness and isomorphism psychology of resources, and formulate the non-cooperative game among users and cooperative game among resource nodes. The model combines the advantages of cooperative and non-cooperative game to make up for deficiencies between them, and improve the fairness index between resources on the basis of ensuring the fairness index of users. The algorithm is superior to other non-cooperative game algorithms according to system efficiency. At the same time, the algorithm reduces the system overhead about the load information maintenance and can achieve energy conservation.(4) This part of dissertation proposes a load balancing allocation strategy which can meet self-organized and self-adaptive requirement from the perspective of consumers based on the principles of economics. In this part, users and resources are put into the competition market environment, and market regulating mechanism is used to make the system to achieve load balancing state. Firstly, according to the microeconomics theory, the formation mechanism of resource price in the market and the behavior of consumers and sale agents in the competitive market are simulated. Simulating the market competition of consumers’ computing resources is to analyze the resource allocation in the steady market state. Results show that the algorithm can reasonably simulate the market behavior of consumers and resource price formation mechanism in the market, this strategy can not only make the resource be in a state of load balance, but also increase the market income of agents (network service providers).