Research On Node And Link Selection Criteria And Embedding Algorithms For Energy Efficient Virtual Networks

As the scale of Internet grows, the excess supply of resources and the redundant design principles deteriorate the overuse problem of energy. Network virtualization, as a supporter for innovation and sustainable development of network, enables us to consider the energy-aware network deployment. One of the core ideas of network virtualization is to construct real-time virtual networks for different requirements over the same infrastructure networks, and how to embed virtual networks to physical networks is the key issue for network virtualization. On the one hand the virtual network mapping should satisfy different requirement, on the other hand it should optimize network resource allocation according to network dynamics, and reduce the energy consumption.An important goal for network virtualization is the energy-aware network deployment. Recently, there exist two main problems that need to be improved, one is energy-aware node and link selection criteria and the other is the organization of energy-aware virtual network embedding algorithms. Then, our contributions are as follows:1. A time and energy aware virtual network embedding(TEAVNE) algorithm is proposed. When selecting the embedding objectives for the virtual nodes and links, we minimize the using duration of physical nodes and paths. Compared with the methods without the consideration of time, the simulation results show that the novel method TEAVNE can reduce the energy consumption more efficiently. Furthermore, TEAVNE can achieve better performance under the scenario with sufficient resources.2. Based on TEAVNE algorithm, we propose an energy efficient virtual network embedding based on constraint concentration mapping(EEVNE-CCM) algorithm. When selecting the embedding objectives for the virtual nodes and links, we minimize the maximum resource utilization. In this way, the nodes and links are turned on in advance, which results in efficient energy-saving. Compared with TEAVNE, the simulation results show that the novel scheme can both improve resources allocation and reduce the energy consumption greatly.3. An adaptive cooperative coevolution based Particle Swarm Optimization Algorithm(ACCPSO) for energy aware virtual network embedding is presented. We proceed to the problem of energy aware virtual network embedding which has a discrete solution set. Through the analysis to the aggregation level of particle swarm, we define the degree of aggregation to guide the process of search. Finally, the ACCPSO algorithm is compared with the original algorithm, and simulation results show that our algorithm can find better solution under the same searching times.