Energy consumption has become a major concern to the widespread deployment of cloud data centers. Many techniques have been devised to help reduce energy consumption for cloud data centers that consist of a large number of identical servers, including dynamic allocation of active servers, consolidating diverse applications, and adjusting the CPU frequency of an active server. However, these techniques normally have a high migration and low resource utilization. CPU and memory is the dominant factors of the performance and energy consumption and whose allocation determines the energy efficiency of cloud system. Leveraging these techniques, we focus on the problem of VM placement, propose a heuristic greedy algorithm to implement VM deployment and live migration to maximize total resource utilization and minimize energy consumption, which is based on energy-aware and quadratic exponential smoothing method to predict the workloads. Our heuristic algorithm makes CPU-intensive services and memory-intensive services mapped to the same physical server more complementary. The experiment results show that there is significant improvement in the aspect of energy saving, workload balancing and scalability, compared with single-objective approaches based on CPU utilization.

The Cyber-physical system is highly desirable for real-time perception and dynamic control, especially in task scheduling and network communication for large and complex engineering systems. In most of applications, the resource scheduling of cyber and physical network are independent to each other, which may cause severe deterioration of the network performance. To implement the cognitive cooperation between the upper layer of industrial services and the underlying network, we propose a co-optimization method in this paper. We adopt the SDN technology to extract the network control logic, which is integrated with the upper-level task scheduling. This co-optimization scheme can dynamically calculate optimal end-to-end virtual paths over the underlying network infrastructures. Meanwhile, it can also adjust the task scheduling of the upper-level network by signals derived from the underlying network. Simulation results show that our proposed scheme outperforms the traditional method in terms of task acceptance rate, average end-to-end delay and network load balance degree.

In multi-end-to-end path request planning, the control plane may not be able to meet all path request requirements under limited bandwidth resources. Moreover, suboptimal path planning can lead to localized network congestion, which in turn causes an overall imbalance in network load. Therefore, the multi-domain control plane needs to consider more network resource states during the path selection, such as link weights, load saturation, and resource occupancy rates, in order to select the optimal paths to maximize the satisfaction of data plane requirements while maintaining network load balance. To address such issues, we first derive a cross-domain communication load balancing objective function based on network modeling. Through collaborative processing among multi-domain controllers, the coordinated planning of cross-domain paths and the collaborative installation of flow tables are achieved. Then, we transform the cross-domain path planning problem into a clique-finding problem under a set of backup paths. Finally, we provide a heuristic approximate solution method for this problem. In terms of cross-domain communication, we adopt a collaborative approach among multiple controllers to achieve coordinated planning of cross-domain paths and collaborative installation of flow tables. Simulation results show that our proposed scheme outperforms the traditional method in terms of path allocation success rate, network load balancing degree, and data transmission delay, especially in cross-domain communication under high-density path requests in SDN networks.