With the wide application of virtualization technology in cloud data centers,how to effectively place virtual machine(VM)is becoming a major issue for cloud providers.The existing virtual machine placement(VMP)solutions are mainly to optimize server resources.However,they pay little consideration on network resources optimization,and they do not concern the impact of the network topology and the current network traffic.A multi-resource constraints VMP scheme is proposed.Firstly,the authors attempt to reduce the total communication traffic in the data center network,which is abstracted as a quadratic assignment problem;and then aim at optimizing network maximum link utilization(MLU).On the condition of slight variation of the total traffic,minimizing MLU can balance network traffic distribution and reduce network congestion hotspots,a classic combinatorial optimization problem as well as NP-hard problem.Ant colony optimization and 2-opt local search are combined to solve the problem.Simulation shows that MLU is decreased by 20%,and the number of hot links is decreased by 37%.
Ethernet link aggregation, which provides an easy and cost-effective way to increase both bandwidth and link availability between a pair of devices, is well suited for data center networks. However, all the traffic splitting algorithms used in existing Ethernet link aggregation are flow-level which do not work well owing to the traffic characteristics of data centers. Though frame-level traffic splitting can achieve optimal load balance and the maximum benefits from aggregated capacity, it is generally deprecated in most cases because of frame disordering which can disrupt the operation of many Internet protocols, most notably transmission control protocol (TCP). To address this issue, we first investigate the causes of frame disordering in link aggregation and find that all of them either are no longer true or can be prevented in data centers. Then we present a byte-counter frame-level traffic splitting algorithm which achieves optimal performance while causes no frame disordering. The only requirement is that frames in a flow are the same size which can be easily met in data centers. Simulation results show that the proposed frame-level traffic splitting method could achieve higher throughput and optimal load balance. The average completion time of different sized flows is reduced by 24% on average and by up to 46%.
Implementing face recognition efficiently to real world large scale dataset presents great challenges to existing approaches. The method in this paper was proposed to learn an identity distinguishable space for large scale face recognition in MSR-Bing image recognition challenge( IRC). Firstly,a deep convolutional neural network( CNN)was used to optimize a 128 B embedding for large scale face retrieval. The embedding was trained via using triplets of aligned face patches from Face Scrub and CASIA-Web Face datasets. Secondly,the evaluation of MSR-Bing IRC was conducted according to a cross-domain retrieval scheme. The real-time retrieval in this paper was benefited from the K-means clustering performed on the feature space of training data. Furthermore,a large scale similarity learning( LSSL) was applied on the relevant face images for learning a better identity space. A novel method for selecting similar pairs was proposed for LSSL. Compared with many existing networks of face recognition,the proposed model was lightweight and the retrieval method was promising as well.
In the cloud data centers,how to map virtual machines(VMs) on physical machines(PMs) to reduce the energy consumption is becoming one of the major issues,and the existing VM scheduling schemes are mostly to reduce energy consumption by optimizing the utilization of physical servers or network elements.However,the aggressive consolidation of these resources may lead to network performance degradation.In view of this,this paper proposes a two-stage VM scheduling scheme:(1) We propose a static VM placement scheme to minimize the number of activating PMs and network elements to reduce the energy consumption;(2) In the premise of minimizing the migration costs,we propose a dynamic VM migration scheme to minimize the maximum link utilization to improve the network performance.This scheme makes a tradeoff between energy efficiency and network performance.We design a new twostage heuristic algorithm for a solution,and the simulations show that our solution achieves good results.
Social tagging is one of the most important characteristics of Web 2.0 services, and social tagging systems (STS) are becoming more and more popular for users to annotate, organize and share items on the Web. Moreover, online social network has been incorporated into social tagging systems. As more and more users tend to interact with real friends on the Web, personalized user recommendation service provided in social tagging systems is very appealing. In this paper, we propose a personalized user recommendation method, and our method handles not only the users' interest networks, but also the social network information. We empirically show that our method outperforms a state-of-the-art method on real dataset from Last.fm dataset and Douban.
