Fangmin Xu

Multiple controllers sleeping management in green software defined wireless networking

Software Defined Networking (SDN) promises the formidable configuration, vigorous evolution, and satisfactory performance, which are helpful to network operators. Hence, researchers in the field of wireless are increasing their attentions in SDN-related orientation. SDN is still in its initial period. Some essential issues still remain not completely resolved, and the scalability of control plane is the most intractable one with the explosive increase of wireless network traffic. To address this issue, many researchers have proposed multiple controllers to realize logically centralized control layer. Along with the number of controllers increasing, the other problem appears, which is the energy consumption in SDN. Especially the energy consumption in the wireless network has been greatly serious. In this paper, taking advantage of multi-controller architecture called Hybridflow, we propose an M-N policy multiple controllers sleeping mode by switching off redundant controllers when system is in the light traffic condition. Simulation results reveal that the proposed M-N policy multiple controllers sleeping mode achieves superior energy efficiency compared to no sleeping mode.

SVM-Based Sentiment Analysis Algorithm of Chinese Microblog Under Complex Sentence Pattern

With the development of Web2.0 era, as local information publishing and social networking platform of Twitter, microblog has become an important medium for people to share and propagate information. Sentiment classification for microblog has also become research hotspot in natural language processing field. By analyzing existing sentiment classification features and complex sentence patterns of microblog and directing at defects of current microblog sentiment classification in feature selection and extraction, this paper combined semantic relation between complex sentences and sentence features of complete sentence based on proposing features of sentence-level fine-grained embedding features and semantic features under complex sentence pattern so as to conduct effective analysis of microblog sentiment features under complex sentence context. It used SVM classification model to conduct comparative experiment, and results indicated that feature selection method proposed in this paper could improve performance of microblog sentiment analysis.

A Novel Space Information Network Architecture Based on SDN

The Space Information Network (SIN) is a full-spatial, full-time, full-frequency, multi-users-oriented information network which has the characteristics of complexity, heterogeneous, and openness. In this paper, we propose a software-defined space information network (SDSIN) to solve above-mentioned problems. This architecture based on the core idea of Software-Defined Network (SDN) that separate control and forwarding panel, and takes full use of the global coverage properties of Geostationary Orbit (GEO), powerful computing capacity of ground station, the predictability and regularity of constellation and the forwarding capacity of inner-satellite links (ISLs). Thus the network can allocate and optimize network resources from a global perspective so as to achieve flexible and efficient network configuration and management, and realize direct and effective control of spatial information network.

People Relation Extraction of Chinese Microblog Based on SVMDT-RFC

People relation extraction is a significant topic in information extraction field. While in traditional study, the feature of extraction lexical and semantic was attached importance to, and the function of classifier was neglected, furthermore, there is great difference between microblog language materials and that of tradition. When it mentioned traditional classification algorithm, its low correctness and the inaccuracy to identification of fuzzy sample become the reason of being used little. In this paper, the traditional classification algorithm was improved. Using SVMDT-Random Forest and we designed, the fuzzy sample classifying ability increased, which remedied the shortcomings of SVM and Random Forest effectively. By testing the microblog language materials, the result indicated that this method can improve the performance of people relation extraction.

Why Did You Opt to Switch off Me? Big Data for Green Software Defined Networking

The pull of Software Defined Networking (SDN) which promises the formidable configuration, vigorous evolution and satisfactory performance is appealing. However, SDN is still in its initial period. Many issues remain to be delved and the most intractable one is the scalability of control plane with the explosive increase of network scales. In order to address this issue, splendid researches have presented multiple controllers to implement logically centralized control layer. Whilst, with the growing number of controllers, another hot potato emerges, which is the energy consumption. In our previous research, we proposed multiple controllers sleeping management which reduces the energy but introduces more delay. In this paper, based on the previous research, we propose the energy saving solution based on big data analysis. We utilize big data to analyse topology data in the system and excavate influence of each controller according to the analytic results. On this basis, the system opts to switch off controllers with smaller influence so as to reduce delay. Simulation results reveal that the improved method with big data analysis achieves superior time delay compared with the previous one.

Congestion-aware signaling aggregation scheme for cellular based Underwater Acoustic Sensor Network

Underwater Acoustic Sensor Network (UASN) is the enabling technology for a wide range of applications including naval surveillance, oil platform monitoring, earthquake forewarning, climate and ocean observation, and water pollution tracking. However, considering the enormous acoustic sensor devices and unique services of UASN, some challenges are emerging to the traditional cellular access and core networks, especially the congestion problem due to simultaneously MTC traffic and signaling. Following this paradigm, the purpose of this paper is to support and optimize the signaling aggregation of UASN services based on cellular network. Taking LTE network as the example access network, a congestion-aware signaling aggregation scheme is designed considering the various requirements of UASN services and the congestion situation in the network entity. Theoretical analysis and experimental simulations show that this scheme can improve the system efficiency and greatly alleviate the signaling congestion, especially for the emergency UASN service.