- Advanced Optical Network Technologies
- Advanced Photonic Communication Systems
- Software-Defined Networks and 5G
- Optical Network Technologies
- Cryptography and Data Security
- Network Traffic and Congestion Control
- Privacy-Preserving Technologies in Data
- Internet Traffic Analysis and Secure E-voting
- Mobile Agent-Based Network Management
- Advanced Fiber Optic Sensors
- Advanced Malware Detection Techniques
- Image Retrieval and Classification Techniques
- Caching and Content Delivery
- Metallurgical Processes and Thermodynamics
- Software Reliability and Analysis Research
- IPv6, Mobility, Handover, Networks, Security
- Energy Efficient Wireless Sensor Networks
- Petroleum Processing and Analysis
- Software Engineering Research
- Advanced Image and Video Retrieval Techniques
- E-commerce and Technology Innovations
- Stochastic Gradient Optimization Techniques
- Image Processing and 3D Reconstruction
- Power Transformer Diagnostics and Insulation
China United Network Communications Group (China)
2013-2022
University of Science and Technology of China
2014-2019
Nanjing University of Science and Technology
2012-2013
Elastic optical networks (EONs) facilitate agile spectrum management in the layer. When coupling with software-defined networking, they function as EONs (SD-EONs) and provide service providers more freedom to customize their infrastructure dynamically. In this paper, we investigate how overcome fragmentation SD-EONs OpenFlow-controlled online defragmentation (DF), conduct system implementations highly-efficient DF. We first consider sequential DF, i.e., scenario that involves a sequence of...
We propose and experimentally demonstrate a control-plane framework to realize online spectrum defragmentation (DF) in software-defined elastic optical networks. Experimental results show that the DF enabled by OpenFlow reduces blocking probability effectively.
In this letter, we present flexible traffic engineering (F-TE), a solution to achieve highly efficient in hybrid network where IPv4 and IPv6 protocols co-exist. By leveraging the programmability of OpenFlow (OF), design OF system facilitate IP forwarding interchanging (i.e., switching packets between at each hop according status). The is implemented real testbed. We conduct experiments with video streaming verify that proposed F-TE can improve throughput effectively adaptive interchanging.
This paper summarized the basic process of software vulnerability prediction using feature-based machine learning for first time. In addition to sorting out related types and basis features definition, advantages disadvantages different methods are compared. Finally, this analyzed difficulties challenges in research field, put forward some suggestions future work.
Computing the ratio of secure summations (RSS) is one most important tools for privacy preserving distributed data mining. It refers to such a problem; given n parties and their respective secret values (x <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">i</sub> ,y ), where i=1, ..., , how can they get value Σ xmlns:xlink="http://www.w3.org/1999/xlink">i=1</sub> <sup xmlns:xlink="http://www.w3.org/1999/xlink">n</sup> x / y without disclosing In...
In this paper, we propose to use IP-forwarding interchanging (i.e., exchanging packets between IPv4 and IPv6 according the network status) enabled by OpenFlow realize quality-of-service (QoS) aware flexible traffic engineering (F-TE) in a hybrid where coexist. Specifically, have different IP domains interconnected switches managed centralized controller, design system facilitate online, adaptive per-flow-based for link utilization optimization with considerations on applications' QoS...
A new non-contact method for predicting the basic oxygen furnace(BOF) end point carbon content is proposed in this study. model applying flame spectrum of converter vessel mouth constructed to carry out prediction. This consists two parts, viz. a classifier based on support vector classification classify whole period one BOF heat into main phases, and relevance machine working at posterior phase predict content. Compared with current methods prediction, can make better use information mouth....
Two main types of TWDM PON are distinguished by the locations AWG and splitter. Three self-protection mechanisms i.e., dual fiber bridging pair (DFBP), single (SFBP), ring (SFBR) designed based on can be directly used TDM part optical distribution networks (ODN). After combined with spectral periodicity property AWG, three schemes in WDM too. Each ONU use monitoring reporting signal to detect failure. DF then get without any service interrupt loss. The calculation results show proposed keep...
In this paper, we introduced the basic concept and framework of Intelligent ODN. Based on our study technology implementations, summarized implementation solutions And based real network operation experience, firstly provide a suit function specifications ODN networks. Provided can be used as guide for realization
As the scalable deployment of PON network and maturity next generation technologies, optical link fault management should be well considered. OTDR is now under study in industry. Based on our real operation experience, we give an analysis use case functional requirements OTDR, which could used as a guide development implementation.
中国科学院 "感知中国" 先导专项子课题 (批准号: XDA06010303)、国家自然科学基金 61371117) 和电子科技大学光纤 传感与通信教育部重点实验室开放课题资助项目 摘要 本文基于软件定义网络 (software defined betworking, SDN) 概念, 利用 OpenFlow 技术设计并实 现了一种新颖的 IP 融合切换技术.该技术基于当前网络状况对 数据流的下一跳转发方式进行灵 活选择, 在 IPv4 与
The idea of ensemble learning can be used to solve problems about privacy preserving distributed data mining conveniently. Owners datasets get an integrated model securely just by sharing and combining their sub models which are built on respective sample sets, generally the is more powerful than any model. However, may cause serious in some cases. So this paper, we present a new method, based holders integrate polynomial regression efficiently without them, optimal combination In addition...
This paper proposes an access network requirements analysis method based on similar service feature values clustering. The data are collected from the real network, including devices and users' information. K-means ++ algorithm is adopted to cluster PON ports, of users connected them. 3 classifiers selected for classification training prediction. They K-nearest neighbors, SVMs perceptron. In case no clustering, neighbor performs better, correct rate about 91.1%. And if divided into 5 groups...
We demonstrate a carrying capability evaluation system, which can evaluate and predict the access network capacity efficiency by extracting detail status trends from hybrid data sources based on machine learning.
The upgrading of PON network is an important part to meet the needs and development user traffic in future. In this paper, data collection style specified by telemetry, partial least squares method used predict future trend, so as formulate cutover strategy._Finally gap between peak OLT different time periods narrowed, load balance achieved.