Cloud computing and service computing are merging together rapidly. This process bring large challenges to data mining and analysis. Our aim is to develop a system for service deployment, analysis, and recommendation to boost API economy. In this paper, we show our partial outcomes.
The rapid development of mobile networks has revolutionized the way of accessing the Internet. The exponential growth of mobile subscribers, devices and various applications frequently brings about excessive traffic in mobile networks. The demand for higher data rates, lower latency and seamless handover further drive the demand for the improved mobile network design. However, traditional methods can no longer offer cost-efficient solutions for better user quality of experience with fast time-to-market. Recent work adopts SDN in LTE core networks to meet the requirement. In these software defined LTE core networks, scalability and security become important design issues that must be considered seriously. In this paper, we propose a scalable channel security scheme for the software defined LTE core network. It applies the VxLAN for scalable tunnel establishment and MACsec for security enhancement. According to our evaluation, the proposed scheme not only enhances the security of the channel communication between different network components, but also improves the flexibility and scalability of the core network with little performance penalty. Moreover, it can also shed light on the design of the next generation cellular network.
Heterogeneous information networks (e.g. cloud service relation networks and social networks), where multiple-typed objects are interconnected, can be structured by big graphs. A major challenge for clustering in such big graphs is the complex structures that can generate different results, carrying many diverse semantic meanings. In order to generate desired clustering, we propose a parallel clustering method for the heterogeneous information net-works on an efficient graph computation system (Spark). We use a multi-relation and path-based method to create similarity matrices, and implement our method based on graph computation model. It is inefficient to directly use existing data-parallel tools (e.g. Hadoop) for graph computation tasks, and some graph-parallel tools (e.g. Pregel) do not effectively address the challenges of graph construction and transformation. Therefore, we implemented our parallel method on the Spark system. The experiment results of clustering show our method is more accuracy.
With the development of the cellular network in the last decade, the number of IoT devices is growing exponentially and IoT applications are becoming more complex with higher requirements for Key Performance Indicators (KPIs) such as latency, accuracy and energy consumption. To address these challenges, the edge computing paradigm is often adopted to push the computing capabilities to the edge servers nearest to end-users. However, the Quality of Experience (QoE) of IoT applications is still hard to guarantee because the nearest edge servers change while users roam around. In this paper, we propose MeFILL, a Multi-edged Framework for Intelligent and Low Latency mobile IoT applications, which reduces the latencies and improves the reliability with the seamless handover of IoT devices between edge servers and leverages the Distributed Deep Learning (DDL) collaboration among edge servers. The comparison experiments show that MeFILL can effectively optimize performance KPIs of mobile IoT applications.
The number of Web services are growing rapidly on the Internet. Topics of services are becoming various. Semantic-based keyword search is used to retrieve proper services for service consumers. According to the semantic information implied in service database, we build a topic model to cluster and management related services. Our service recommendation approach can extract service patterns from correlated topics in semantic service descriptions. We use Latent Dirichlet Allocation to obtain the service patterns; and use Concept lattice to model the correlation between the extracted topics. Higher precision results are obtained in the experiments.
With the continuous increase of SaaS (Software as a service) market demand, the competition of SaaS platforms is also increasing. Using deep learning to conduct sentiment analysis on users' reviews can provide a better reference for users to choose SaaS services. But some recent studies have pointed out that deep learning models are easily affected by attackers. When an attacker adds a very small disturbance to the input sample, the deep learning model will give wrong classification results. In order to further study the vulnerability of deep learning models in the field of sentiment analysis, we designed a model of review intention recognition based on DIKW Pyramid, which can recognize the intention of the review, and we also propose an attack algorithm based on synonym substitution of evaluation words and evaluation reasons, which can generate usable and efficient adversarial samples under black box attack. We designed and implemented adversarial attacks on three deep learning models of LSTM, Bi-LSTM and CNN, and combined with Random adversarial attack algorithm, population based adversarial attack algorithm and DeepWordBug adversarial attack algorithm. The experimental results proved that the adversarial attack algorithm based on synonym substitution of evaluation words and evaluation reasons is better than the other three attack algorithms in terms of attack success rate and word replacement rate.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)