Hua Zou

Web service composition algorithm based on TOPSIS

Abstract Independent quality of service (QoS) registration center is difficult to monitor QoS with lots of requests in Internet environment with a large number of services and lots of service requests. It is able to reduce the load and decrease the failure rate by using multiple QoS registration centers to monitor QoS of some services with lots of requests. In addition, the heterogeneous QoS data close to users habitual expression and adapts to the environment with complexity and uncertainty. This paper introduces a novel Web service composition algorithm based on technique for order preference by similarity to an ideal solution (TOPSIS) (WSCA_TOPSIS) to solve the above difficulties for the first time. WSCA_TOPSIS can support multiple decision-makers and heterogeneous QoS data. It includes three main steps: normalizing decision matrix, evaluating alternatives synthetically and evaluating group alternatives synthetically. Other contributions of the paper include a novel algorithm of aggregating QoS and a set of experiments that demonstrate the benefits and effectiveness of our approach. Experimental results show that the proposed algorithm can better support Web service composition with heterogeneous QoS data and multiple decision-makers.

Multi-QoS Effective Prediction in Web Service Selection

The rising development of service-oriented architecture makes Web service selection a hot research topic. However, there still remains challenges to design accurate personalized QoS prediction approaches for Web service selection, as existing algorithms are all focused on predicting individual QoS, without considering the relationship between them. In this paper, we propose a novel Multi-QoS Effective Prediction (MQEP for short) problem, which aims to make effective Multi-QoS prediction based on Multi-QoS attributes and their relationships. To address this problem, we design a novel prediction framework Multi-QoS Effective Prediction Approach (MQEPA for short). MQEPA first takes use of Gaussian method to normalize the QoS attribute values, then exploits Non-negative Matrix Factorization to extract the feature of Web services from Multi-QoS attributes, and last predicts the Multi-QoS of unused services via Multi-output Support Vector Regression algorithm. Comprehensive empirical studies demonstrate the utility of the proposed method.

A Novel Approach to Optimize Information Dissemination in IMS Presence System

Presence is an important service in IMS. The user in IMS would use presence as a necessary component. But the more user uses, the harder presence information transmits. Previous researches show that the traffic increases rapidly with the presentity number grows. In this paper, we propose an approach to optimize information dissemination in IMS Presence System. Comparing with previous researches, a new architecture is set up. In this architecture, presence information transmission is modeled as a growing tree. A k sequence method is proposed to construct the tree. And some key algorithms are introduced for the tree adjustment. With a mathematical analysis, the new method is more suitable for presence information dissemination. And in the experiment, the result shows that the method can do a better job with smaller traffic and jitter.

Small business-oriented index construction of cloud data

With the development of cloud computing, data owners (businesses and individuals) are motivated to outsource their local complex database systems to public cloud for flexibility and economic savings. But for the consideration of users privacy, personal data has to be special treatment locally before outsourcing to the cloud server. Considering the large number of data users and documents in cloud, it is crucial for data owner to construct an index for their data collection, which increases the cost of the data owner. Related works focus on the searches on encrypted database but rarely consider the overhead of the index construction for data owner and the extensions of the index. Although traditional index construction methods of information retrieval have been widely studied, direct application of these methods would not be necessarily suitable for our scenario. Thus, enabling an efficient index construction service is of paramount. In this paper, we define and solve the problem of index construction on small business (SBIC). Among various index methods, we choose inverted index method. An inverted index is an index data structure storing a mapping from content to its locations in a set of documents. The purpose of it is to allow fast full text searches.We firstly propose a basic SBIC scheme using Lucene (an open source project for web search engine), and then significantly improve it to meet efficient keyword extraction requirement and multi-type files demand. Thorough analysis design goals(see section 2.3) of proposed schemes is given, extensive experimental results on the dataset further show proposed scheme indeed introduce low overhead on time and space.

Using ‘word of mouth’ to find community structure in an online IMS system

Abstract Presence is a feature that is used widely in lots of system, especially in IMS. The presence system in IMS also provides application developer and end user with a chance to new style applications. But the presence system in IMS has some problems. Previous researches show that the traffic increases rapidly with the presentity number grows. In this study, we would treat the system in a different way by introducing some social networking concepts to the architecture. Moreover, we would focus on how to apply and generate community in the dynamic and online environment. A word of mouth marketing principle would be used to model the community generating procedure. The presence system would works like a marketing environment. Experiment with simple math analysis show this approach is effective for distributed community generating.

A Novel Path Selection Algorithm Based on Graph-Cut to Ensure Security in SDN.

To ensure the safety of data transmission in SDN network, a practical approach is to forward data by time division in multiple reachable paths. Considering multiple data transmission processes between the sender and the receiver based on the mechanism above, the overlap of the paths is inevitable, which may make the utilization rate of some switches or critical paths too high and other switches are relatively idle, then the transmission efficiency would be too low. This paper put forward a path selection algorithm based on graph-cut to solve the problem. Also, taking switch weight under consideration. In addition, effectiveness is taken into account. Considering to the large scale of actual networks and uneven occupancy rate of switches, this algorithm reflects its advantage.

A SDN Security Transmission Service Using Multi-connection and Time-Slot Scheduling

SDN (Software-Defined Networking) provides a new approach to improve network security. In this paper, we design and implement a security service of transmission channels in SDN, which will efficiently lower the percentage of information leakage on the data plane of SDN networks. The core of the service is forwarding strategies based on multi-connection and time-slot scheduling, which functions on SDN controllers. According to the time-slot strategy, every data flow is transmitted on one path in a network at a certain moment, then in next time-slot it will be transmitted on another path, and so on. The multi-connection strategy allows to transmit data through multiple links at a same time. There is a client-side subsystem deployed in user terminals to process the data to send or receive. The processing of data including splitting and encrypting in application layer, coordinating with the forwarding strategies, reduces probability of the whole data flow being intercepted.

Assessment of Performance in Data Center Network Based on Maximum Flow

Recently, data center networks (DCN) have received significant attention from the academic and industry. However, researches of DCN are mainly concentrated on the improvement of network architectures and the design of routing protocols, or the performance evaluation from the perspective of node importance. In contrast to existing solutions, in this paper, we propose using maximum-flow theory to assess the network performance. Firstly, we abstract two kinds of typical DCN architectures and then formulate and convert the performance analysis of those architectures into a maximum-flow problem including a supersource and a supersink. Secondly, we get the value of maximum-flow by using Edmonds and Goldberg algorithm. Last but not the least, based on the theory of maximum-flow and Minimal cut sets, we get the critical edges for each architecture. Extended experiments and analysis show that our method is effective and indeed introduce low overhead on computation. In addition, the method and issues observed in this paper is generic and can be widely used in newly proposed DCN architectures.

Efficient Service Deployment by Image-Aware VM Allocation Strategy

Cloud computing centers face the key challenge of rapidly provisioning diverse virtual machines VMs to efficiently deploy services in an elastic and scalable manner. VMs are created by instantiating VM images which encapsulate the required applications or services with their support environment. Thus, the rapid provisioning of VM images is an important factor of efficient service deployment. However, rapid service deployment is determined not only by efficiently VM images fetch but also by request patterns, placement and caching strategies. Thus, in this paper, we explore those effects, and present an image-aware VM allocation strategy to speed up service deployment. Our experiments show that our proposed algorithm largely improves the deployment efficiency. Moreover, the algorithm improves the utilization of cache space and alleviates network traffic.

Study on a session description model for cognitive communication services adaptation in converged network

In future, user-centric cognitive communication services under convergent environments will take appropriate adaptations to enhance service experience according to users needs and service contexts. By analyzing constituent elements of session and connection between them, a novel Adaptive Session Description Model—ASDM, is proposed, consisting of a three-layer session structure: node session, terminal session and service session. This model gives the concept of session isomorphism and corresponding judge conditions. This paper compares the ASDM with other existent model, and takes a real adaptive session scenario as an example to validate practicability and correctness of this model.