Lianchen Liu

An integrated resource management and scheduling system for grid data streaming applications

Grid data streaming applications are novel from others in that they require real-time data supply while the processing is going on, which necessitates harmonious collaborations among processors, bandwidth and storage. Traditional scheduling approaches may not be sufficient for such applications, for they usually focus on only one aspect of resources, mainly computational resources. A resource management and scheduling system for such applications is developed in this paper, which is responsible for enabling their running based on Globus toolkit. An integrated scheme is proposed, including admission control, application selecting, processor assigning, allocation of bandwidth and storage, with corresponding algorithms elaborated. Evaluation results show excellent performance and scalability of this system.

A three-layered method for business processes discovery and its application in manufacturing industry

Business process discovery is a basis for analyzing an enterprises current business and is a critical key to enterprises seeking to be more process-centric today. Challenges in business process discovery include the complexity of the enterprise and the interactions among its units, the inaccuracy and incompleteness of available business information, and the dynamic changing businesses in the enterprise. In this paper, a three-layered architecture mixing the top-down and bottom-up methodology is proposed to address the above issues. Technical details of the three layers, namely the Component layer, the Operation Integration layer and the Operation layer are discussed. Moreover, a model checking based verification approach is also integrated in the method to verify the discovered business process against the desired business requirements from analysts. Finally, a case study in a furniture manufacturing enterprise in China is illustrated to prove the effectiveness of our method.

The study on simulation grid technology for equipment resource sharing system

It is difficult to realize equipment resources sharing on network. Based on grid and simulation modeling technique, simulation-grid architecture is presented. Simulation models of sharing equipments are established in grid. Manipulating and optimally scheduling remote equipments are realized by operating the simulation models in simulation-grid system. The proposed architecture provides a method to equipment resources sharing and management on network and would be widely used.

Research on Equipment Resource Scheduling in Grids

In order to resolve equipment resources scheduling problem in grid we develop an equipment resource scheduling system with the equipment grid portal and scheduler including scheduling algorithm, information service and its architecture. The proposed technique provides a method to equipment resources sharing and management on network and will be widely used.

How Overlapping Community Structure Affects Epidemic Spreading in Complex Networks

Many real-world networks exhibit overlapping community structure in which vertices may belong to more than one community. It has been recently shown that community structure plays an import role in epidemic spreading. However, the effect of different vertices on epidemic behavior was still unclear. In this paper, we classify vertices into overlapping and non-overlapping ones, and investigate in detail how they affect epidemic spreading respectively. We propose a SIR epidemic model named ICP-SIR (Inner-Community Preferred Susceptible-Infective-Recovered) where the inner-community and inter-community spreading rates are different. We consider the case where epidemic process is started by immunizing and infecting multiple overlapping or non-overlapping vertices. The epidemic model is applied on both synthetic and real-world networks. Simulation results indicate that compared to non-overlapping vertices, overlapping vertices play a vital role in spreading the epidemic across communities. The result of our research may provide some reference on epidemic immunization in the future.

WSCN: Web Service Composition Based on Complex Networks

Currently the research of service science mainly focuses on the composition of distributed and heterogeneous Web services. Most of the approaches proposed to tackle this problem are based on cross-platform workflow and AI planning. However, until very recently, little attention is paid to the approaches based on complex networks. In the era of Big Data, Web services are created and updated in a fast way and their relationships are becoming more complicated. Therefore, studying the basic characteristics of services networks based on complex network theory will be meaningful. In this paper, we propose a framework to model the services network as a directed complex network, in which nodes represent the atomic Web services and the data communicated between them while edges represent the dependent relationship between data and service nodes. Based on the modeled network, we further apply graph search algorithms to solve the problem of Web services composition. Experimental results are given to validate the flexibility and efficiency of our framework.

Performance Optimization of Temporal Reasoning for Grid Workflows Using Relaxed Region Analysis

With quick evolution of grid technologies and increasing complexity of e-science applications, reasoning temporal properties of grid workflows to ensure reliability and trustworthiness is becoming a critical issue. Relaxed region analysis (RRA) is proposed in this work for performance optimization of grid workflow verification by decomposing workflows into separate standard regions with parallel branches. The approach is implemented in GridPiAnalyzer, a pi calculus based formal verifier for grid workflows, and validated using gravitational wave data analysis workflows. Detailed experimental results illustrate that RRA can dramatically reduce CPU and memory usage of verification processes.

Fast Autotuning Configurations of Parameters in Distributed Computing Systems Using Ordinal Optimization

Conventional autotuning configuration of parameters in distributed computing systems using evolutionary strategies increases integrated performance notably, though at the expense of consuming too much measurement time. An ordinal optimization (OO) based strategy is proposed in this work, combined with neural networks to improve system performance and reduce measurement time, which is fast enough to autotune configurations for distributed computing applications. The method is compared with a well known evolutionary algorithm called Covariance Matrix Algorithm (CMA). Experiments are carried out using high dimensional rastrigin functions, which show that OO can reduce one to two orders of magnitude of simulation time while at the cost of an acceptable scope of optimization performance. We also carried out experiments using a real application system with three-tier web servers. Experimental results show that OO can reduce 40% testing time on average at a reasonable and slight cost of optimization performance.

Performance improvement of distributed systems by autotuning of the configuration parameters

The performance of distributed computing systems is partially dependent on configuration parameters recorded in configuration files. Evolutionary strategies, with their ability to have a global view of the structural information, have been shown to effectively improve performance. However, most of these methods consume too much measurement time. This paper introduces an ordinal optimization based strategy combined with a back propagation neural network for autotuning of the configuration parameters. The strategy was first proposed in the automation community for complex manufacturing system optimization and is customized here for improving distributed system performance. The method is compared with the covariance matrix algorithm. Tests using a real distributed system with three-tier servers show that the strategy reduces the testing time by 40% on average at a reasonable performance cost.

Formal verification of temporal properties for reduced overhead in formal verification of temporal properties for reduced overhead in grid scientific workflows

With quick development of grid techniques and growing complexity of grid applications, it is becoming critical for reasoning temporal properties of grid workflows to probe potential pitfalls and errors, in order to ensure reliability and trustworthiness at the initial design phase. A state Pi calculus is proposed and implemented in this work, which not only enables flexible abstraction and management of historical grid system events, but also facilitates modeling and temporal verification of grid workflows. Furthermore, a relaxed region analysis (RRA) approach is proposed to decompose large scale grid workflows into sequentially composed regions with relaxation of parallel workflow branches, and corresponding verification strategies are also decomposed following modular verification principles. Performance evaluation results show that the RRA approach can dramatically reduce CPU time and memory usage of formal verification