Bin Gong

Specification and enforcement of flexible security policy for active cooperation

Interoperation and services sharing among different systems are becoming new paradigms for enterprise collaboration. To keep ahead in strong competition environments, an enterprise should provide flexible and comprehensive services to partners and support active collaborations with partners and customers. Achieving such goals requires enterprises to specify and enforce flexible security policies for their information systems. Although the area of access control has been widely investigated, current approaches still do not support flexible security policies able to account for different weighs that typically characterize the various attributes of the requesting parties and transactions and reflect the access control criteria that are relevant for the enterprise. In this paper we propose a novel approach that addresses such flexibility requirements while at the same time reducing the complexity of security management. To support flexible policy specification, we define the notion of restraint rules for authorization management processes and introduce the concept of impact weight for the conditions in these restraint rules. We also introduce a new data structure for the encoding of the condition tree as well as the corresponding algorithm for efficiently evaluating conditions. Furthermore, we present a system architecture that implements above approach and supports interoperation among heterogeneous platforms.

Resource Discovery Algorithm Based on Small-World Cluster in Hierarchical Grid Computing Environment

Grid is an environment of seamless and integrated sharing and collaboration. The heterogeneous, dynamic, various and autonomous characteristics of resources make grid resource discovery a challenging issue. Hierarchical grid computing environment (hierarchical grid for short) with many advantages is far and wide used. This paper introduces small-world cluster into hierarchical grid in which intra-cluster adopts centralized management and cluster center nodes form small-world network. The architecture strikes a balance between high efficiency of total centralized management and good scalability of absolute distributed disposal. In the process of constructing small-world network, a new construction method on the basis of NW (Newman-Watts) model is presented, that is, short range contacts and long range contacts are represented by routing tables in a logic and dynamic way not by Manhattan distance. SWRD (Small-World Resource Discovery) algorithm pertinent to user classification and data retrieval way is proposed. Results from simulation experiments demonstrate SWRD algorithm greatly prunes search space and improves query efficiency.

Cloud Model-Based Security-Aware and Fault-Tolerant Job Scheduling for Computing Grid

The uncertainties of grid nodes security are main hurdle to make the job scheduling secure, reliable and fault-tolerant. The fixed fault-tolerant strategy in jobs scheduling may utilize excessive resources. In this paper, the job scheduling decides which kinds of fault-tolerance strategy will be applied to each individual job for more reliable computation and shorter makespan. And we discuss the fuzziness or uncertainties between TL and SD attributes by the subjective judgment of human beings. Cloud model is a model of the uncertain transition between qualitative concept and its quantitative representation. Based on the cloud model, We propose a security-aware and fault-tolerant jobs scheduling strategy for grid (SAFT), which makes the assess of SD and SL to become more flexibile and more reliable. Meanwhile, the different fault-tolerant strategy has been applied in grid job scheduling algorithm by the SD and job workload. Moreover, much more important, we are able to set up some rules and active each qualitative rule to select a suitable fault-tolerant strategy for a scheduling job by input value (the SD and job workload) to realize the uncertainty reasoning. The results demonstrate that our algorithm has shorter makespan and more excellent efficiencies on improving the job failure rate than the fixed fault-tolerant strategy selection.

Parallel computing for option pricing based on the backward stochastic differential equation

The Backward Stochastic Differential Equation (BSDE) is a robust tool for financial derivatives pricing and risk management. In this paper, we explore the opportunity for parallel computing with BSDEs in financial engineering. A binomial tree based numerical method for BSDEs is investigated and applied to option pricing. According to the special structure of the numerical model, we develop a block allocation algorithm in parallelization, where large communication overhead is avoided. Runtime experiments manifest optimistic speedups for the parallel implementation.

Parallel Option Pricing with BSDE Method on GPU

The development of the hardware changes program structure. Now the Graphic Processing Unit (GPU) has evolved into an extremely flexible, powerful and cost-efficient processor, which is specialized for compute intensive, massively data parallel computation. In the field of financial derivatives pricing and risk management, the Backward Stochastic Differential Equation (BSDE) is a robust tool. The aim of this paper is the efficient use of GPU acceleration for option pricing with BSDEs. Experimental results show that a GPU can achieve a superior performance, greater than 230x, compared with the CPU-only case.

Energy-Optimization Scheduling of Task Dependent Graph on DVS-Enabled Cluster System

At present, power management in High-Performance Computing (HPC) environment is becoming a hot topic owning to its high operation cost, low reliability and environmental impact. In this paper, we investigate energy minimization scheduling algorithm of data dependent tasks in DVS-Enabled cluster system. Considering the data-intensive characteristics, the proposed EOTD (Energy Optimization scheduling for Task Dependent graph) algorithm adopts task clustering to reduce data transmission time and communication energy consumption. In order to decrease dynamic power of processing elements, it uses one of the power-saving techniques in system level—Dynamic Voltage Scaling while not violating the deadline users specify. Moreover, on the premise that application execution is predictive and exclusive for processing elements, we employ Dynamic Power Management and Binary Search technique to reduce the static power of idle processing elements and last find the optimal number of processing elements. EOTD algorithm not only optimizes the energy consumption of task dependent graph, but also satisfies the QoS requirements service level agreement gives. Compared with VM and LJ-VM algorithm, experimental results demonstrate that EOTD algorithm can achieve larger energy optimization in less optimizing time.

The design and implementation of Render Farm Manager based on OpenPBS

With the extensive use of the 3D computer graphics technology, the computation and the processing data are rapidly increased . Computer animation rendering is a long time process. In this paper, the animation rendering implementation based on OpenPBS is proposed. The OpenPBS-based animation rendering application provides the uniform rendering job management. Render Farm Manager is designed to provide OpenPBS a rendering module interface to manage the rendering job, and to propose GOS Portal a rendering job xml template for animators to submit the rendering job. Two Animation models are experimented and tested to show the efficiency of the OpenPBS-based animation rendering application.

Option Pricing on the GPU with Backward Stochastic Differential Equation

In this paper, we develop acceleration strategies for option pricing with non-linear Backward Stochastic Differential Equation (BSDE), which appears as a robust and valuable tool in financial markets. An efficient binomial lattice based method is adopted to solve the BSDE numerically. In order to reduce the global memory access frequency, the kernel invocation is avoided to be performed on each time step. Furthermore, for evaluating the affect of load balance to the performance, we provide two different acceleration strategies and compare them with running time experiments. The acceleration algorithms exhibit tremendous speedup over the sequential CPU implementation and therefore suitable for real-time application.

Marginal Pricing Based Scheduling Strategy of Scientific Workflow Using Cost-Gradient Metric

Service-oriented architecture builds software applications from large numbers of loosely coupled distributed services, while workflow technology coordinates these distributed services to achieve a shared task or goal. Scientific workflow is becoming researching hotspot along with the e-science and cyber-infrastructure applications, which run in the grid environment. Therefore, selection and scheduling of services in scientific workflow are complex and challenging. In utility grid, scientific application uses scheduling strategy to minimize its cost, while service provider adopts pricing schema to maximize its profit. Based on our previous work: pricing schema based on marginal principle (marginal pricing for short in this paper), we introduce a cost-gradient metric to provide criteria for service selection, and then propose a cost minimization scheduling strategy with specified deadline. Moreover, our scheduling environment is totally commercialized by scientific workflow scheduling market model. Results from experiments demonstrate our strategy can optimize usage cost of different types of workflows while satisfying QoS requirements.

Scheduling mobile collaborating workforce for multiple urgent events

Despite the advancement of wireless technologies that allows collaboration at different places, under emergencies, professionals are often still required to arrive at the scene to carry out critical tasks. Under many practical constraints, how to schedule mobile collaborating workforce for urgent event requirements becomes a challenging problem. In this paper, we study the optimal mobile workforce assignment problems for multiple events and propose an efficient algorithm to find an optimal workforce arrangement with respect to quick response under qualification and location constraints. A practical example is given to illustrate how our method works. We also study the exception case where there are not enough qualified users. We allow a user to take on multiple qualified tasks previously assigned to different users. But each person is restricted within one event location so as to reduce traffic transfer between different places for the quick response purpose. We analyze the computational complexity of the problem of finding an optimal assignment of mobile workforce under such restraints and solve it by means of integer linear programming.