Earticle

The resources in a grid are distributed in multiple autonomic domains. Hence, it’s always a time consuming process to perform the resource co-allocation in grids especially for parallel jobs due to the diversity of local domain policies and the dynamic resource workloads. Even worse, the resource competing among multiple parallel jobs may result in resource allocation deadlock. In this paper, a novel synchronized resource co-allocation model for cross-domain parallel jobs, called virtual job model (VJM), is proposed. VJM uses virtual jobs to co-reserve the earliest available resources for real parallel jobs with considering the local resource policies and the current resource availability in multiple domains. The co-reservation algorithm implemented in VJM is not only able to detect potential deadlocks, but able to relieve the resource competing via the resource reorganization mechanism. At last, we demonstrate that VJM is able to reduce the time cost of resource co-allocation significantly via experiments.

Multi-thread is a mature and widely-used programming mode for multitasking applications developing, especially in Java. However java.lang.Thread is not designed for HPC parallel programming. Java multi-thread program is confined within single compute host. CommonJ Work Manager is a joint specification proposed by IBM and BEA. Work Manager is a higherlevel alternative to using Java Thread. Originally, Work Manager is used to solve the constraint of opening independent threads in J2EE container. Although there are some enterprise servers such as WebLogic and WebSphere which have supported Work Manager, they are not aim at large scale parallel applications. They do not have a grid-enabled and expandable architecture, and the multi-thread programs are still executed on single compute host. Currently, grid-enabled parallel programming model like MPICH-G2 is limited with C/C++ implementations. There are no official bindings for MPI with Java. In this paper, we present a grid-enabled multi-thread programming model: SymWorkManager to enable multithread programs to run on distributed environments. SymWorkManager is designed as a high performance computing (HPC) implementation of CommonJ Work Manger specification using Platform Symphony. It provides J2EE developers a simpler way to program parallel multi-thread applications under a HPC platform.

Grid computing is becoming more and more attractive for coordinating large-scale heterogeneous resource sharing and problem solving. Although Globus toolkit version 4 which is developed based on WSRF can successfully deal with the heterogeneity of resources and the diversity of applications, the current grid environments based on Globus toolkit version 4 have several limitations regarding dynamic reconfiguration of virtual organization (VO) and remote hot deployment of grid services. This paper proposes WIVO (Web-based user Interface for Virtual Organization), the web-based user interface using web service. Using WIVO, user requested sub-VO can easily be configured on top of grid environment based on Globus toolkit version 4, and grid services can also be deployed remotely for better utilization of resources and applications. It is expected that the VO reconfiguration function as well as the remote grid service deployment function provided in WIVO can improve resource and application utilization.

This paper presents the design principle for integrating quality management on Web service registry developed in UDDI specification and Web service quality management system (WSQMS). WSQMS, developed by NIA 1 can measure and collect the quality information of Web services by its agency system installed on the Web service system. Web service registry is core system for registering and searching WSDL(Web Service Description Language). On the registry, there are no methods that users search Web services on the basis of quality information because the registry is operating without any relationship of WSQMS. So, it is required an integration method where WSQMS and registries share the quality data on the basis of classified quality data. For representing Web service quality information, we adopted the WSQDL(Web Service Quality Description Language), which published by WSQM technical committee in OASIS. In more detail, this paper also presents the scheme to compose the classification scheme for quality data and to modify the necessary data structure of the registry.

Regularization theory presents a sound framework to solving supervised learning problems. However, there is a gap between the theoretical results and practical suitability of regularization networks (RN). Radial basis function networks (RBF) that can be seen as a special case of regularization networks have a rich selection of learning algorithms. In this work we study a relationship between RN and RBF, and show that theoretical estimates for RN hold for a concrete RBF applied to real-world data, to a certain degree. This can provide several recommendations for strategies on choosing number of units in RBF network.