Kohei Ichikawa

Building cyberinfrastructure for bioinformatics using service oriented architecture

Cyberinfrastructure makes the development and deployment of bioinformatics applications easier by providing the framework and components that may be loosely coupled using service oriented architecture. Here we describe an end to end prototype environment that allows existing applications to run on the grid, taking advantage of open source software that provides a portal interface using GridSphere, with transparent GSI authentication using GAMA, a Web service wrapper using Opal, a metascheduler using CSF4, a virtual file system using Gfarm, and a grid-enabled cluster environment using Rocks. Solutions to complex problems may be developed using workflow tools such as Kepler that coordinate different interoperable services. The availability of this type of cyberinfrastructure suggests that new applications should be designed with the grid in mind, using service oriented architecture for interoperability and efficiency. This approach may enable bioinformaticians to focus on their problems of interest, and make use of the emerging cyberinfrastructure through virtually drag and drop deployment.

OPAL OP: AN EXTENSIBLE GRID-ENABLING WRAPPING TOOL FOR LEGACY APPLICATIONS

This paper describes a new approach (Opal OP: Opal Operation Provider) to wrap existing legacy applications as Grid services. In order to expose, with minimal effort, existing applications as Grid services, Opal OP provides a method for wrapping a legacy application as a program module, or as an operation provider. Traditional wrapping methods usually restrict the way to implement Grid services because these methods provide only a suite of interfaces necessary for using the wrapped application. The proposed Opal OP, on the other hand, doesn’t restrict the way to implement a Grid service from a legacy application. Opal OP is implemented as an operation provider that wraps the application and thus can be used as a module in the Grid service. Application developers can easily develop their own services where legacy applications are wrapped through the utilization of Opal OP. In this paper, we show some scientific applications including a bio-molecular simulation system developed as Grid services using Opal OP. The results show the usefulness and effectiveness of Opal OP.

A Prototype of Network Failure Avoidance Functionality for SAGE Using OpenFlow

A tiled display wall (TDW), which is a single large display device composed of multiple sets of computers and displays, has recently gained the attention of scientists. In particular, SAGE, which is a middleware for building TDW, allows scientists to browse a multiple series of visualized results of computer simulation and analysis through the use of network streaming. Each visualized result can be generated on a different remote computer. For this reason SAGE has been increasingly hailed as a promising visualization technology that will solve the geographical distribution problem of computational and data resources. SAGE depends heavily on a network streaming technique in its architecture, but does not have any recovery mechanism against cases of network problems. This research, therefore, aims at realizing a network failure avoidance functionality for SAGE, focusing on OpenFlow against SAGE vulnerability to network failure. Specifically, the functionality is designed and developed as a composition of the following three functions: network failure detection, network topology understanding, and a packet forwarding control function. The key concept behind our design is that the network control function from OpenFlow should be built into SAGE. The evaluation in the paper confirms that the proposed and prototyped network failure avoidance functionality can detect failures on network routes and then reroute network streaming for visualization on TDW.

Development of Data Mining Platform MUSASHI Towards Service Computing

The objective of this paper is to introduce further development of a data mining tool, MUSASHI (Mining Utilities and System Architecture for Scalable processing of HIstorical data), for service computing. Recent advances in information systems have allowed us to gather enormous amounts of data on marketing. However, these gathered data have been individually stored at each company, and have never been integrated because of a lack of techniques to analyze the data in an integrated way and to handle the large amount of data efficiently. To address this issue, we are currently investigating a way to provide a data mining platform as a service so that users can apply various data mining techniques to their marketing data with ease and at a low cost. For this purpose, we have developed an ASP platform leveraging distributed computing technology represented by Cloud computing. This paper describes the ASP platform for data mining services and introduces an empirical application of data mining using our platform.

A Built-in Application Control Module for SAGE

Recently, the amount of scientific data has been increasing because of technological advancement in measurement and high-performance computing. As the result, visualization technologies are becoming increasingly important to analyze the scientific data. Among visualization technologies, SAGE is gathering attention of scientists in these days. It not only enables the visualization of tens of million pixel-class high-resolution image but also has the potential to realize the collaboration environment where scientists can work together for their common research purpose. However, there remain still technical issues in SAGE, since it is still in the process of development. A problem of application operability, which prevents SAGE users from efficiently operating various applications on SAGE, is targeted in this paper. This problem is caused from the architectural design that SAGE users have to manage window-related operation and application-related operation from different interfaces. For solving the problem, we have developed a built-in application control module for SAGE. Our built-in application control module unifies the separate interfaces for window management and application operation. As a result, users can control application from a unified user interface, SAGE UI.

Social Network Rebuilder: A Tool to Estimate a Social Network of Financial Crisis Propagation

The objective of this paper is to find a new method to estimate real social networks based on observed data collected by questionnaire surveys. Studies on social networks have been increasing in order to analyze social phenomena from a micro viewpoint. Most social phenomena can be explained by micro-level interactions among people. Spread of rumor and pandemics are typical example of micro interaction? However, there has not been much work on an analysis of real social networks based on observed data. This study tries to establish a methodology that exploits a genetic algorithm to rebuild a social network based on the data observed indirectly from real social networks. This paper introduces our proposed method, which allows us to rebuild a social network to some extent from degree distributions of a target real social network.

A P2P-based virtual cluster computing using PIAX

Providing large scale, stable and robust computing environments is indispensable to various fields including scientific researches as well as business. Grid computing has gathered considerable attention as a means to efficiently and effectively utilize computational resources among multiple organizations. However, the current Grid technologies do not always provide the computational environment enough to satisfy users requirements because of the heterogeneity and disconnect among computational resources on the Grid. For the reason, we have been attempting to develop a new middleware that leverages newly emerging Cloud technologies and allows for a virtual cluster composed of virtual computational resources by Xen as a virtualization technology on top of an overlay network realized by PIAX as an overlay networking technology. In this paper, we propose a system that allows Xen-based virtual computational resources composing a virtual cluster to communicate with each other toward the future development of Grid and Cloud middleware technology. The evaluation shown in the paper indicates that such a virtual cluster is feasible and could work effectively for computationally intensive applications.

A Proposal of P2P-Based Resource Discovery Mechanism for the Grid

Resource discovery is a fundamental service of the Grid to find out resources suitable for executing user jobs. Traditional resource discovery services of the Grid have static structure and centralized or hierarchized, such as Globus MDS. However, with the number of resources composing the Grid increased, this approach has two problems, namely, on dynamic changes of the structure in the Grid and obtaining the latest resource information. To solve these problems, we have been seeking for the possibility of a resource discovery mechanism adopting Peer-to-Peer approach, which is characterized by real-time retrieval of information using an overlay network spanned on physical computational resources. In this paper, we present the preliminary results of our research activities so far.

A Communication Relay Mechanism toward Virtual Cluster Computing on the Grid

Recently, Grid computing has gathered the attention as a means to efficiently and effectively utilize computational resources among multiple organizations. However, the current Grid technologies do not always provide the computational environment enough to satisfy users requirements because of the heterogeneity and disconnect among computational resources on the Grid. For the reason, we have been attempting to develop a new Grid middleware technology that allows for a virtual cluster composed of virtual computational resources by Xen as a virtualization technology on top of an overlay network realized by PIAX as an overlay networking technology. In this paper, we propose a communication relay mechanism that allows Xen-based virtual computational resources composing a virtual cluster to communicate with each other toward the future development of such a Grid middleware technology. The evaluation shown in the paper indicates that such a virtual cluster is feasible and could work effectively for computationally-intensive applications.

A Framework for Meta-Scheduling WSRF-based Services

The maturity and stability of WSRF and its related technologies is producing a new demand for meta-scheduling WSRF-based services. In fact, many scientific institutions and universities are looking into the possibility of using WSRF-based services as production-level services. For this purpose, they need a meta-scheduling solution that selects one of the WSRF-based services, each of which is deployed in an organization for fault tolerance and load balancing. However, methods for meta-scheduling services among multiple sites have not been discussed well. This paper describes a new framework (MSSA: Meta-Scheduling Service Architecture) for scheduling and selecting WSRF-based services deployed over multiple sites. This framework provides a transparent scheduling method to users. Our method takes advantage of the factory pattern used typical WSRF services. The framework builds scheduling processes into the factory pattern and hides the complexity of the scheduling process from users. MSSA defines a meta-factory service that selects a suitable factory service and delegates the users request to the factory service. By using this meta-factory service instead of using the original factory services, users can utilize meta-scheduling features transparently. MSSA can build hierarchical structures of meta-factory services to construct scalable scheduling environment, and can also switch scheduling policies easily. In this paper, the design and implementation of MSSA is discussed.