Manabu Higashida

A History-Based Job Scheduling Mechanism for the Vector Computing Cloud

The wide-area vector meta computing infrastructure named a vector computing cloud has been proposed as a next generation high-performance computing infrastructure. However, in the vector computing cloud, the difference in site policies between organizations causes inefficient usage of vector computing resources. To achieve fairness and efficient job scheduling on the vector computing cloud, this paper presents a history-based job scheduling mechanism for a queue system. The proposed mechanism estimates the time to start the job execution in a queue system from the history of job-execution on vector supercomputers. Based on the estimation, the job scheduling mechanism automatically allocates the job to an appropriate site, which can execute the job earlier. The simulation results show that the proposed job scheduling mechanism improves the utilization efficiency of vector computing resources, compared to the conventional round-robin scheduling mechanism. In addition, the experiment using a prototype of the vector computing cloud indicates that the proposed job scheduling mechanism has enough potential for transparently executing jobs between the two SX-9 systems.

An Application of the NAREGI Grid Middleware to a Nationwide Joint-Use Environment for Computing

In the Cybermedia Center, Osaka University (CMC), we have applied the NAREGI grid middleware to our latest large-scale computing system for nationwide joint-use, and have made grid computing technology available on them coexisting with the conventional computing style. In this paper, we mention the problems of resource provisioning, password/passphrase management, issuance of grid certificate and grid operations in user management, and a solution is discussed.

Designing of a Network-Aware Cloud Robotic Sensor Observation Framework

This paper proposes a network-aware cloud networked robotic sensor observation framework based on the topic-based publisher/subscriber messaging paradigm. Cloud network robotics systems have some advantages in making use of rich computation or storage resources of clouds, low maintenance cost, and easy development of applications. But a problem occurs when systems that need fast feedback from application modules are run over networks with intermittent connection. In this framework, the placement of the robot application modules is changed according to the QoS provided by the physical network. We also propose a map cell-based robotic sensor observation algorithm that suits to the framework. A prototype system of a data center monitoring robot that is based on the proposed framework and algorithm is implemented and a preliminary experiment was carried out to demonstrate the benefit of the system. The experiment result shows that the performance of the data center monitoring could be improved and hot spots can be found faster by making the cloud networked robot network-aware.

An Authentication System Development in Application of Grid Middleware to a Nationwide Joint-Use Environment for Computing

In the Cybermedia Center, Osaka University (CMC), we have applied the NAREGI grid middleware to our latest large-scale computing system for nationwide joint-use, and have made grid computing technology available on them coexisting with the conventional computing style. In this paper, we will describe an authentication system solving the problems which we have encountered in construction of the grid environment.

The Complementary Registration Method for Vocal Tract of Sibilant /s/

We analyzed the sibilant /s/, an unvoiced consonant, by considering turbulence and vortex sound. The morphology of an oral tract is required when doing numerical analyses, and because magnetic resonance imaging (MRI) and cone-beam computed tomography (CBCT) are complementary, we use them as the registration technique in our proposed method. Calculating the approximate function to represent the morphological features of these geometries enabled us to associate the MRI and CBCT datasets with each other by referring to the functions in our proposal method. We applied the method to an actual person’s 4-dimensional MRI and CBCT datasets, and as a result were able to determine the appropriate registration result. The result was validated in a special configuration in which the result image was compared with the CBCT image of an actual person’s oral tract while pronouncing sibilant /s/. The validation demonstrates how our proposed method accurately registers the MRI and CBCT datasets and can thereby be used to perform numerical analyses in speech science research.

The Grid Middleware on SX and Its Operation for Nation-Wide Service

The Cybermedia Center of Osaka University runs a large-scale, high-performance computing system that combines a vector supercomputer with PC clusters. In order that all registered users can equally access this system from a grid environment, the center has been looking at using NAREGI grid middleware. We enhanced the virtual machine feature of the middleware-known as the GridVM (grid virtual machine)-and researched how to effectively supply the resources of the NEC SX Series vector supercomputer to the grid environment.

The Vector Computing Cloud: Toward a Vector Meta-Computing Environment

Aiming at realizing a high performance computing (HPC) cloud with vector supercomputers, this paper presents the world’s first prototype of wide-area vector meta computing infrastructure named a vector computing cloud by virtualizing remote vector computing resources as an HPC service over the Internet. The prototype system consists of two remote NEC SX-9 nodes connected through a long fat-pipe network (LFN), and each node is located at Tohoku University and Osaka University with a distance of 800 km. The vector computing cloud also provides a single sign on environment and jobs are automatically assigned to appropriate sites. Wide-area co-operation of distributed vector supercomputers is realized by adopting the NAREGI Grid Middleware, and a virtual machine for NEC SX vector supercomputer series, job scheduling algorithms, and an MPI operating environment are newly developed to enhance the job and resource management capabilities of the NAREGI Grid Middleware. In addition, to achieve fairness and efficient job scheduling on the vector computing cloud, this paper presents a history-based job scheduling mechanism for a queue system. Based on the estimation, the job scheduling mechanism automatically allocates the job to an appropriate site, which can execute the job earlier. The operation tests and experiment results indicate that the prototype system realizes single sign on for multiple vector resources and has enough potential for transparently operating jobs between the two SX-9 systems. This paper also evaluates and discusses the performance of the proposed job scheduling mechanism and MPI operation between both SX-9 systems using the High Performance Linpack (HPL) benchmark.