Hiroyuki Matsunaga

Data transfer over the wide area network with a large round trip time

A Tier-2 regional center is running at the University of Tokyo in Japan. This center receives a large amount of data of the ATLAS experiment from the Tier-1 center in France. Although the link between the two centers has 10Gbps bandwidth, it is not a dedicated link but is shared with other traffic, and the round trip time is 290ms. It is not easy to exploit the available bandwidth for such a link, so-called long fat network. We performed data transfer tests by using GridFTP in various combinations of the parameters, such as the number of parallel streams and the TCP window size. In addition, we have gained experience of the actual data transfer in our production system where the Disk Pool Manager (DPM) is used as the Storage Element and the data transfer is controlled by the File Transfer Service (FTS). We report results of the tests and the daily activity, and discuss the improvement of the data transfer throughput.

Performance of a disk storage system at a Tier-2 site

It is expected that a WLCG Tier-2 site plays a major role in the user analysis in addition to generating Monte Carlo simulation data. For the user analysis, the data storage system should be designed properly for sustaining massive data access from a large number of analysis jobs by the end users within the LAN, and also for the data transfer from the Tier-1 site over theWAN. The Tokyo Tier-2 centre provides 400 TB of disk and 1000 kSI2K of CPU to the ATLAS VO. As the data storage manager, the gLite DPM is deployed, which is common at the Tier-2 sites. In this talk, we present the architecture of our DPM storage system and its performance as well as operational experiences.

Grid Operation at Tokyo Tier-2 Centre for ATLAS

International Centre for Elementary Particle Physics, the University of Tokyo, has been involved in the Worldwide LHC Computing Grid since 2003. After extensive R&D experience of the PC computing farm, disk and tape storage systems, network technology and the integration of these components, it is now operating a regional centre for the ATLAS data analysis. The regional centre includes an ATLAS Tier-2 site which is running the gLite middleware developed by the Enabling Grids for E-sciencE (EGEE) project. One of the biggest challenges at the regional centre is efficient data transfer between the Tier-2 site in Tokyo and other sites, in particular the associated Tier-1 site in France, because the large round trip time due to the long distance makes it difficult to transfer data at a high rate. We have been studying to achieve a good performance of the data transfer, and some results of network tests and ATLAS data transfer are described. Hardware and software components and the operational experience are also reported in this article.

A local network integrated into a balloon-borne apparatus

Abstract A local network is incorporated into an apparatus for a balloon-borne experiment. A balloon-borne system implemented in the apparatus is composed of subsystems interconnected through a local network, which introduces modular architecture into the system. The network decomposes the balloon-borne system into subsystems, which are similarly structured from the point of view that the systems is kept under the control of a ground station. The subsystem is functionally self-contained and electrically independent. A computer is integrated into a subsystem, keeping the subsystem under the control. An independent group of batteries, being dedicated to a subsystem, supplies the whole electricity of the subsystem. The subsystem could be turned on and off independently of the other subsystems. So communication among the subsystems needs to be based on such a protocol that could guarantee the independence of the individual subsystems. An Omninet protocol is employed to network the subsystems. A ground station sends commands to the balloon-borne system. The command is received and executed at the system, then results of the execution are returned to the ground station. Various commands are available so that the system borne on a balloon could be controlled and monitored remotely from the ground station. A subsystem responds to a specific group of commands. A command is received by a transceiver subsystem and then transferred through the network to the subsystem to which the command is addressed. Then the subsystem executes the command and returns results to the transceiver subsystem, where the results are telemetered to the ground station. The network enhances independence of the individual subsystems, which enables programs of the individual subsystems to be coded independently. Independence facilitates development and debugging of programs, improving the quality of the system borne on a balloon.