Search

We demonstrate high-speed data transfer GridFTP using a multipath control mechanism with SDN (Software-Defined Networking). GridFTP is a typical tool that has been developed and widely used for bulk data transfer over a wide area network in the field. GridFTP supports a parallel high-speed data transfer scheme using multiple TCP streams. However, one of the shortest paths is used solely for data transfer in the default IP routing while there are multiple network paths (multipath) exist between widely-distributed sites. In this study, we propose a system that distributes the parallel TCP streams of GridFTP into multiple network paths by a traffic engineering technique brought by SDN. Our proposed system has achieved approximately 20% better performance than the conventional method in the best case in a global-scale real enviroment.

A large amount of scientific data needs to be transferred from one site to another as fast as possible in the computational science fields. High-speed data transfer between sites is very important, especially in the Grid computing field, GridFTP has been widely used for bulk data transfer over a wide area network. GridFTP achieves greater performance by supporting parallel TCP streams. Using parallel TCP streams improves the throughput of slow-start algorithms and lossy networks even on a single path. This research proposes a traffic engineering technique that increases the data transfer performance by using multiple paths simultaneously for the parallel TCP streams. For this purpose, we use Software-Defined Network (SDN) technology and its implementation, OpenFlow. This paper presents the design and implementation of the proposed system. Our performance evaluation demonstrates that our proposed system can accelerate GridFTP Transfer in both virtual and real global-scale environments.

In grid computing, a key issue is how limited network resources can be shared by communications by various applications more effectively in order to improve application-level performance, e.g., by reducing the completion time for an individual application and/or set of applications. Communication by an application changes the condition of the network resources, which may, in turn, affect communications by other applications, and thus may degrade their performance. In this paper, we examine the characteristics of traffic generated by typical grid applications, and the effect of the round-trip time and bottleneck bandwidth on the application-level performance (i.e., completion time) of these applications. Our experiments showed that the impact of network conditions on the performance of various applications and the impact of application traffic on network conditions differed considerably depending on the application. These results suggest that effective allocation of network resources must take into account the network-related properties of individual applications.

This paper describes the £ 120M UK 'e-Science' (http://www.research-councils.ac.uk/and http://www.escience-grid.org.uk) initiative and begins by defining what is meant by the term e-Science. The majority of the £ 120M, some £ 75M, is funding large-scale e-Science pilot projects in many areas of science and engineering. The infrastructure needed to support such projects must permit routine sharing of distributed and heterogeneous computational and data resources as well as supporting effective collaboration between groups of scientists. Such an infrastructure is commonly referred to as the Grid. Apart from £ 10M towards a Teraflop computer, the remaining funds, some £ 35M, constitute the e-Science 'Core Programme'. The goal of this Core Programme is to advance the development of robust and generic Grid middleware in collaboration with industry. The key elements of the Core Programme will be outlined including details of a UK e-Science Grid testbed. The pilot e-Science projects that have so far been announced are then briefly described. These projects span a range of disciplines from particle physics and astronomy to engineering and healthcare, and illustrate the breadth of the UK e-Science Programme. In addition to these major e-Science projects, the Core Programme is funding a series of short-term e-Science demonstrators across a number of disciplines as well as projects in network traffic engineering and some international collaborative activities. We conclude with some remarks about the need to develop a data architecture for the Grid that will allow federated access to relational databases as well as flat files.

The StarPlane project enables users to dynamically control network photonic paths. Applications running on the Distributed ASCI Supercomputer (DAS-3) can manipulate wavelengths in the Dutch research and education network SURFnet6. The goal is to achieve fast switching times so that when the computational pattern in the computing clusters changes, the underlying network topology adapts rapidly to the new traffic flow. StarPlane develops: the software to perform optimal traffic engineering; the management plane to map the users’ request into directives for the network control plane, as well as the integration to the DAS-3 Grid middleware. We present here some preliminary results obtained with few selected Grid applications that make use of StarPlane.

This paper introduces a procedure called traffic engineering for grids for enabling grid networks to self-adjust to resource availability. The proposal is based on monitoring the state of resources and on task migration. It involves several layers of the Internet architecture. Experiments executed in NS-2 are used to illustrate the efficacy of the procedure proposed.

Multi-domain resource sharing is the fundamental feature of computing and data grid applications. In this paper, we address the issue of multi-domain optical network resource sharing and present a network management system designed for end-to-end lightpath provisioning across multiple independently managed domains. Our network management system, called user-controlled lightpath provisioning (UCLP) system, is based on the grid services concept and built on the Jini and JavaSpaces technologies. The UCLP system provides a traffic engineering approach to control network infrastructure for computing and data grid applications.

We describe the implementation of a transport simulation in a parallel environment. The implementation is based on a graphical parallel programming environment called P-GRADE. The transport simulator, called MadCity, simulates a specific road network of a city and shows cars moving on the roads. To achieve scalability of the traffic simulation, the use of templates is necessary. This helps to control the number of participating processes required for the simulation without making modifications to the simulator's source code. Performance results are collected from four, eight and sixteen nodes of the Parsifal cluster and compared with the sequential execution results of the simulator. The implementation of the transport simulator is extended further to support the simulation of multiple cities within the same cluster and on the Grid.