Agent based Grid Computing

at

7th IEEE International Symposium on Cluster Computing and the Grid

(CCGrid 2007)

Rio de Janeiro, Brazil, 154-17 May  2007


http://ccgrid07.lncc.br/

Please register at the CCGrid 2007 website.

PROGRAM

ACCEPTED PAPERS

IMPORTANT DATES:

Submission Deadline (EXTENDED!):
15 December 2006  (See submission details)

Notification to authors: 
22 January 2007  (UPDATED)

Camera ready versions: 
5 February 2007 (hard deadline)

DESCRIPTION

Grid computing is an active research area which promises to provide a flexible infrastructure for complex, dynamic and distributed resource sharing and sophisticated problem solving environments. The Grid is not only a low level infrastructure for supporting computation, but can also facilitate and enable information and knowledge sharing at the higher semantic levels, to support knowledge integration and dissemination.

As Grid infrastructures mature and its complexity increases, deciding which services to use, where the data resides for a particular application domain, how to migrate the data to the point of computation (or vice versa), and performance rates required to maintain a particular application “behaviour” become a significant challenge. There are still issues to be tackled such as: semantically enhanced service descriptions and specification of resources, autonomy, dynamic service composition, collaboration, self-organisation and learning, intelligence and adaptability.

Some of these research areas have already been addressed within the multi-agent systems community. Agent and multi-agent technologies provide a promising approach to make Grid technologies smarter, more flexible, and adaptable. However, Grid computing environments and applications, with their inherent distribution, heterogeneity and multi-institutional setups, impose new challenges to agent-based approaches. Proposed solutions must be able to scale, be (self)organised from a diversity of agents with different roles and policies into dynamic Virtual Organizations (VOs), adapt to policies in place on a given VO, and be able to change roles or migrate between VOs to adapt to changing conditions and requirements. All of this while maintaining stringent levels of performance.

Therefore agents could play an important role in Grid computing, and Grid computing can offer useful test-beds for investigating agent related technologies. The aim of this workshop is to bring together infrastructure developers, applications developers, and researchers to inspire and encourage collaboration between these two communities.

The workshop is based on six similar events that have been organised over the last few years.

Five of these have been organised alongside CCGrid conferences in the past:

http://www.cs.cf.ac.uk/agc2006
http://www.iw.uni-karlsruhe.de/ccgrid/
http://users.cs.cf.ac.uk/O.F.Rana/agent-grid-2003/
http://users.cs.cf.ac.uk/O.F.Rana/agent-grid-2002/
http://users.cs.cf.ac.uk/O.F.Rana/agent-grid/

and one at last year Autonomous Agents and Multi-Agent Systems conference:
 http://www.iw.uni-karlsruhe.de/sgt05/

PUBLICATION

Workshop papers will be distributed to all CCGrid participants on a  CD published by IEEE Computer Society. The papers will also be  available electronically in the IEEE digital library.

TOPICS OF INTEREST

Authors are encouraged (but not restricted) to submit in the following areas:

  • Agent frameworks and infrastructure to support the Grid
  • Performance analysis/modelling of multi-agent based Grid applications and infrastructures 
  • Integrating Peer-2-Peer and mobile infrastructure with multi-agent based grid systems for pervasive gird environments
  • Software engineering methodologies and tools to support the development of multi-agent based gird systems
  • Support for managing and establishing virtual organizations using agents
  • Supporting service discovery, service composition and service management using agent-based approaches
  • Integrating ontologies with Grid services description for service discovery, composition and management
  • Agent based tools for scientific problem solving (ontologies, service composition, application monitoring)
  • Agent-based support  for high performance computing (load balancing, process scheduling, process migration, application optimization, communication optimization)
  • Agent-based autonomic systems
  • Agent-based monitoring, performance prediction and workload characterisation
  • Data migration and management via agents
  • Grid markets mechanisms and agent-based computational economics
  • Agent application in Grid simulation environments
  • Empirical results using agent-based simulation environments
  • Environment description and metric definitions for Grid simulations
  • Knowledge  generation and processing in complex decision situations using autonomous agents
  • Alternative approaches for managing Grid systems (such as nature-inspired approaches)

Experience reports and application demonstrators are particularly welcomed.
PROGRAM

Date: May 15, 2007,  10:30 - 12:30

10:30-10:45 - Introduction
10:45 - 11:15 - Paper: Token Exchange System as incentive mechanism for the e-Science Grid community
11:15 - 11:45 - Paper: RABC: A Pervasive Infrastructure for Browser Computing based on Ajax technologies
11:45 - 12:15 - Paper: Prototype an e-Science workflow bus using Jade agent framework
12:15 - 12:30 - Discussions
12:30 - END


ACCEPTED PAPERS

Token Exchange System as incentive mechanism for the e-Science Grid community
Arun Anandasivam and Dirk Neumann
Abstract: 
e-Science has a large demand for Grid resources. Research institutes of biology, chemistry or physics are interested in sharing their computational resources. Currently, Grid resources are not shared between institutes of different research areas as they do not know each other. A high communication overhead at a political level is required to start and maintain cooperation. In this paper we propose a mechanism called Token Exchange System (TES). Our aim is to facilitate exchange of resources among the different research institutes without any face-to-face agreement. Furthermore, the mechanism has to set the right incentives so that members will share their resources fairly and free-riders are avoided. The approach presented here is the combination of reputation mechanism and the payment mechanism in one coherent system. Since the e-Science community abhors real money as an exchange mechanism, this approach provides a solution to set incentives like payment mechanisms and to implement the advantages of reputation mechanisms.

RABC: A Pervasive Infrastructure for Browser Computing based on Ajax technologies
Fumikazu KONISHI, manabu ishii and shingo ohki.
Abstract:  We have developed the RIKEN Ajax Browser Computing (RABC) system. This system enables the simple realization of large scale Global Computing systems,like the SETI@home project, that take advantage of unused computing resources. Its greatest advantage is that an agent program, which has been traditionally required, does not need to be installed. Volunteers can participate by simply browsing to a webpage provided by the RABC system. This system allows greater CPU cycle resources to be acquired than traditional methods. We have evaluated a performance difference between JavaScript and legacy computer languages.

Prototype an e-Science workflow bus using Jade agent framework
Zhiming Zhao, Adam Belloum and Cees de Laat
Abstract:  The development of an effective Scientific Workflow Management System (SWMS) requires profound knowledge on both application domains and network programming, and it is often time consuming. Most of the existing SWMSs are driven by applications from specific domains and are developed in academic projects. It is challenging to introduce an existing SWMS to a new domain; not only the workflow model and description language do not easily fit in new problem domains, but also the unstable development state of existing systems does not provide all functionality required by the new applications and thus gives high risk for the development. Integrating domain specific SWMS improves reusability of workflow resources and promotes a generic framework for different e-Science domains. In this paper, we discuss an agent based approach to realize a bus like runtime infrastructure for interconnecting different workflow engines.

WORKSHOP CHAIRS/ORGANISERS
Pablo Chacin
Universitat Politècnica de Catalunya, Departament d'Arquitectura de Computadors, UPC Campus Nord, Despatx D6-212, Jordi Girona, 1-3, 08034 Barcelona, Spain
Tel: +34 (0) 93 4011055
Fax: +34 (0) 93 4017055
email: pchacin@ac.upc.edu
Liviu Joita
Cardiff University, School of Computer Science and Welsh e-Science Centre, Queen's Buildings, 5 The Parade, Cardiff CF24 3AA, UK
Tel: +44 (0) 29 20874000 (ext. 77399)
Fax: +44 (0) 29 20874598
email: L.Joita@cs.cardiff.ac.uk
Isaac Chao
Universitat Politècnica de Catalunya, Departament d'Arquitectura de Computadors,
UPC Campus Nord, Despatx C6-002, Jordi Girona, 1-3,
08034 Barcelona, Spain
Tel: +34 (0) 93 4011055
Fax: +34 (0) 93 4017055
email: ichao@lsi.upc.edu

STEERING COMMITTEE
 Torsten Eymann,
University of Bayreuth, Germany
Omer F. Rana,
Cardiff University, UK
Daniel Veit,
University of Karlsruhe, Germany

PROGRAMME COMMITTEE (
PRELIMINARY)
Mark Baker, University of Reading, UK

Bernhard Bauer, University of Augsburg, Germany

Rajkumar Buyya, The University of Melbourne, Australia

Felix Freitag, University of Catalonia, Spain

Gianfranco Giulioni, Universita Politecnica delle Marche Ancona, Italy

Sverker Janson, SICS, Sweden

Heiko Ludwig, IBM Research, USA

Beniamino Di Martino, Second University of Naples, Italy

Dirk Neumann, University of Karlsruhe, Germany

Bruno Schulze, National Laboratory for Scientific Computing - LNCC, Brazil
Line Catherine Pouchard, Oak Ridge National Laboratory, USA

Giselher Pankratz, The FernUniversitaet Hagen, Germany

Fehti Rabhi, University of New South Wales, Australia

Simon Chong-Wee See, Sun Microsystem Inc., Singapore

Jan Staellaert, University of Connecticut, USA

York Sure, University of Karlsruhe, Germany

Niranjan Suri, Institute for Human & Machine Cognition – IHMC, USA

Floriano Zini, ITC-irst, Trento, Italy

Geoffrey Fox,  University of Indiana, USA
Authors are encouraged (but not restricted) to submit in the following areas
    Agent frameworks and infrastruc

SUBMISSION DETAILS

For submissions, please go to the web page http://www.easychair.org/AGC2007  and register as an author. Once registered, follow the "new submission"  option in the main page's menu to submit your paper, including all the authors, the title, the abstract and the paper's document. Please notice that, by default, the first author will be the corresponding author and will receive all notifications, but  you can select other author(s).

Authors should submit papers of not more than 6 pages of double column text using single spaced 10 point size type on 8.5 x 11 inch pages, as per IEEE 8.5 x 11 manuscript guidelines (see http://www.computer.org/portal/pages/cscps/cps/cps_forms.html for instructions).

  • All papers must be original work authorized to be released.

  • No submission of material that has been previously published.

  • Papers must be written in English language.

Authors should submit a PostScript (level 2) or PDF file that will print on a PostScript printer. Please do not embed specialist fonts (such as Asian fonts) in the PDF. Hard copies should be sent only if electronic submission is not possible. Submission implies the willingness of at least one of the authors to register to the conference and workshop and present the paper. 

In addition, authors should submit an ASCII abstract, with the following information: title of paper; names and affiliations of authors; name, email, snail mail, phone number, and fax number of primary contact. The same information should be included on the first page of submitted papers.

Submitted papers will be reviewed by the program committee. The papers are intended to be reviewed by at least 3 reviewers.