Introduction to the 2nd International Workshop on Middleware for Grid Computing (MGC)

As network performance has outpaced computational power andstorage capacity, a new paradigm has evolved to enable the sharingand coordinated use of geographically distributed resources,popularly known as "Grid" computing. The aims are to coupledistributed resources and offer consistent and inexpensive accessto resources irrespective of their physical location. Gridcomputing provides scalable and secure remote access to computing,data, and other resources throughout the Internet. Thesetechnologies enable the clustering of a wide variety ofgeographically distributed resources, i.e., supercomputers, storagesystems, data sources, and special devices and services, that canthen be used as a unified resource. Much of the focus of Grid until recently has been withinresearch organizations. Currently the involvement of commercialorganization has increased tremendously. That leads to a greaterfocus on integration with commercial computing models and more workon the security and resource management models required forautomated and secure access negotiation for remote resources. Inorder to control such a broad base of resources, Grid Computing hasa middleware layer controlling the distributed execution ofapplications. Under the initial approach based on dedicatedresources or opportunistic resources, were basically related toconstraints on budget and complexity. However by exploring thepossibility of a wider range of applications specially based ondata grids and content networks, a larger diversity of resourcesand devices can be thought participating in a grid system. The first edition of the MGC workshop was held on June 2003, inconjunction with the Middleware Conference, in Rio de Janeiro,Brazil. MGC'2003 saw a good combination of the many differentflavors in Middleware for Grid Computing, with 16 technicalpresentations and discussions on several stimulating topicsincluding, among others, Classic Grids, object orientedtechnologies, Service-based Grids, Open grid Service Architecture,Agent Grid, Interactive Grid, Grid Economy / Scheduling, andPortlets. After the workshop, extended and thoroughly revisedversions of the papers were invited to a Special Issue ofConcurrency and Computation: Practice and Experience Journalpublished in 2004. MGC'2003 generated substantial interest in thecommunity and we hope to build on this tradition this year. Farfrom exhausting the topics of interest, it paved the way for asecond edition of the workshop. The goal of the second edition of the MGC workshop is to bringtogether researches in the field of middleware, addressinglargescale and real world problems in grid environments, includingthe most interesting and stimulating topics emerged last year, andsome novel ones as Strategies and Protocols for obtaining Qualityof Services, Virtualization, Wireless Grids, Data Grid Middleware,Semantic Grid Middleware, Dependability and Fault Tolerance in GridMiddleware and Managing Information. The second workshop received an unexpectedly high number ofquality submissions and fifteen papers were chosen for theseproceedings out of the forty eight originally submitted.Additionally nine posters have been invited. The presentershighlight issues and solutions in one or more of the themesidentified for the workshop. As full papers there were contributions addressing differenttopics and common issues related to Resource Management andScheduling by Chen and Agrawal, Talwar et al., Andreozzi et al.,Araki, and Venugopal, Buyya and Winton. They focus on streamingdata, remote desktop sessions in utility grids, monitoringconnectivity in grids, autonomic www servers, and data-orientedapplications. Addressing Security there were papers by Dodonov,Guardia and Sousa, and Detsch et al. They examine security in thecontext of securing hosts from malicious greedy applications and inpeer-to-peer base grid computing. Addressing aspects on Strategiesand Protocols for QoS there was a paper by Z Cai et al.,investigating a network aware middleware for interactive large-dataapplications. Examining issues related to Data Grid Middleware andServices there were papers by Fontes et al., Kosar, Kola and Livny,and Aloisio et al. They examine a data and program integrationservice, data pipelines enabling large scale multiprotocol datatransfers, and advanced delivery mechanisms. On ManagingInformation there was a paper by Zhao et al., presenting a gridmiddleware service for virtual data discovery, composition andintegration. On Programming Models, Tools and Environments, therewere contributions by Vianna et al., Cicerre, Madeira and Buzato,and Camargo et al. They explore hybrid scheduling algorithms,hierarchical process execution, and checkpointing based rollbackrecovery. Additionally there were poster contributions by Coulson et al.,Ziviani et al., and Araújo, Cirne and Mendes on ResourceManagement and Scheduling, focusing on a management gridkit,calculations and usage of network measurements, and hidingresources behind brokers. Nassif et al., andAssunção, Koch and Westphall present posters oncommons issues on Agent based Grid Middleware, using agent-basednegotiation and resource allocation, and grid-based network /systems management. Cecchet, Quema and Boutaleb present a poster onresource-driven component deployment in the topics of Classic andData Grid Middleware and Services. Biswas and Pal present a posteron Programming Models, Tools, and Environments, using a dataflowapproach for grid computing. Finally, there were contributions bySchaeffer Filho et al., and Yamolenko et al. on Evolution of andExperiences with Grid systems, describing a practical usage insequence alignment and pi-calculations. We would like to thank the authors and the participants forpresenting their work and contributions on research in middlewarefor grid computing. Additionally we would like to thank all thereviewers for providing constructive reviews and specially thankall members of the program committee for helping to shape theworkshop program. Bruno Schulze, Distributed Scientific Computing Group, National Lab for Scientific Computing LNCC, BRRadha Nandkumar, National Center for Supercomputing Applications NCSA, University of Illinois Urbana-Champaign UIUC, US