How to bring together fault tolerance and data consistency to enable Grid data sharing

High Performance Computing – HiPC 2007

Caron

Desprez

et al. 2007

Abstract:As grids become more and more attractive for solving complex problems with high computational and storage requirements, the need for adequate grid programming models is considerable. To this purpose, the GridRPC model has been proposed as a grid version of the classical RPC paradigm, with the goal to build NES (Network-Enabled Server) environments. Paradoxically enough, in this model, data management has not been defined and is now explicitly left at the user's charge. The contribution of this paper is to enhance data management in NES by introducing a transparent data access model, available through the concept of grid data-sharing service. Data management (persistent storage, transfer, consistent replication) is totally delegated to the service, whereas the applications simply access shared data via global identifiers. We illustrate our approach using the DIET GridRPC middleware and the JUXMEM data-sharing service. Experiments performed on the Grid'5000 testbed demonstrate the benefits of the proposed approach.

Section: The Concept Of Data-sharing Servicementioning

confidence: 99%

Section: Towards a Transparent Data Access Model For The Gridrpc Paramentioning

confidence: 99%

Towards a Transparent Data Access Model for the GridRPC Paradigm

High Performance Computing – HiPC 2007

Caron

Desprez

et al. 2007

“…In order to support these events, the data sharing service relies on replication techniques and failure detection mechanisms in order to enhance data availability despite disconnections and failures [8]. These aspects are beyond the scope of this paper.…”

Section: Fault Tolerancementioning

confidence: 99%

Enabling Transparent Data Sharing in Component Models

Sixth IEEE International Symposium on Cluster Computing and the Grid (CCGRID'06)

Bouziane

Breuil

et al. 2006

Self Cite

Abstract:The fast growth of high-bandwidth wide-area networks has encouraged the development of computational grids. To deal with the increasing complexity of grid applications, the software component technology seems very appealing since it emphasizes software composition and re-use. However, current software component models only support explicit data transfers between components through remote procedure call (RPC), remote method invocation (RMI) or event based ports. The distributed shared memory paradigm has demonstrated its utility by enabling a transparent access to data via a globally shared data space. Data localization, replication and transfer as well as synchronization of concurrent accesses are delegated to an external data-sharing service. This paper proposes to extend software component models with shared memory capabilities: a) a transparent access to shared data, and b) the possibility to use shared data as parameters of operations provided by component ports. The proposed model is instantiated as an extension of the CORBA Component Model (CCM) and the Common Component Architecture (CCA), using the transparent data access model provided by the JUXMEM grid data-sharing service. Key

“…As the cost of disk read/write operations may significantly limit the performance of data accesses, the use of faster-access RAM storage appears as a promising approach. The concept of grid-data sharing service [1,2] explores this idea by providing the abstraction of a globally shared memory space, built by aggregating the RAM storage made available by thousands of grid nodes. However, the overall storage capacity is limited to the aggregated RAM storage available.…”

Section: Introductionmentioning

confidence: 99%

Building Hierarchical Grid Storage Using the Gfarm Global File System and the JuxMem Grid Data-Sharing Service

Lecture Notes in Computer Science

Cudennec

Ghareeb

et al. 2008

Abstract. As more and more large-scale applications need to generate and process very large volumes of data, the need for adequate storage facilities is growing. It becomes crucial to efficiently and reliably store and retrieve large sets of data that may be shared at the global scale. Based on previous systems for global data sharing (global file systems, grid data-sharing services), this paper proposes a hierarchical approach for grid storage, which combines the access efficiency of RAM storage with the scalability and persistence of the global file system approach. Our proposal has been validated through a prototype that couples the GFARM file system with the JUXMEM data-sharing service. Experiments on the Grid'5000 testbed confirm the advantages of our approach.