Peer-to-peer-based resource discovery in global grids: a tutorial

Abstract

he last few years have seen the emergence of a new generation of distributed systems that scale over the Internet, operate under decentralized settings, and are dynamic in their behavior (participants can leave or join the system). One such system is referred to as grid computing; other similar systems include peer-to-peer (P2P) computing [1], semantic Web [2], pervasive computing [3], and mobile computing [4, 5]. Grid computing [6] provides the basic infrastructure required for sharing diverse sets of resources, including desktops, computational clusters, supercomputers, storage, data, sensors, applications, and online scientific instruments. Grid computing offers its vast computational power to solve grand challenge problems in science and engineering such as protein folding, high energy physics, financial modeling, earthquake simulation, climate/weather modeling, aircraft engine diagnostics, earthquake engineering, virtual observatory, bioinformatics, drug discovery, digital image analysis, astrophysics, and multi-player gaming. Grids can primarily be classified [7] into various types, depending on the nature of their emphasis: computation, data, application service, interaction, knowledge, and utility. Accordingly, grids are proposed as the emerging cyber infrastructure to power utility computing applications. Computational grids aggregate the computational power of globally distributed computers (e.g., TeraGrid, ChinaGrid, and APAC-Grid). Data grids emphasize global-scale management of data to provide data access, integration, and processing through distributed data repositories (e.g., LHCGrid, Gri-PhyN). Application service (provisioning) grids focus on providing access to remote applications and modules; libraries hosted on data centers or computational grids (e.g., NetSolve and Grid-Solve). Interaction grids focus on interaction and collaborative visualization between participants (e.g., AccessGrid). Knowledge grids aim toward knowledge acquisition, processing, and management, and provide business analytics services driven by integrated data mining services. Utility grids focus on providing all the grid services, including compute power, data, and service to end users as IT utilities on a subscription basis, and provide the infrastructure necessary for negotiation of required quality of service, establishment and management of contracts, and allocation of resources to meet competing demands. To summarize, these grids follow a layered design IEEE Communications Surveys & Tutorials • 2nd Quarter 2008 6

The Gridbus Middleware for Market‐Oriented Computing

“…Therefore, our design of a Grid-Federation focuses on two important aspects: a distributed resource discovery system [21,25] and a marketbased resource allocation system [26]. Grid-Federation allows cooperative sharing of topologically and administratively distributed Grid resources.…”

“…26.15). In general, a GFA service undertakes two basic types of queries [21]: (1) a resource lookup query (RLQ)-a query issued by a GFA service to locate resources matching the user's application QoS requirement and (2) a resource update query (RUQ), which is an update query sent to a resource discovery by a GFA (on behalf of the Grid site owner) about the underlying resource conditions. Since a Grid resource is identified by more than one attribute, a RLQ or RUQ is always multidimensional.…”

Survey on Grid Resource Allocation Mechanisms

“…(a) computational Grid (e.g., TeraGrid, ChinaGrid, and APACGrid [107]) that combines the computational power of distributed resources including clusters, desktops, and supercomputers that provide services for high performance computing [55][56][57], (b) access Grid [106,125] that offers limited specific resources for a short period of time, (c) data Grid (e.g., LHCGrid, GriPhyN [107]) is used by data-intensive tasks that consists of distributed data repositories providing the facilities to store enormous amount of data that can be accessed, moved, and processed as if they were small files [65,138], and [122] . .…”

SARIDS: A Self-Adaptive Resource Index and Discovery System

“…However, the centralized or hierarchical approach has some limitations in terms of scalability, fault tolerance, and self-organizing. To overcome these drawbacks, the decentralized model [12,17] is proposed for indexing and discovering resource information, e.g., the Peer-to-Peer manner.…”