Ontology-based Grid resource management

Balachandar R. Amarnath, Thamarai Selvi Somasundaram, Mahendran Ellappan, Rajkumar Buyya
<span title="2009-12-10">2009</span> <i title="Wiley"> <a target="_blank" rel="noopener" href="https://fatcat.wiki/container/afve5b3tavbfzch46mdtazbqla" style="color: black;">Software, Practice &amp; Experience</a> </i> &nbsp;
Grid resources are typically diverse in nature with respect to their software and hardware configurations, resource usage policies and the kind of application they support. Aggregating and monitoring these resources, and discovering suitable resources for the applications become a challenging issue. This is partially due to the representation of Grid metadata supported by the existing Grid middleware which offers limited scope for matching the job requirements that directly affect scheduling
more &raquo; ... isions. This paper proposes a semantic component in conventional Grid architecture to support ontology-based representation of Grid metadata and facilitate context-based information retrieval that complements Grid schedulers for effective resource management. Web Ontology language is used for creating Grid resource ontology and Algernon inference engine has been used for resource discovery. This semantic component has been integrated with conventional Grid schedulers. Several experiments have also been carried out to investigate the performance overhead that arises while integrating this component with Grid schedulers. However, there was no technology to support flexible and controlled sharing of various types of resources that are needed to solve computationally intensive applications. To address this issue, an extended distributed computing technology, termed as 'Grid', was proposed which facilitates the aggregation of distributed computational resources that spans beyond organizational boundaries, and their coordinated utilization to meet the requirements of advanced science and engineering applications [1] [2] [3] . Consequently, Grid middleware has been proposed, that performs basic authentication and authorization of the participants of the Grid and governs execution of job and monitors the dynamic state of the participating resources. However, due to the diverse nature of participating Grid resources, Grid middlewares are limited in functionality with respect to coordinating and managing the resources for application execution. Further, Grid applications often require computational resources with different software and hardware configurations. Hence, discovering suitable resources that match the application requirements are really a difficult task. In such an environment, the role of Grid schedulers is very important as they often need to manage various resources information and their status during application scheduling. For instance, as soon as a job arrives in the job queue, the Grid scheduler aggregates Grid resource information and performs matchmaking to find out suitable Grid resource information. A sophisticated Grid scheduling mechanism also considers resource usage policies before scheduling application to the resources, and establishes service level agreement to ensure the proper delivery of the requested Quality of Service (QoS) parameters. However, the representation of these metadata supports the keyword-based discovery of resources that often miss relatively capable resources. For instance, an application requests Linux Operating System for execution. If this resource is not available, conventional Grid schedulers put the application in pending state due to non-availability of suitable resource. However, in most cases, this application can also be executed in resource that has Fedora or other Unix-based Operating Systems. Unfortunately, currently existing Grid schedulers do not possess the capability to infer the relationship between the two operating systems. This is due to the fact that the underlying Grid middleware such as Globus [4], gLite [5] and Unicore [6] defines and implements mechanisms for resource discovery and monitoring, which supports traditional service matching based on symmetric, attribute-based matching and does not support semantic descriptions of Grid resources. Meanwhile, semantic web technology offers greater support in describing a particular domain that makes it possible to infer semantic relationship between two entities. Exploiting such technologies in Grid resource management provides greater flexibility in making scheduling decisions such as resource discovery for application execution, policy negotiation and establishing Service Level Agreements (SLA) between the users and Grid resource providers. This initiative has led to the emergence of Semantic Grid technology. The Semantic Grid is an extension of the current Grid in which information and services are given well-defined meaning through machine understandable descriptions which helps in making intelligent scheduling decisions with a high degree of automation in computational Grid infrastructure [7] . Metadata in Grid includes available Grid resource information, job execution status and information related to resource usage policies. Representing such information with the help of semantic web technologies would support context-based information retrieval that can complement Grid schedulers while making scheduling decisions such as resource allocation and policy management. The objective of semantic Grid's approach is to make information 'understandable' by computers. Information must therefore be described in such a way that computers can interpret it and derive its meaning. The concept of ontology helps to describe the information by expressing the relationship between them. The semantic representation of information is exploited while retrieving information using an ontology reasoning language. Hence, The semantic Grid architecture interacts with the underlying Grid resources and creates its ontology representation to support semantic-based information retrieval. Further, it should not impose any special requirements to be fulfilled by the participating Grid resources. In addition, semantic component must be able to incorporate with different Grid middlewares. To meet the above requirements, the implementation of semantic component is not tightly coupled with the underlying Grid middleware. It uses the standard way of interacting with the Grid middleware and automatically creates ontology description of Grid resources. A four-layered semantic Grid architecture is modeled with the knowledge layer sitting at the top of the Grid scheduler in the high-level Grid middleware layer as shown in Figure 1 . Fabric layer: The Grid Fabric layer provides the resources to which shared access is mediated by Grid protocols. The resources may be computational resources, storage systems, catalogues, network resources and sensors or may be a logical entity, such as a distributed file system, computer cluster or distributed computer pool. Figure 1. Semantic grid architecture.
<span class="external-identifiers"> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1002/spe.942">doi:10.1002/spe.942</a> <a target="_blank" rel="external noopener" href="https://fatcat.wiki/release/cq6d2ljzfrhxvb2dxvictm24e4">fatcat:cq6d2ljzfrhxvb2dxvictm24e4</a> </span>
<a target="_blank" rel="noopener" href="https://web.archive.org/web/20110308093909/http://www.buyya.com/papers/OntologyGrid-SPE2009.pdf" title="fulltext PDF download" data-goatcounter-click="serp-fulltext" data-goatcounter-title="serp-fulltext"> <button class="ui simple right pointing dropdown compact black labeled icon button serp-button"> <i class="icon ia-icon"></i> Web Archive [PDF] <div class="menu fulltext-thumbnail"> <img src="https://blobs.fatcat.wiki/thumbnail/pdf/63/ee/63ee360a426ece3cec75b867a0495b0d382e379b.180px.jpg" alt="fulltext thumbnail" loading="lazy"> </div> </button> </a> <a target="_blank" rel="external noopener noreferrer" href="https://doi.org/10.1002/spe.942"> <button class="ui left aligned compact blue labeled icon button serp-button"> <i class="external alternate icon"></i> wiley.com </button> </a>