|Title:||A Distributed and Heuristic Policy-based Management
Architecture for Large-Scale Grids
|Committee:||David Hausheer (UniZH), Laurent Lefevre (INRIA),
Joan Olmos (UPC), Leandro Navarro (UPC), Juan L. Gorricho (UPC)
PhD project description
Nowadays, software applications are developed to solve multi-disciplinarian problems in wide different contexts, such as weather forecast, aerospace experiments, simulation of drugs behaviour, etc. Computing resources are used to accelerate or fully automate the processing, quantification and analysis of large amounts of high-information content. These applications need powerful computing systems to properly execute their computational requirements. Unfortunately, computing demands from novel applications have increased faster than the development of powerful computational resources. Moreover, the access to high-throughput systems is very expensive and limited by several conditions. On the other hand, network connectivity between geographically dispersed computing systems (workstations) is every day faster and cheaper. Therefore, novel network technologies started to develop distributed technologies to execute high-throughput applications in order to facilitate, increase and improve the use of any computational resource in multi-disciplinarian research areas. Grid Computing is defined as a heterogeneous, distributed and shared system where applications are solved in dynamic, multi-institutional virtual organizations (VOs). This key concept involves the inherent ability to negotiate resource-sharing arrangements among a set of participating parties (providers and costumers) and then to use the resulting resource pool for some purpose. Basically, Grids should integrate and coordinate resources and users that live within different control domains. Besides, it is built from multi-purpose protocols and interfaces that address such fundamental issues as scheduling, security, resource discovery, and resource allocation. Finally, Grid allows its constituent resources to be used in a coordinated fashion to deliver various qualities of service, relating for example to response time, throughput, availability and/or co-allocation of multiple resource types to meet complex user demands, so that the utility of the combined system is significantly greater than that of the sum of its parts [Foster_02]. Large-scale Grids are formed by thousands of nodes sharing multiples kind of resources (computing, networking, software applications, high-technology instruments, etc) and therefore, the total amount of shared resources become millions of individual entities that most be adequately integrated and coordinated for solving multi-disciplinarian problems. In this dynamic, heterogeneous and geographically dispersed environment, Resource Management (RM) is regarded as a vital component of the Grid Infrastructure. Grid Resource Management (GRM) coordinates and shares multiple kinds of resources efficiently; and, in the above mentioned environments, GRM faces several challenges that make the implementation of practical systems a very difficult problem. Among these challenges we would like to emphasize the fact that GRM systems must fulfil strict functional requirements from heterogeneous, and sometimes conflicting, domains (e.g., the users’, applications and networks domains). In addition, GRM systems must adhere to non-functional requirements that are also rigid, such as reliability and efficiency in terms of time consumption and load on the host nodes. The aim of this thesis is to design and implement a new Grid Resource Management methodology, where non-massive resources owners would be able to share their resources and integrate human collaboration across multiple domains regardless of network technology, operative platform or administrative domain. This thesis proposes a distributed and heuristic policy-based resource management architecture for large-scale Grids. The resource management architecture proposed herein is composed of three main phases: resource discovery and monitoring, resource scheduling and jobs allocation. The resource discovery and monitoring phase is supported by the introduction of SNMP-based Balanced Load Monitoring Agents for Resource Scheduling (SBLOMARS), in which network and computational resources are monitored by autonomous agents. This allows for a flexible, heterogeneous and scalable monitoring system. The resource scheduling phase is based on the Balanced Load Multi-Constraint Resource Scheduler (BLOMERS). This heuristic scheduler represents an alternate way of solving the inherent NP-hard problem for resource scheduling in large-scale distributed networks. The jobs allocation phase is supported by means of a Policy-based Grid Management Architecture (PbGMA). This architecture is able to consider service needs arising from diverse sources during the deployment and management of Grid Services, such as requirements demanded by customers, applications and network conditions. The synergy obtained by these components allows Grid administrators to exploit the available resources with predetermined levels of Quality of Service (QoS), reducing computational costs and makespan in resource scheduling while ensuring that the resource load is balanced throughout the Grid. The makespan of a schedule is the time required for all jobs to be processed when no one job could be interrupted during its execution and each node can perform at most one operation at any time. This new approach has been successfully tested in a real large-scale scenario such as Grid5000.
- E. Magaña, Laurent Lefevre and J. Serrat, “SBLOMARS: SNMP-based Balanced Load Monitoring Agents for Resource Scheduling in Grids”, GCA'07: The 2007 International Conference on Grid Computing and Applications. Las Vegas, Nevada, USA. June 25-28, 2007.
- E. Magaña, and J. Serrat, “Distributed and Heuristic Policy-based Resource Management System for Large-scale Grids”, AIMS 2007: The ACM Conference in Autonomous Infrastructure, Management and Security. Oslo University College, Norway. June 21-22 2007.
- E. Magaña, Masum Hasan and J. Serrat, "BLOMERS: Balanced Load Multi-Constrained Resource Scheduler”, The Third International Conference on Networking and Services (ICNS 2007). Athens, Greece. June 19-25, 2007 -
- E. Magaña, Laurent Lefevre and J. Serrat, "Autonomic Management Architecture for Flexible Grid Services Deployment Based on Policies", ARCS'07: Architecture of Computing Systems, Swiss Federal Institute of Technology (ETH) Zurich, Switzerland March 12-15, 2007.
- E. Magaña and J. Serrat, "Services and Applications Management Middleware for Autonomic Grid Computing", 1st IEEE/IFIP International Workshop on Autonomic Grid Networking and Management (Manweek2005), Universitat Politècnica de Catalunya, Barcelona, Spain. October 24-26, 2005.
- E. Magaña and J. Serrat, "QoS Aware Policy-Based Management Architecture for Service Grids", 14th IEEE International Workshops on Enabling Technologies: Infrastructures for Collaborative Enterprises (WETICE/ETNGRID 2005). Linkoping University, Sweden, June 13-15, 2005.
- E. Magaña and J. Serrat, "QoS Aware Resource Management Architecture for OGSA Services Deployment", 9th IFIP/IEEE International Symposium on Integrated Network Management (IM 2005). Nice, Acropolis, France. June 19-25, 2007. (Poster)
- E. Magaña and J. Serrat, "Adaptive Management Middleware for Grid Services and Applications Based on Policies", Workshop on Adaptive Grid Middleware (AGridM2004), Antibes Juan-les-Pins, France, September 29, 2004.
- E. Magaña, E. Salamanca, J. Serrat, "A Proposal of Policy-Based System Architecture for Grid Services Management", Active and Programmable Grids Architectures and Components (APGAC'04). Kraków, Poland. June 19-25, 2007.
- C. Tsarouchis, S. Denazis, Chiho Kitahara, J. Vivero, E. Salamanca, E. Magaña, A. Galis, J. Mañas, Y. Carlinet, et.al., "Policy-Based Management Architecture for Active and Programmable Networks", IEEE Network, May 2003, Vol.17, No.3.
This thesis will be defended on May 30th, 2008 in Barcelona The thesis manuscript is available to EMANICS members here]
- [_URL_ Homepage] of Edgar Magaña
- Publications of Edgar Magaña, as indexed by DBLP