Conference Paper

Grid allocation and reservation - Grid capacity planning with negotiation-based advance reservation for optimized QoS.

Conference: Proceedings of the ACM/IEEE SC2006 Conference on High Performance Networking and Computing, November 11-17, 2006, Tampa, FL, USA
Source: DBLP
4 Reads
  • Source
    • "medium, low, very low}, the approach proposed by Mumtaz et al. in [17] is adopted. Based on a pre-defined semanticsbased hierarchical structure of all possible values of a nonnumerical QoS parameter, each level of the hierarchy is associated with a numerical value. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Cloud-based software applications (Software as a Service - SaaS) for multi-tenant provisioning have become a major development paradigm in Web engineering. Instead of serving a single end-user, a multi-tenant SaaS provides multiple end-users with the same functionality but with potentially different quality-of-service (QoS) values. The service selection for such a SaaS is a complex decision-making process which involves a number of stakeholders with different QoS requirements. SaaS developers need to compose services with different QoS values to meet end-users' different multidimensional QoS constraints for the SaaS. Furthermore, they also need to satisfy SaaS providers' optimisation goals for the SaaS, such as least resource cost and best system performance. Existing QoS-aware service selection approaches are oriented at a single tenant. They do not consider the characteristics of multi-tenant SaaS and hence are ineffective and inefficient when applied to compose multi-tenant SaaS. In this paper, we introduce a novel QoS-driven approach for helping SaaS developers select the services for composing multi-tenant SaaS, which achieves SaaS providers' optimisation goals while fulfilling the end-users' different levels of QoS constraints. The proposed approach is evaluated using an example SaaS synthetically generated based on a dataset of real-world Web services. Experimental results show that our approach significantly outperforms existing approaches in terms of both effectiveness and performance.
    5th IEEE 2012 International Conference on Cloud Computing (CLOUD 2012), Honolulu, Hawaii, United States, 24-29 June 2012; 01/2012
  • Source
    • "In the context of Grid or HPC scheduling the benefits of using overbooking are pointed out, but no solutions are provided [14] [15]. Overbooking is also foreseen in a three layered protocol for negotiation in Grids [16]. Here, the restriction is made that overbooking is only used for multiple reservations for workflow sub-jobs which were made by the negotiation protocol for optimal workflow planning. "

    Proceedings of the 2008 International Conference on Grid Computing & Applications, GCA 2008, Las Vegas, Nevada, USA, July 14-17, 2008; 01/2008
  • Source
    • "Clusterbased Real-time divisible load scheduling is investigated in [8], [9], [6], [10], [7]. However, in a cluster, advance reservations have been largely ignored due to the underutilization concerns and lack of support for agreement enforcement [11]. In a cluster with no reservation, resources are allocated to tasks until they finish processing. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Providing QoS and performance guarantees to arbitrarily divisible loads has become a significant problem for many cluster-based research computing facilities. With the emergence of Grid applications that require simultaneous access to multi-site resources, supporting advance reservations in a cluster has become increasingly important. While progress is being made in scheduling arbitrarily divisible loads, current approaches either do not support advance reservations, or are not efficient. In this paper, we propose a linear algorithm for real-time divisible load scheduling that supports advance reservations in a cluster. Unlike existing approaches, the new algorithm relaxes the tight coupling between the task admission controller and the task dispatcher. By eliminating the need to generate exact schedules in the admission controller, the algorithm avoids high overhead. Our approach also addresses the under-utilization concerns raised by advance reservations.
    Real-Time Systems 01/2008; 48(3). DOI:10.1007/s11241-011-9145-6 · 1.00 Impact Factor
Show more