Harmony: a desktop grid for delivering enterprise computations

Abstract

This paper describes Harmony, a practical grid infrastructure built using personal computer resources. Harmony addresses the key concerns of end users for responsiveness, privacy and protection by isolating the grid computation in a virtual machine on the PC and by implementing a layered resource management architecture to divert workload to unutilized computers from those currently experiencing high levels of interaction. The use of a virtual machine separates the interactive workload software environment from that for the grid workload. Harmony also addresses the key concerns of enterprise IT by automating initial resource assignment and by automatically reallocating workload so as to meet quality of service goals. We have implemented a prototype of Harmony and demonstrated its capability to protect interactive performance. Our preferred grid workload is transactional -a key characteristic of commercial applications. The implementation uses Web Service-based interfaces, so the programming model of Harmony is compatible with and familiar to enterprise developers. We believe that Harmony demonstrates practical exploitation of a hitherto underutilized resource of considerable capability, with the potential to complement, or even in some cases replace, dedicated server-based resources.

Adaptive Resource Sharing in a Web Services Environment

“…This are described in more detail in [11]. The CMCs in the Logical Resource Layer act as coordinators between the Grid Service Layer and Physical Resource Layer.…”

“…In this paper, we describe an approach that uses hypervisor-based virtual machines. This is an extension of our earlier work described in [11]. Specifically, we make use of the VMWare Workstation product [4] to isolate the two types of workloads.…”

PAC-MEN: personal autonomic computing monitoring environment

“…This is particularly useful in situations where the PC is unattended for example a web server, and the user may be notified via a peer PC that the machine is in difficulty. Another useful situation is when machines in the peer group are sharing work load, for example via Harmony PC grid services [11]; a peer is notified in advance of immanent danger and can recover data and relocate work to another peer. Such an approach is more proactive than responding once the machine has hung, and as such offers fuller potential for autonomic capabilities.…”

“…The contribution of autonomic computing to personal computing is different-being much less about achieving optimum performance or exploiting redundancy and more about simplifying use of the equipment and the associated services involved [3]. Within autonomic personal computing, examples of current research include: (i) automatic system configuration, especially for application or communication settings, as in Prism, which uses a rulebased approach (ABLE) [10]; (ii) sharing workloads through a PC-based local Grid [11]; and (iii) autonomic management of wearable health devices [8].…”

Dynamic resource management heuristics for minimizing makespan while maintaining an acceptable level of robustness in an uncertain environment

“…One of the areas where this work is directly applicable is the development of resource allocations in enterprise systems that support transactional workloads sensitive to response time constraints, e.g., time sensitive business processes [22]. Often, the service provider in these types of systems is contractually bound through a service level agreement to deliver on promised performance.…”