Scheduling in the context of parallel systems is often thought of in terms of assigning tasks in a program to processors, so as to minimize the makespan. This formulation assumes that the processors are dedicated to the program in question. But when the parallel system is shared by a number of users, this is not necessarily the case. In the context of multiprogrammed parallel machines, scheduling refers to the execution of threads from competing programs. This is an operating system issue, involved with resource allocation, not a program development issue. Scheduling schemes for multiprogrammed parallel systems can be classified as one or two leveled. Single-level scheduling combines the allocation of processing power with the decision of which thread will use it. Two level scheduling decouples the two issues: first, processors are allocated to the job, and then the job's threads are scheduled using this pool of processors. The processors of a parallel system can be shared i...

Abstract not found

. Jobs that do not require all processors in the system can be packed together for gang scheduling. We examine accounting traces from several parallel computers to show that indeed many jobs have small sizes and can be packed together. We then formulate a number of such packing algorithms, and evaluate their effectiveness using simulations based on our workload study. The results are that two algorithms are the best: either perform the mapping based on a buddy system of processors, or use migration to re-map the jobs more tightly whenever a job arrives or terminates. Other approaches, such as mapping to the least loaded PEs, proved to be counterproductive. The buddy system approach depends on the capability to gang-schedule jobs in multiple slots, if there is space. The migration algorithm is more robust, but is expected to suffer greatly due to the overhead of the migration itself. In either case fragmentation is not an issue, and utilization may top 90% with sufficiently...

Abstract not found

This paper presents a method for qualitative and quantitative analysis of load sharing algorithms, using a number of well known examples as illustration. Algorithm design choices are considered with respect to the main activities of information dissemination and allocation decision making. We argue that nodes must be capable of making local decisions, and for this efficient state dissemination techniques are necessary. Activities related to remote execution should be bounded and restricted to a small proportion of the activity in the system. The quantitative analysis provides both performance and efficiency measures, including consideration of the load and delay characteristics of the environment. To assess stability, which is also a precondition for scalability,we introduce and measure load sharing hit-ratio, the ratio of remote execution requests concluded successfully. Using our analysis method, we are able to suggest improvements to some published algorithms.

A process is an operating system abstraction representing an instance of a running computer program. Process migration is the act of transferring a process between two machines during its execution. Several implementations

Data-intensive, interactive applications are an important class of metacomputing (Grid) applications. They are characterized by large dataflows between data providers and consumers, like scientific simulations and remote visualization clients of simulation output. Such dataflows vary at runtime, due to changes in consumers' data needs, changes in the nature of the data being transmitted, or changes in the availability of computing resources used by flows. The topic

This report details a survey of systems providing process or object migration. After an introduction to the topic a number of systems are covered in detail. Four chapters deal with systems providing migration over unmodified UNIX, those providing "classical" process migration and containing modified or new kernels, migration over modern microkernel based operating systems and finally those providing more general "object" migration. The coverage of each system contains a general overview, a detailed examination of the migration subsystem and load balancing support, if any. Migration support is sectioned into an overview, an examination of the migration mechanisms, the effects of migration on inter (task) communication, and details of any environment and associated resource transfer. The summary provides comments on the major design issues identified within the report. It notes that solutions running over raw UNIX are the least complete, and that some level of kernel modification is gene...

Peer-to-peer (P2P) technology has undergone rapid growth, producing new protocols and applications, many of which enjoy considerable commercial success and academic interest. Yet, P2P applications are often based on complex protocols, whose behavior is not completely understood. We believe that in order to enable an even more widespread adoption of P2P systems in commercial and scientific applications, what is needed is a modular paradigm in which well-understood, predictable components with clean interfaces can be combined to implement arbitrarily complex functions. The goal of this paper is to promote this idea by describing our initial experiences in this direction. Our recent work has resulted in a collection of simple and robust components, which include aggregation and membership management. This paper shows how to combine them to obtain a novel load-balancing algorithm that is close to optimal with respect to load transfer. We also describe briefly our simulation environment, explicitly designed to efficiently support our modular approach to P2P protocol design.

This paper presents a distributed file organization for record-structured, disk-resident files with key-based exact-match access. The file is organized into buckets that are spread across multiple servers, where a server may hold multiple buckets. Client requests are serviced by mapping keys onto buckets and looking up the corresponding server in an address table. Dynamic growth in terms of file size and access load is supported by bucket splits and bucket migration onto other existing or newly acquired servers. The significant and challenging problem addressed here is how to achieve scalability so that both the file size and the client throughput can be scaled up by linearly increasing the number of servers and dynamically redistributing data. Unlike previous work with similar objectives, our data redistribution considers explicitly the cost/performance ratio of the system by aiming to minimize the number of servers that are acquired to provide the required performance. A new server i...