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

Mentat is an object-oriented parallel processing system designed to simplify the task of writing portable parallel programs for parallel machines and workstation networks. The Mentat compiler and run-time system work together to automatically manage the communication and synchronization between objects. The run-time system marshals member function arguments, schedules objects on processors, and dynamically constructs and executes large grain data dependence graphs. In this paper we present the Mentat runtime system. We focus on three aspects --- the software architecture, including the interface to the compiler and the structure and interaction of the principle components of the runtime system; the run-time overhead on a component by component basis for two platforms, a Sun SparcStation 2 and an Intel Paragon; and an analysis of the minimum granularity required for application programs to overcome the run-time overhead. Keywords: object-oriented, parallel processing, dataflow, distribu...

The coming of giga-bit networks makes possible the realization of a single nationwide virtual computer comprised of a variety of geographically distributed high-performance machines and workstations. To realize the potential that the physical infrastructure provides, software must be developed that is easy to use, supports large degrees of parallelism in applications code, and manages the complexity of the underlying physical system for the user. This short paper briefly describes our approach to constructing and exploiting such "metasystems". Our approach inherits features of earlier work on parallel processing systems and heterogeneous distributed computing systems. In particular, we are building on Mentat, an object-oriented parallel processing system developed at the University of Virginia. A more detailed presentation can be found in technical report CS 94-21, "Legion: The Next Logical Step Towards a Nationwide Virtual Computer".

. We compare two load balancing techniques for Cayley graphs based on information and load exchange between neighboring vertices. In the first scheme, called natural diffusion, each vertex gives (or receives) a fixed part of the load difference to (from) its direct neighbors. In the second scheme, called Cayley diffusion, each vertex successively gives (or receives) a part of the load difference to (or from) direct neighbors incident to the edges labeled by the elements of the generator set of the Cayley graph. We prove that the convergence of the Cayley diffusion is faster than the natural diffusion, at least for some particular graphs (cube, circuit with an even number of vertices, graphs from the symmetric group). Furthermore we compute the fastest possible way to distribute load in a circuit using local load balancing strategies. Topics covered. Theory of Parallel and Distributed Computation, Parallel Algorithms, Load Balancing, Cayley Graphs, Complexity. MSC Mathematics Subject ...

Resource allocation is a difficult and challenging problem. With increasing interest in using networks of workstations as high performance, general-purpose multi-user computing platforms it is even more so. In order to utilise these systems, it is important for resource allocation to dynamically maximise performance, whilst maintaining fairness between competing processes. In this paper, we present a low-overhead proportion-share resource allocator which maintains fairness between competing processes by allocating resources dependent on the process' relative importance. Combined with this, we extend access to the resource allocator through a distributed marketplace to provide processes with an environment of the demands on their surrounding resources. Using this environment processes are provided with the information on which they base their resource decisions. The use of this marketplace to enable processes to evaluate their resource surroundings in order to improve their performance ...

. This paper describes a model which uses market ideas to allow uniform, scalable allocation of resources by multiple distributed applications in distributed systems. In this situation, using environment unaware manual placement of work by applications will lead to poor load-balancing. Traditionally, such allocation needs to take place in the operating system using heuristics in the presence of incomplete information. We assert that environment aware applications have far better knowledge of the resources they should acquire than any operating system ever could. Using a market model, each process attempts to obtain the specific resources it requires at market price from a limited supply using restricted funds. Similarly, resource managers attempt to maximise their revenues based on the demands for their services. The conflict between these two determines the price and hence a process's preference for each resource. We present a model which utilises market principles, and demonstrate bo...

The Tandem NonStop is a loosely-coupled multi-computer system managed by Guardian 90, a message-based distributed operating system designed to provide an environment for online transaction processing. One of the benefits of a loosely-coupled architecture is its inherently distributed character. A distributed architecture allows many components to be applied scalably in parallel to a large data-intensive task, such as batch or query processing, or to many small independent tasks, such as online transaction processing. However, achieving good performance in a loosely-coupled architecture presents some challenges relative to a tightly-coupled architecture. These challenges include: • Overcoming the high cost of inter-process communication. • Performing load-balancing without shared memory. • Solving the client-server priority inversion problem. Overcoming these challenges has required complex performance-oriented optimizations in Guardian. These optimizations- aimed at reducing message traffic, avoiding software

Legion was the first integrated grid middleware architected from first principles to address the complexity of grid environments. Just as a traditional operating system provides an abstract interface to the underlying physical resources of a machine, Legion was designed to provide a powerful virtual machine interface layered over the distributed, heterogeneous, autonomous, and fault-prone physical and logical resources that constitute a grid. We believe that without a solid, integrated, operating system-like grid middleware, grids will fail to cross the chasm from bleeding-edge supercomputing users to more mainstream computing. This paper provides an overview of the architectural principles that drove Legion, a high-level description of the system with complete references to more detailed explanations, and the history of Legion from first inception in August 1993 through commercialization. We present a number of important lessons, both technical and sociological, learned during the course of developing and deploying Legion. Keywords—Distributed object system, grid, grid architecture, grid design philosophy, large-scale distributed system, metaoperating systems, metasystems. I.