Introduction Like many organisations we are interested in managing computer clusters as resources for carrying out computationally intensive tasks. In common with many University organisations our resources are heterogeneous and have to serve many purposes. Although we have been able to develop computational clusters where a tranche of nodes was entirely dedicated we believe it is highly attractive to support cluster management of resources that may fluctuate in number or availability at different times. This is important for management of local resources but is also an important foundation idea for sharing resources between different organisations. In this paper we report on our prototype cluster management framework using a tuple spaces model and show how a coordination management system can be set up independent of any communications mechanism needed for achieving high performance amongst nodes working on parallel jobs. The tuple space model widely popularised by Gelernter

Efficient and robust metacomputing requires the decomposition of complex jobs into tasks that must be scheduled on distributed processing nodes. There are various ways of creating a schedule and implementing it efficiently, depending upon global system state knowledge. Many computations may be structured as process networks, where data is either pushed from the source node to the target node where it will be used, or is pulled from source to target at the instigation of the target. We have developed a metacomputing infrastructure to investigate this idea, which employs the concept of a rich data pointer, the DISCWorld Remote Access Mechanism (DRAM), which can point to either data or services, and can be traded in a client/multiple-server model. We present an extension of the DRAM concept and implementation to represent and describe data that has not yet been created, the "DRAM Future" (DRAMF). We show how the use of the DRAMF facilitates efficient metacomputing scheduling and runtime optimisation on high performance distributed systems. We present a recursive algorithm for determining the optimal placement of a job's components in the presence of partial system state information. This algorithm uses only a selected subset of all available processing nodes, and we implement it using DRAMFs. There are many research issues to consider when designing a robust and general algorithm for scheduling and process placement on distributed systems. We address some of these issues in our Distributed Information Systems Control World project[6] as do other research projects[2].

We describe distributed and parallel algorithms for processing remotely sensed data such as geostationary satellite imagery. We have built a distributed data repository based around the client-server computing model across wide-area ATM networks, with embedded parallel and high-performance processing modules. We focus on algorithms for classification, geo-rectification, correlation and histogram analysis of the data. We consider characteristics of image data collected from the Japanese GMS5 geostationary meteorological satellite, and some analysis techniques we have applied to it. As well as providing a browsing interface to our data collection, our system provides processing and analysis services on-demand. We are developing our system to carry out processing and data reduction services at-a-distance, enabling remote users with limited bandwidth, access to our system, to obtain useful derived data products at the resolution they require.Our target hardware consists of a heterogeneous collection of distributed workstations, multi-processor servers and massively parallel computers at locations throughout Australia. These platforms are connected by ATM-based LANs and also through ATM switches across long distance WANs such as Telstra's Experimental Broadband Network, connecting Adelaide, Melbourne and Canberra. Our particular interest in constructing remote data access and processing services is the potential to utilise such resources as are available to a given user, yielding the best performance compromise of data processing and data delivery. To this end, we are building a set of resource scheduling and management utilities that will integrate the processing modules we describe. We have considered a number of software frameworks for building our integrated system and are focusing on a distributed object model using WWW protocols and the Java language.