Paul McKee

Efficient optical elements to generate intensity weighted spot arrays: design and fabrication.

We report on the design and fabrication of fully 2-D surface relief diffraction elements that can split a single collimated beam into many beams in an arbitrary intensity distribution. These splitters were designed by computer using simulated annealing, and made into phase gratings by electron-beam lithography followed by plasma etching into quartz glass. Both two and four phase level gratings have been fabricated, allowing a wide range of uniform and weighted spot patterns to be generated. These grating elements have a measured diffraction efficiency of over 74%, with the beam intensity ratios accurate to within 1% of their target values.

Dynamic Allocation of Servers to Jobs in a Grid Hosting Environment

As computational resources become available for use over the Internet, a requirement has emerged to reconfigure servers to an optimal allocation to cope with an unpredictable flow of incoming jobs. This paper describes an architecture that allows the dynamic reconfiguration of servers to process incoming jobs by switching servers between conceptual pools. The system makes use of heuristic policies to make close-to-optimal switching decisions. A prototype system which builds on existing resource management software has been developed to demonstrate some of the concepts described.

Behavioural specification using XML

Active networks are an example of a wide area distributed system. The objects in this system require platform independent machine-readable, flexible behaviour specifications. Current mechanisms in CORBA, DCOM and similar technologies are not adequate. We demonstrate for the first time how specifications with the required properties can be written in XML, the new language of the WWW.

Policy-based management for ALAN-enabled networks

This paper presents the architecture, policy schema, and policy specifications necessary to accomplish effective management of the application level active networking (ALAN) environment. Using ALAN, developers can engineer applications through the network by utilising platforms (active servers) on which 3rd party software (Proxylets) can be dynamically loaded and run. Redirection of packets destined for active processing at the servers is performed by active routers. Management of such large, dynamic systems presents challenges to centralised approaches. Management based on policies locally interpreted in the context of local state is gaining acceptance as an alternative. The IST project ANDROID uses a flexible generic specification for policies, represented in XML, allowing a wide range of policies to be expressed and processed in a common framework. Policies given here focus on management of routers for VPN scenarios, the resource and security management of active servers running the Proxylets, and management of the information distribution mechanism. Preliminary results were demonstrated during the trial which included the scenario involving the inter-site connectivity and active server resource and security management.

Active Robust Resource Management in Cluster Computing Using Policies

We present an implementation of a policy-based management architecture for emerging communications and computing paradigms such as Active Networks and the Grid. To manage such open, highly distributed and decentralized environments, an approach based on policy concepts is adopted, allowing support for active, dynamic adaptability in network elements, services and end-user applications, as well as achieving decentralization and distribution. We present our flexible, extensible policy and event specifications in XML, and describe our management architecture. One key feature of our approach is the distributed infrastructure: the Directory and the Management Information Distribution system. The second feature is the Resource and Security Management elements residing on the multi-node managed systems. These combine to provide a light-weight, self-organizing management architecture. As an applications example, we describe the implementation of our management system applied to the Application Level Active Networking (ALAN) environment, implemented in the European Commission Information Society Technologies (IST) project ANDROID.

Dynamic Allocation of Servers in a Grid Hosting Environment

A grid hosting environment is described, where servers may be reconfigured dynamically from one type of work to another in response to changes in demand. The problem of carrying out these reconfigurations in the most efficient manner is addressed by means of stochastic modelling and optimization. A heuristic policy, which is close to optimal over a wide range of parameters, is introduced. A prototype system based on existing resource management software has been developed to demonstrate the concepts described

Allocation and Admission Policies for Service Streams

A service provisioning system is examined, where a number of servers are used to offer different types of services to paying customers. A customer is charged for the execution of a stream of jobs; the number of jobs in the stream and the rate of their submission is specified. On the other hand, the provider promises a certain quality of service (QoS), measured by the average waiting time of the jobs in the stream. A penalty is paid if the agreed QoS requirement is not met. The objective is to maximize the total average revenue per unit time. Dynamic policies for making server allocation and stream admission decisions are introduced and evaluated. The results of several simulations are described.

Active management of multi-service networks

Future multi-service networks will be extremely large and complex. Novel management solutions will be required to minimise management costs. Active networking enables management to be delegated to network users. An active network based management solution for multi-service networking is presented.

A Policy Based Management Architecture for Large Scale Active Communication Systems

An initial design for a policy based management system combining conventional hierarchical control and significant local autonomy is described. A critical part of the design is a scheme of partial guarantees for policy distribution and execution. This scheme renders explicit the nondeterminism that is implicit in policy based control schemes that include conflict resolution, and to some extent replaces the need for conflict resolution. Some preliminary implementations of the design are described, and implications for further work are discussed.

Research Directions in Distributed Systems

This paper considers the technical characteristics of current distributed systems technology as exemplified by CORBA. In the context of large-scale globally distributed applications, potential problems are highlighted, and areas for further research and development are suggested.