Video surveillance is a key technology for enhanced protection of facilities such as airports and power stations from various types of threat. Networks of thousands of IP-based cameras are now possible, but current surveillance methodologies become increasingly ineffective as the number of cameras grows. Constructing software that efficiently and reliably deals with networks of this size is a distributed information processing problem as much as it is a video interpretation challenge. This paper demonstrates a software architecture approach to the construction of large scale surveillance network software and explores the implications for instantiating surveillance algorithms at such a scale. A novel architecture for video surveillance is presented, and its efficacy demonstrated through application to an important class of surveillance algorithms.

There exists a wide spectrum of coherency models for use in distributed shared memory (DSM) systems. The choice of model for an application should ideally be based on the application's data access patterns and phase changes. However, in current systems, most, if not all of the parameters of the coherency model are fixed in the underlying DSM system. This forces the application either to structure its computations to suit the underlying model or to endure an inefficient coherency model. This paper introduces a unique approach to the provision of DSM based on the idea of compliance. Compliance allows an application to specify how the system should most effectively operate through a separation between mechanism, provided by the underlying system, and policy, pro-vided by the application. This is in direct contrast with the traditional view that an application must mold itself to the hard-wired choices that its operating platform has made. The contribution of this work is the definition and implementation of an architecture for compliant distributed coherency management. The efficacy of this architecture is illustrated through a worked example.

A new mechanism for achieving isolated train detection in the distributed train algorithm is presented. The train algorithm is a general strategy for garbage collection, and the task of isolated train detection is a portion of the strategy which allows complete collection. When the train algorithm is deployed in a distributed system, a distributed algorithm is required to solve isolated train detection. We analyse the distributed train algorithm in the context of its isolated train detection aspect. We determine how it can be construed within the paradigm underlying distributed termination detection algorithms. From this we form a technical framework which allows any algorithm from this class to be used, essentially unmodified, to solve the distributed isolated train detection problem. We demonstrate this by deriving a new mechanism from a distributed termination detection algorithm.

Little work has been done in garbage algorithms for distributed shared memory systems. Mosaic is a safe and complete garbage collection system that collects garbage in object systems that are implemented above page-based distributed shared memory systems. It is non-intrusive in its impact on application performance. Mosaic partitions the virtual address space into separately managed regions, which form the basic unit of object storage. Garbage collection operates by moving objects between these partitions in a manner that associates live objects together, leaving unreachable objects behind so that partitions may be reused. To provide for safe operation of the collector, a distributed updated protocol maintains sufficient local knowledge of pointer duplication and destruction to allow safe determination of object reachability without the need to globally trace the object space. Mosaic exploits the semantics of pointer operations in type-safe object languages to allow for a weakened consistency model of pointer update during garbage collection.

The changing needs of modern application systems demand new and radical software architectures to support them. The attraction of persistent systems is that they define precisely the extent to which they are open, thereby allowing the dynamically changing resource requirements of applications to be tracked accurately within the persistent environment. Thus, an ever-growing body of work is being established to study the nature of running applications, and to use the information gleaned, to improve the run-time execution of these applications. Here we propose a new architectural approach to constructing persistent systems that accommodates, and thus is compliant to, the needs of particular applications. By separating policy from mechanism in all components, the architecture may be tailored to the policy needs of the application. (Policy may be regarded as strategy for achieving a goal, such as a cache eviction algorithm, whereas mechanism is the method by which the objective is achieved, such as the physical movement of the cache lines. As we see later, policy and mechanism are composable to form new mechanism.) We first propose a generic architecture for compliance, and then show how it may be instantiated. Finally, we describe an example of how the architecture operates in a manner that is compliant to a target application. We postulate, since we have not yet measured, that the benefits of compliant architectures will be a reduction in complexity, with corresponding gains in flexibility, portability, understandability in terms of failure semantics, and performance. Copyright © 2000 John Wiley & Sons, Ltd.

We demonstrate the use of a hyper-programming system in building persistent applications. This allows program representations to contain type-safe links to persistent objects embedded directly within the source code. The benefits include improved efficiency and potential for static program checking, reduced programming effort and the ability to display meaningful source level representations for first class procedure values. Hyper-programming represents a completely new style of programming which is only possible in a persistent programming system.

A new garbage collection algorithm for distributed object systems, called DMOS (Distributed. Mature Object Space), is presented. It is derived from two previous algorithms, MOS (Mature Object Space), sometimes called the train algorithm, and PMOS (Persistent Mature Object Space). The contribution of DMOS is that it provides the following unique combination of properties for a distributed collector: safety, completeness, non-disruptiveness, incrementality, and scalability. Furthermore, the DMOS collector is non-blocking and does not use global tracing.