Elizabeth A. Kendall

Agent software engineering with role modelling

Due to their autonomy and social behavior, agents will play important roles in future emerging enterprises. They will fill key positions and provide essential capabilities. We propose role modelling as a software engineering technique for specifying, analyzing, and designing systems on the basis of the roles that the agents will play. Our approach builds on our earlier research in patterns [4], [5] of agent systems. Object-oriented role models can be extended to represent patterns of agent interaction that can then be employed to engineer agent systems.

Distributed network file storage for a serverless (P2P) network

The design of a persistent peer-to-peer (P2P) network storage application is presented in the paper. SNS separates the application from the underlying P2P network, i.e. the serverless layer. The serverless layer is responsible for routine networking tasks such as self-organisation and maintaining the network state information. The application specific functions are implemented through the file information protocol e.g. maintaining and synchronizing file and disk space information by using lightweight XML-formatted messages. The distributed storage is made to appear as a single large network drive, with all the main features of a conventional file system being made available to the client nodes.

A Model Driven Exception Management Framework for Developing Reliable Software Systems

Programming languages provide exception handling mechanisms to structure fault tolerant activities into software systems. However, the use of exceptions at this low level of abstraction can be error-prone and complex leading to new programming errors. In this paper, we present a model-driven framework to support the iterative development of reliable software systems. This framework is comprised of UML-based modeling notations and a transformation engine that supports the automated generation of exception management features for a software system. It leverages domain specific exception modeling languages, patterns, modeling tools and framework libraries. The feasibility of this approach is demonstrated through the development of a case study business application, known as Project Tracker

A software engineering process for BDI agent-based systems

Extensive research for BDI (belief, desire, intention) agents has been carried out. To employ BDI agents in an application such as life cycle costs (LCC) information gathering, a methodology is required to analyze, design and specify agents from system requirements. Although some methodologies are available for development of agent-based systems, none provide a systematic approach for BDI agent-based systems involving all phases of the software development life cycle, especially for addressing agent specifications. This paper exploits synergy across different research areas such as use cases, BDI agents, role models, and IDEF, and proposes a methodology for BDI agent-based systems. This methodology presents a step-by-step process for agent-oriented software engineering. This process has been applied to designing a BDI agent-based information gathering system for LCC.

Algorithm Exchange of a Security Control System for Web Services Applications

It has been determined that the complexity of Internet infrastructures, such as in a distributed system (Web services), can conceal the potential risks of attacks, and subsequently become disadvantageous to Internet users, applications and institutions. We have considered that a better way to cope with such problems can be achieved through utilisation of a systems approach, rather than an analytical approach. In this article the development architecture will be described, and the implementation results of algorithm exchange as part of the TTSN architecture (the project that is currently being undertaken) will be presented. Since Web services applications perform XML-based transactions, which are very vulnerable against an interception, algorithm exchange is a superior alternative than key exchange, the currently common strategy to protect confidentiality, integrity and authentication of data. The reason why key exchange is no longer appropriate for the current Internet technologies will also be discussed.

Dynamic data protocol exchange to handle passive attacks on trusted transient simple network (TTSN) transactions

The most difficult part of network security is protecting the traffic content from passive attacks. Passive attacks cannot be avoided because the data structure (data protocol) of Internet transactions is well known and of fixed standard. We have considered an alternative to handle passive attacks through utilising dynamic data protocol exchange. The utilisation of dynamic data protocol exchange within the proposed TTSN security control system can significantly increase the time consumed by a passive attack, i.e. security cryptanalysis or brute-force, on encrypted Internet traffic. Such a dynamic data protocol exchange mechanism allows both parties (requester and provider) to replace or modify the data structure of the traffic package as often as required. In this article we describe the development architecture, and we also present results from implementing the dynamic data protocol exchange as part of the TTSN architecture. Since Internet applications are utilising readable script languages, which are very vulnerable against an interception, dynamic data protocol exchange is a better strategy to protect confidentiality and integrity of data.

CDHT: the design of a clustered DHT routing scheme for serverless (P2P) networks

Avoiding hot spots and handling dynamic network environments are the inherent weaknesses of the DHT-based (dynamic hash table) p2p networks. The solution lies in isolating the DHT overlay from the network uncertainty and seeking an efficient and complementary lookup pattern outside the DHT. We present the design of a hybrid-overlay architecture for p2p networks in this paper that forms the DHT overlay based on the serverless layer clusters. The architecture delegates the handling of the network uncertainty to the network-aware serverless Layer; thus a more reliable network and content structure is constructed as the foundation of the DHT overlay. In addition, a new heuristic approach to content location is made possible to exploit network proximity and content popularity to balance network traffic and to improve the lookup efficiency for popular content.

Multimedia authoring framework with ILOG TM solver

Multimedia presentation delivery characteristics are impacted by the delivery environment including network bandwidth, latency, local disk storage, etc., and this may invalidate the presentation schedule. These varying characteristics are not addressed adequately in the authoring process. We propose a new approach to multimedia presentation scheduling that utilizes constraint technology and a flexible set of scheduling heuristics. The Constraint Satisfaction Authoring Framework aims at increasing the expressive power of a multimedia authoring paradigm with regard to resource capacity conflicts. The approach deals with domain uncertainty and incorporates it into the presentation schedule.

CDHT: a clustered DHT communication scheme for serverless (P2P) networks

The structured model or distributed hash table (DHT) based peer-to-peer (P2P) networks are favoured in research but not in real applications. Its argued that the unstructured model is more suitable for the real world. This paper compares the advantages and disadvantages of the structured and the unstructured model, and proposed a clustered DHT (CDHT) communication scheme which addresses some of the fundamental problems of DHT by combining the structured model with the unstructured one. This combination leads to a new content location scheme which balances network traffic and improves the lookup efficiency for popular content. The experimental results show CDHTs superiority over DHT in handling dynamic network conditions and avoiding hot spots.

Constructing personalized catalogs for speech-enabled applications

Organization of product information is important to help consumers find out what they want when shopping on the Internet. This issue is even more crucial to the success of voice-enabled m-commerce applications, where users are often in a time-critical situation and follow voice prompts to search for items of interest. The present study proposes a method combining techniques of decision tree and user profiling to generate personalized catalogs for voice-enabled m-commerce applications. Such catalogs can be used to construct interactive dialogs to guide consumers while finding needed products.