System Research Group Seminars (SRGs)

These seminars provide an informal means for members of the School to keep in touch with what others are doing. They allow research students to see what lecturers are working on. They allow lecturers to see what research students are getting up to. And they give everyone practice in presentation skills. They also give people a chance to be exposed to awkward questions before giving talks at conferences! They are normally scheduled at Wednesday lunch time during the term time.

It is expected that all PhD students and Research Associates will present an SRG at least once during their stay, and that lecturing staff will present such talks every few years. All research students are strongly encouraged to attend, and lecturers are also recommended to do so.

At the moment there are no SRGs scheduled

Past SRGs

July 21, 2011 1pm, CLT.701
Exploring Visualization Techniques to Enhance Privacy Control UX for User-Managed Access/SmartAM
Domenico Catalano

Managing Data sharing policy for distributed resources could be mentally difficult for the end-user. The difficulty is to maintain control, correlate data and resources, and assign privileges for specific scope to own contacts or subject that they want access to the data. This could cause of lost of privacy, exploit of online personal data and significative security breach. To mitigate these risks, an advanced visualization approach is necessary to help the end-user during the definition of the data-sharing policy and in the lifecycle of the itself. This presentation summarize the 4 weeks of research on UX to enhance Privacy control for User-Manager Access applied applied at Smartam project. It will be presented the analysis approach, requirements and the results in terms of new UX concepts and interactions scenarios, including wireframes.

May 17, 2011 2pm, CLT 701
Agility: Building federated networks for sharing (and saving the world whilst we're about it)
Steve Caughey, Arjuna Technologies

Arjuna Technologies have recently been working with the School of Computing Science to reduce energy consumption. Steve Caughey, MD of Arjuna, will report on progress and explain how the company's 'Agility' product has been used within the project. Agility allows the creation of a federated network within an organisation and is intended to allow multiple departments, groups or teams to share IT resources flexibly whilst retaining ultimate control of those resources. The presentation will describe how Agility uses service agreements between parties and policy management to achieve these goals.

May 17, 2011 14:00-15:00, CLT701
Building federated networks for sharing (and saving the world whilst we're about it)
Steve Caughey, Arjuna Technologies

Arjuna Technologies have recently been working with the School of Computing Science to reduce energy consumption. Steve Caughey, MD of Arjuna, will report on progress and explain how the company's 'Agility' product has been used within the project. Agility allows the creation of a federated network within an organisation and is intended to allow multiple departments, groups or teams to share IT resources flexibly whilst retaining ultimate control of those resources. The presentation will describe how Agility uses service agreements between parties and policy management to achieve these goals.

April 19, 2011 2pm, CLT 701
How to ensure tallying integrity of large-scale electronic voting
Dr Feng Hao, Newcastle University

In this talk, I will first briefly introduce the research work I had done in the past. It covers three main areas: biometrics, cryptography and the interface between the two. In particular, I'll focus on my latest work on how to ensure tallying integrity in large-scale DRE-based electronic voting. The Direct Recording Electronic (DRE) is a widely deployed voting system that commonly uses touch-screen technology to directly record votes. In the USA, 32% voting machines had been electronic in 2008. India moved to fully DRE election in 2004 and Brazil started its first fully DRE election in 2002. However, a lack of tallying integrity has been considered the most contentious problem with the DRE system. In the talk, I will present a cryptographic protocol to ensure the tallying integrity -- even if the DRE machine is completely corrupted. Essentially, this protocol shifts from having to trust DRE software to "software independence". In other words, it does not matter how the software is written inside the system, by checking the cryptographic output of the software, voters can get strong assurance that the software is tallying votes correctly.

April 4, 2011 2pm, CLT 701
Workflows As Verifiable Analytical Middleware And Their Application To Clinical Guidelines
Dr. Vasa Curcin, Postdoctoral fellow, Department of Computing, Imperial College London

While workflows have attracted a great deal of attention as application middleware, mainly because of their intuitive representation of potentially complex processes and tasks in both business and scientific settings, their formal verification has been largely untackled until the last few years, when several models have been developed. These range from Petri Nets, via various flavours of process algebras to functional languages, each with their own associated model checking and reasoning techniques. In this talk, I will offer an overview of some of these approaches, and discuss the associated challenges in selecting the right model for a workflow language. Also, I will look into the applicability of workflows to computerized representation of clinical guidelines - an ongoing effort in medical informatics that aims to achieve verifiable medical software tools, such as clinical decision support systems. The experiences from the workflow field may significantly contribute here, and I will discuss the shared challenges.

March 1, 2011 2pm, CLT 701
UK Smart Metering and Home Automation
Prof Andy Stanford-Clark, IBM

Prof Andy Stanford-Clark, IBM's Chief Technology Officer for Smarter Energy, will talk about the emerging landscape of smart metering in the UK, its aims and directions, and will discuss the infrastructure and communications challenges around the roll-out of 47 million domestic smart meters. He will go on to talk about technology and approaches to changing consumer behaviour in the home, with examples and experiences from his own automated home ("The House that Twitters").

February 25, 2011 2pm, CLT 701
Policy Languages For User-Managed Access
Dr Michael Huth, Imperial College London

User Managed Access promises systems in which users can control access to their resources with ease and precision. Policies are objects that specify who should get access to what resources under which circumstances. Policies written in formal policy languages obtain clean analysis and implementation methods from the formal semantics of the underlying language. In this talk, we explore important design principles for formal policy languages and indicate how they may be brought to bear in bespoke policy languages for User Managed Access.

February 15, 2011 2pm, CLT 701
Knowledge extraction from provenance of data: initial ideas for techniques and potential applications
Dr Paulo Missier

The term provenance, in relation to information products, refers to structured representations of metadata that describe how the information was produced, along with any available attribution information, i.e., the actors that participated in the production process. This broad definition applies to a wide variety of information products, ranging from the outcomes of scientific experiments (e.g. models, simulations), to self-published newsbites that propagate across the Web in the form of blogs, tweets, or other social media, to traditional citations in published literature. Although interest in provenance management is on the increase, so far efforts have largely focused on the basic data infrastructure, namely a consensus data model for provenance representation, along with the capability to observe and record provenance events from a variety of process types, for example workflows. My interest is in demonstrating the potential of large corpora of provenance traces as a rich source of knowledge, an area of data management research that I refer to as "provenance mining". Such knowledge can find its use in areas as diverse as process optimisation, derivation of quality properties of data, correction of data errors by partial process rerun, and systematic information sharing across science networks. In this talk I will provide some background on the state of the art in provenance management, and present initial ideas on applications of provenance mining, including to ongoing projects within the School.

February 1, 2011 2 pm, Room 701, Claremont Tower
Machine Learning Multi-parametric Classification of Traumatic
Dr. Benjamin Aribisala, Institute of Cellular Medicine, Newcastle University

Traumatic brain injury (TBI) is ranked as the fourth highest cause of death in the developed world. The majority of patients sustain mild TBI, and a significant number suffer persistent neuropsychological problems. Conventional neuroimaging methods (CT, MRI) do not reveal abnormalities consistent with the cognitive symptoms. Imaging methods offering prognostic information in acutely injured patients are therefore required. Here we applied advanced quantitative MRI techniques (T1, T2 mapping and diffusion tensor MRI) in 24 mild TBI patients and 20 matched controls. We applied a machine learning algorithm based on support vector machine (SVM) to classify the quantitative MRI data. Univariate classification was ineffective due to overlap between patient and control values, however multi-parametric classification achieved sensitivity of 88% and specificity of 75%.

January 13, 2011 2 pm, Room 701, Claremont Tower
Mining Flexible-Receptor Molecular Docking Data
Ana Trindade Winck, visiting researcher School of Computing Science

With the growth of biological experiments, solving and analyzing the massive amount of data being generated has been one of the challenges in bioinformatics, where one important research area is the rational drug design (RDD), centered on examining possible interactions between receptors and ligands, usually performed by molecular docking. Very little is known about the effectiveness of considering the flexibility of a receptor in current molecular docking simulations when this flexibility is modeled by a molecular dynamics (MD) simulation. This methodology generates vast amounts of data which need to be explored to generate useful information about receptor-ligand interactions. To better interpret these data we developed a comprehensive repository that integrates features of all snapshots from a MD simulation trajectory with related data about receptor-ligand interactions from the docking simulations results. Based on the receptor residues-ligand distances we preprocessed all docking results to generate appropriate input to mine data. Data preprocessing was done by calculating the shortest inter-atomic distances between the ligand and the receptor’s residues for each docking result, considering the estimated free energy of binding (FEB) value as the target attribute. By performing data mining applying regression decision trees algorithms on such data and conveniently postprocessed the induced models, were able to select a set of promising receptor conformations. Although we had reached important results, they can be improved. Hence, we propose an approach to predict data from MD simulations, based on previous docking simulations. We aim at evolving the M5P algorithm so that it can read and treat 3D properties and induce a proper model tree that predict probably best atoms regions that can produce good estimated FEB values.

November 16, 2010 2 pm, Room 701, Claremont Tower
Performance Benchmarking and Statistical Analysis of Response Time
Anatoliy Gorbenko. National Aerospace University. Ukraine

The talk reports a practical experience of performance measurement of the complex system biology web service BASIS, and investigates the instability of its behaviour and the delays induced by the communication medium. The results of statistical data analysis and finding distribution laws which fit and predict the response time instability typical of Service-Oriented Architectures (SOAs) built over the Internet are presented. The experimental work supports a claim that dealing with the uncertainty inherent in the very nature of SOA and WSs is one of the main challenges in building dependable service-oriented systems. In particular, this uncertainty exhibits itself through very unstable web service response times and Internet data transfer delays that are hard to predict. Our findings indicate that the more experimental data is considered the less precise distributional approximations become. The talk concludes with a discussion of the lessons learnt about the analysis techniques to be used in such experiments, the validity of the data, the main causes of uncertainty and possible remedial actions.

November 15, 2010 2 pm, Room 701, Claremont Tower
From biochemical species to pathway to tissue as process
Professor Muffy Calder, School of Computing Science, University of Glasgow

Over the last decade, computer scientists have explored the application of process algebraic concepts to modelling bio-molecular processes, with the main abstraction being molecule-as-process (Regev, Cardelli). We have adopted an alternative abstraction: species-as-process and a reagent-centric style of modelling that offers a distributed view of a system. The system is easily represented in a state-based formalism where variables represent levels of species and all reagents (consumers, producers, catalysts etc.) synchronise (multi-way) on reaction events. The underlying semantics is continuous time Markov chains. Two further abstractions are described. Pathway-as-process allows us to investigate how intracellular signalling pathways interact with each other (known as cross-talk), and how we can detect and characterise the interactions. Tissue-as-process allows us to investigate how species concentrations interact with spatial aspects to produce patterns at a tissue level, and how the same pattern can arise from different biochemistries. The overall motivation of this work is to model mechanisms underlying diseased and non-diseased pathways, and possible interventions.

November 10, 2010 2 pm, 701, Claremont Tower
Introduction to UMA/j – User Managed Access framework
Maciej Machulak & Lukasz Moren

Web 2.0 technologies have made it possible to migrate traditional desktop applications to the Web, resulting in a rich and dynamic user experience and in expanded functionality. Similarly, RESTful Web services have become a commonly used model for exposing services and their APIs on the Web. As a result, a potentially large amount of personal, sensitive, and valuable data is put online, spread across various Web services. With the increasing number of resources and services accessible over the Web, software developers are forced to provide more sophisticated and more user-friendly authentication and authorization mechanisms. Support for delegated authentication, such as OpenID, is well-supported by existing frameworks and libraries. Authorization, on the other hand, lacks a comparable support and is commonly built into existing applications or delegated using protocols that are not suitable for the open Web environment. User Managed Access (UMA), a standardisation effort from the Kantara Initiative that is closely related to the recently emerging OAuth V2.0 protocol from IETF, proposes a new approach to access management for Web resources that includes a user as a core part of its model. UMA lets an individual control the authorization of data sharing and service access made between online services on the individual's behalf. It provides benefits for both users and application developers. The latter ones, in particular, can benefit from providing authorization mechanisms and sharing capabilities to their new or existing applications without the need of building these mechanisms by themselves. In this talk, we aim to review existing solutions to authorization for Web applications and Web services. We will introduce UMA and explain how it builds on such protocols as OAuth V2.0 and OAuth WRAP. We will then present UMA/j – a purely Java-based open-source framework that helps developers to delegate authorization from their Web applications and Web services to specialized software components. We will also demo an exemplar UMA system built with the UMA/j library that is planned to be deployed within Newcastle University to provide access control mechanisms for over 19,000 students and 4,500 staff members.

November 5, 2010 10 am, Room 701, Claremont Tower
Security Analytics - A Rigorous Approach to Security Management
Simon Shiu, Hewlett-Packard Laboratories, Bristol, UK

The evolving threat environment and increasing complexity are making risk governance and security management more challenging for large enterprises. Following a series of case studies with the security teams of some very large HP customers we have developed model based methodology and tools to improve security investment and policy decisions. In this talk I will give examples of our model based approach. I will also describe ongoing work looking at validation of the approach, the way we are extending it to look at security governance challenges with cloud computing, and trusted virtualization.

October 7, 2010 13.00, Room 701, Claremont Tower
Can we use network analysis methods to discover functionally important method calls in software systems by considering dynamic analysis data?
Anjan Pakhira

It is difficult to analyse large-scale integrated software systems with the purpose of improving their functionality through maintenance and evolution. Such systems contain many interactions between their components and can be represented as complex interaction networks similar to complex biological and socio-technical systems. Here we aim to check whether the combination of dynamic analysis and network analysis can determine method calls of high functional importance in a software system. We use as a test case the JHotDraw 6.01b software and predict the method calls with high functional importance using network analysis methods. We validate the predictions by disabling the methods predicted to have high functional importance and evaluating the behaviour of the software following this. Our results indicate that some of the considered network analysis methods are relatively good in predicting method calls of high functional importance.

August 10, 2010 14:00 - 15:00, Room 701, Claremont Tower
Cheating and Fairness in First Person Shooters: Doing science under fire
Mike Bond, University of Cambridge

First person shooters make up a large proportion of games releases in the console and PC entertainment markets, which gross more than the film industry nowadays. Multiplayer support is at its heart, but multiplayer games lose their fun when cheats hack the game, or unfairness in the game causes peoples tempers to fray. But far worse than cheating/unfairness is **perceived** cheating and unfairness, and for this the games makers are to blame. This talk opens a window on how multiplayer games work under the hood, and explores how the tricks the games studios use to create the sensation of participation in shared world over imperfect network connections end up creating as many problems as they solve. In particular, the talk discusses "Neo-tactics": game strategies for beating opponents evolved from low-level network properties of the game [1]. The talk ends with an academic challenge. The fundamental assumptions of science include these axioms: (i) there is an independent reality which we can all observe and understand (ii) our senses and observations present a coherent and accurate view of that reality Multiplayer games deliberately break both of these: they present different realities to different players, and they deliberately try to trick the players to believe false things, in order to cover up the effects of network latency. How then can we safely conduct experiments in these worlds?

August 5, 2010 14:00 - 15:00, Room 701, Claremont Tower
Probabilistic Aspects of Flash Filestores
Zoe Andrews

Flash filestores have a variety of unique features that lead to interesting design constraints. One challenge is that of ensuring that each block of the drive is used and erased evenly, known as "wear-levelling". This talk presents a novel approach to the analysis of wear-levelling algorithms using probabilistic specification and analysis techniques. A simplified version of a real wear-levelling algorithm used in an actual flash filestore is given as an illustration. The expected lifetime of a flash filestore implementing such an algorithm is derived using probabilistic proof techniques.

July 26, 2010 14:00 - 15:00, Rm. 701, Claremont Tower
Location-Aware Distributed Sensor-Actuator Systems
Prof. Dr. Matthias Kranz

Ubiquitous Computing is characterised by a huge variety of heterogeneous, distributed sensing and actuator systems - including sensing systems (such as wireless sensor nodes, mobile phones, and intelligent environments and cities) as well as actuators such as ambient information systems or pro-active agents. Location is one of the most important cues for context-awareness in such distributed sensor-actuator systems. In the first part of the talk we will hence discuss selected aspects of location-awareness for distributed sensor-actuator systems. We will then present an approach that uses off-the-shelf cordless telephones - technology that has become ubiquitous and now remains completely invisible in the background - for indoor localisation. This will lead on to an application scenario for indoor localisation and navigation that will illustrate how position, navigation and interaction design can be combined to develop useful and usable ubiquitous computing systems for human users. At the end of the talk we will briefly discuss distributed sensor-actuator environments, which are equally inhabitable by human users and by robots.

July 6, 2010 2pm, Room 701, Claremont Tower
Service Availability Standards for Carrier-Grade Platforms
Dr. Francis Tam, Nokia Research Centre, Finland

Mobile network equipment manufacturers have been exploiting carrier-grade platforms as reusable assets to provide various support functions in the products creation. The term carrier-grade refers to a class of systems used in public telecommunications network that deliver five or six nines (99.999% or 99.9999%) availability. The convergence of communications and information technology in the industry has led to more competition and pressure to reduce development efforts. By creating a service availability standard, off the shelf software can be integrated into a carrier-grade platform, enabling a company to focus on the core business and concentrate the resource investment onto new innovations. This talk explores the kind of service availability support needed for such a carrier-grade platform. Service Availability Forum's Availability Management Framework and Software Management Framework standards will also be introduced. About the speaker: Dr. Tam is a Principal Engineer at Nokia Research Centre in Finland. He is currently undertaking a visiting fellowship in the School of Computing Science.

July 1, 2010 14.00, Room 701, Claremont Tower
Code Mobility Mechanisms in Managing Internet Resources and the Role of Behavioural Contracts
Dr. Mario Bravetti, University of Bologna

Constructs involving code mobility, such as client-side application execution typical of web 2.0 and service deploy/workflow update mechanisms that allow for dynamic reconfiguration, are at the heart of modern internet resource providers and play a fundamental role in service oriented and cloud computing. In this context, it is important to base service level agreement between providers and users on formal behavioural specifications and to have a theory for establishing behavioural compatibility of such behavioural specifications.Technically, we present a middleware for managing files/resources and executing programs (in the form of client-side/server-side components/services) in Web Operating Systems and decidability results based on process algebraic representations of service interface behaviours accounting for workflow update mechanisms.

June 24, 2010 2pm, Room 701, Claremont Tower
On Fault Tolerance Modelling and Reuse during Refinement
Ilya Lopatkin

Complex modern applications have to be developed to be dependable to meet their requirements and expectations of their users. An important part of this is their ability to deal with various threats (such as faults in the system environment, operator's mistakes, underlying hardware and software support problems). Development of modern applications is complicated by the need for systematic and rigorous integration of fault tolerance measures. In our talk, we will focus on reuse of fault tolerance modelling. Firstly, I will introduce a modelling approach which captures the user's view on system behaviour with respect to faults. The approach has a beta-version tool support with a small case study in Event-B. Secondly, I will show ideas in progress on using libraries of Event-B modelling patterns allowing developers to systematically integrate specific fault tolerance mechanisms (e.g. recovery blocks, checkpoints, exception handling) into the models.

June 17, 2010 2pm, Room 701, Claremont Tower
Its Cloudy at IBM
Dr Graeme Dixon, IBM

This talk will cover the many challenges (technical and non-technical) we face at IBM as we work to build out our cloud computing support. It will describe a number of announced or delivered products that are our first steps at providing private and public cloud support and how we are reconciling these with future hardware and software products and leveraging them to build dev/test clouds and publicly accessible clouds. About the speaker: Dr. Dixon is CTO and Chief Architect, STG Platform Management at IBM. He leads a large development team in the IBM's Systems and Technology Group (ie. hardware division) building the core of most of IBM's cloud computing products and works closely with the software and services divisions on cloud and many other areas of technology.

June 10, 2010 2pm, Room 3.13, Daysh Building
Text Retrieval support for Concept Location in Software
Dr Andrian Marcus, Dept of Computer Science, Wayne State University, Detroit, Michigan,

Abstract: Concept (or feature) location in software is the process of identifying the parts of the source code that correspond to a specific functionality. This process is a prerequisite to program comprehension and is one of the most common activities undertaken by developers during software evolution. We are proposing an approach to concept location in software based on indexing the source code and additional data using Text Retrieval methods. This approach allows users to formulate queries in natural language and obtain results in form of software components related to the query. One of the problems common to all concept location techniques is filtering and ranking the results such that relevant parts of the source code are quickly identified by the developers. To address this problem we propose three separate approaches, one that clusters the results of a search using formal concept analysis, one that merges the results with execution traces, and one where the search results are used in conjunction with software dependencies. The advantages and disadvantages of each approach are discussed with respect to results obtained from empirical studies on existing large software systems. Bio: Dr. Andrian Marcus is associate professor in the Department of Computer Science at Wayne State University in Detroit, USA. He received his PhD in Computer Science from Kent State University, USA in 2003. His research interests include software evolution, program comprehension, and software visualization, focusing on the management of unstructured information during evolution of large scale software systems. Dr. Marcus served on the Steering Committee of the IEEE International Conference on Software Maintenance (ICSM) between 2005-2008 and on the steering Committee of the IEEE International Workshop on Visualizing Software for Understanding and Analysis (VISSOFT) since 2007. He is the Program Co-Chair of the 17th IEEE International Conference on Program Comprehension (ICPC 2009) and the Program Co-Chair of the 26th IEEE International Conference on Software Maintenance (ICSM 2010). His research is currently funded through grants from the US National Science Foundation, the US National Institute of Health, IBM, etc. Dr. Marcus' publications earned a Best Dissertation Paper Award at the IEEE ICSM in 2004 and two Best Paper Awards at the IEEE ICPC in 2006 and 2007 respectively. He is also the recipient of a Fulbright Junior Research Fellowship in 1997 and the US NSF CAREER Award in 2009.

May 20, 2010 14.00, Room 701, Claremont Tower
Assessing the Attack Resilience Capabilities of a Fortified Primary-Backup System
Dylan Clarke

Primary-Backup service replication does not constrain that services must be built as deterministic state machines. It is meant to tolerate crashes, not intrusions. We consider an approach, called FORTRESS, for adding intrusion-resilience capability to a primary-backup server system. It involves using proxies that block clients from directly accessing servers, and periodically randomizing the executables of proxies and servers. We argue that proxies and proactive randomization can offer sound defense against attacks including de-randomization attacks. we use simulations to compare the attack resilience that FORTRESS adds to a primary-backup server system with that attainable through state machine replication (SMR) that is fit only for deterministic services.

April 29, 2010 14.00, Room 3.13, Daysh Building
Computational Approaches to Synthetic Biology
Goksel Misirli

Large-scale engineering of biological systems requires computational methods. Modelling and the use of data integration are of high importance in this process. As we try to create biological parts with standard interfaces in vitro, having complementary virtual parts in the form of in silico mathematical models are highly desirable. When annotated properly, the model constituents can be used to derive the final sequence for the fabrication. As the data integration could be used to inform these models, it could also be used to constrain the huge design space and eliminate unnecessary evaluations. Furthermore, semantically well-structured integrated data could be used to reveal new information.

April 1, 2010 14.00, Room 701, Claremont Tower
Privacy preserved event correlation for the MASTER monitoring infrastructure
Su-Yang Yu

MASTER infrastructure is designed to monitor compliance level of existing SOAs. A layer of signaling components are used as wrappers around existing services, they generate raw events based on the status of these services and send them to a monitoring layer. The monitoring layer is a network of complex event processors (CEP), they correlate these raw events in order to produce high level events. These high level events are then processed by higher level layers (assessment and enforcement) to evaluate overall compliance level of the business process, and if necessary perform correctional feedback to the SOA(s). Due to the process intensity of CEPs, any security solutions should be lightweight. However, traditional cryptographic techniques require the CEPs to first perform decryption on all messages before there can be any correlation. For this, we surveyed existing privacy preservation techniques, and present a set which allow specific event correlations to be performed directly on their resultant cipher-texts.

March 25, 2010 1pm, Devonshire G21/22
Optimistic Fault Models for Low-Cost Byzantine-Fault Tolerant Systems
Marco Serafini

We are increasingly relying on computer systems to carry out critical tasks with very high degrees of dependability. For example, safety-critical systems have long been used in the aerospace domain. The trend towards storing and managing personal information online is extending the need for dependable services to a growing range of applications, from emailing, to calendars, to the storage of photos. High dependability requirements can be met with higher assurance by systems that are able to tolerate worst case failures, which are often named as Byzantine. Byzantine Fault Tolerance (BFT) is required by certification authorities for most safety-critical applications and is being advocated for critical Internet services too. BFT can be achieved by replicating logical functionalities over multiple physical nodes of a distributed system. However, the entailed replication costs are quite high. This talk points out that it is possible to reduce these costs by using optimistic fault models. In fact, Byzantine faults represent the exception rather than the rule since faults are often of more benign nature. The talk will review three modern technology trends which can be leveraged for cost reduction using optimistic fault models. The first trend is the foreseen raise in the ratio of transient faults compared to permanent faults in modern hardware due to reduced geometries and lower voltage levels. The second is the development of trusted nodes which only fail in a restricted, benign manner. The third is that replication clusters are relatively small, even in large-scale systems. The talk will then focus on the last trend and present a novel algorithm, called Scrooge, which can be used to achieve low-cost BFT in large scale or geographically-distributed systems.

March 18, 2010 2pm, CLT.701
CCSdp: A Process Algebra for Modelling and Analysis of Dynamic Reconfiguration with Interference
Ani Bhattacharyya

The traditional way of modelling dynamic reconfiguration of a system is to use two formalisms: one to model computations by the system, and the other to model the system's structural change. This approach is valid if the reconfiguration is instantaneous, or if computations are suspended or aborted during the reconfiguration interval. However, the approach is unsuitable for systems where availability of service is a critical requirement, so that computation and reconfiguration activities must proceed concurrently; which raises the issue of their interference. In this talk, I will describe the first version (CCSdp) of a new process algebra I am developing for modelling and analysis of both computations and process reconfiguration, explain the rationale behind its design, and identify the issues still to be resolved.

February 25, 2010 3pm, CLT.701
Consistent Diagnosis for Distributed Embedded Systems.
Raul Barbosa, Chalmers University of Technology, Gothenburg, Sweden

The presentation will describe a protocol for consistent diagnosis in distributed systems and explain the usage of formal verification to ascertain its correctness. The protocol allows nodes in a synchronous system to maintain consensus on the set of operational nodes, i.e., the membership, in the presence of omission failures and node restarts. It relies on nodes observing the transmissions of other nodes to detect failures. Consensus is maintained by exchanging a configurable number of acknowledgements for each node's message; increasing this number makes the protocol resilient to a greater number of simultaneous or near-coincident failures. The SPIN model checker was used to formally verify the correctness of the protocol. The presentation will describe how the protocol was modeled and provide the results of the exhaustively verified model instances.

February 15, 2010 2pm, Room 701, Claremont Tower
High quality P2P service provisioning with decentralized trust management
Le-Hung Vu

Trust management is essential in fostering cooperation and high quality service provisioning in several peer-to-peer (P2P) applications. Among those applications are customer-to-customer (C2C) trading sites or markets of services implemented on top of centralized systems, P2P, and online social networks. Under these application contexts, existing work does not really address the heterogeneity of the problem setting in practice. This heterogeneity includes the different approaches employed by the participants to evaluate trustworthiness of their partners, the diversity in contextual factors that influence service provisioning quality, as well as the variety of possible behavioral patterns. This talk presents the design and usage of appropriate computational trust models to enforce cooperation and ensure high quality P2P service provisioning, considering the above heterogeneity issues. First, I propose a graphical probabilistic framework for peers to model and evaluate trustworthiness of the others in a highly heterogeneous setting. The framework targets many important issues: the multi-dimensionality of trust, the reliability of different rating sources, and the personalized modeling and computation of trust on a participant based on the quality of services it provides. Next, an analysis on the effective usage of computational trust models in environments where participants exhibit various behaviors, e.g., honest, rational, and malicious, will be presented. I provide theoretical results showing under which conditions cooperation emerges when using trust learning models with given accuracy and how cooperation can still be sustained while reducing cost and accuracy of those models. As a another contribution, I designed and implemented a general prototyping and simulation framework for reputation-based trust systems. The developed simulator can be used for many purposes, such as to enable empirical simulation to discover new phenomena or to test various hypotheses in complex settings. Two potential applications of computational trust models are then discussed: (1) the selection and ranking of (Web) services based on quality ratings from reputable users, and (2) the use of a trust model to choose reliable delegates in a key recovery scenario in distributed social networks. Finally, I will identify a number of various issues in building next-generation reputation-based trust systems as a future direction of my research. Note: This talk comprises the result of my PhD research at EPFL over recent years in collaboration with my colleagues from several institutions.

December 3, 2009 1pm, Room 701, Claremont Tower
Model-based Dependability Analysis & System Architecture
Yiannis Papadopoulous, University of Hull

The increasing scale and complexity of safety critical systems pose significant challenges in the safety assessment of such systems which becomes increasingly more expensive, error prone and difficult to complete. To address these challenges, the past fifteen years research has focused on automating the synthesis of predictive models of system failure from design representations. In one approach known as compositional safety analysis, system failure models such as fault trees and Failure Modes and Effects Analyses (FMEAs) are being constructed from the topology of a system and component failure models using a process of composition. In this seminar, I focus on Hierarchically Performed Hazard Origin and Propagation Studies (HiP-HOPS) - one of the more advanced and well supported compositional safety analysis techniques. I discuss the principles that underpin this technique and show how these principles have been effectively used to deliver a number of scientific and practical contributions which include: - Fast algorithms for automatic synthesis of Fault Trees and multiple failure mode FMEAs from design representations of a system. - An extension to the Fault Tree notation and Fault Tree Analysis (FTA) technique that enable assessment of the effects of sequences of faults. - A language for specification of inheritable and reusable component failure patterns. - Pareto-envelope genetic algorithms that enable automatic optimisation of system models in respect of dependability (i.e. safety, reliability, availability) and cost via application of automatic model transformations. I summarise those contributions and demonstrate application of HiP-HOPS on small examples discussing the strengths and limitations of the method. HiP-HOPS offers less automation that formal safety analyses techniques, e.g. those using model-checking. However, it is computationally less expensive, not prone to combinatorial explosion, scales up and can be easily iterated throughout the design lifecycle. For the same reason, it enables exhaustive assessment of combinations of failures and design optimisation using computationally greedy meta-heuristics. I conclude this discussion by pointing to the future of this work as I outline current plans with Newcastle towards harmonisation and integration of HiP-HOPS with two architecture description languages emerging in the Avionics and Automotive industries.

October 29, 2009 2pm, CLT 701
E-Surface: An Exploration of Tabletop Interaction for E-Science
Tom Bartindale

E-Science annotation and collaborative investigation using tabletop computing. Demonstration of software on the Surface, and a short introduction and tutorial on Surface development and the Surface SDK. Tom Bartindale is a PhD student within our school http://www.cs.ncl.ac.uk/people/t.l.bartindale

October 15, 2009 14.00, Room 701, Claremont Tower
Capturing process: The promise and challenges of connecting experimental records to the semantic web
Dr Cameron Neylon

The promise of the semantic web in its broadest terms to integrate structured data, and in particular scientific data, is breathtaking. Yet our ability to deliver on that promise remains severely limited. On one side, battles are fought over vocabularies and structured description languages, while on the other computer scientists argue over design and architecture. In the middle the experimental scientist is usually left bemused as to what all the fuss is about and how it can have any relevance to their research, which is naturally far too specialised and ground-breaking to fit into any framework that an outsider could ever provide. The challenge lies in finding a pathway that moves towards satisfying the needs and interests of these disparate groups while providing both value and a usable interface to the end user. We have adopted a blog based approach to capturing the record of research in an experimental laboratory. The use of a free form framework along with templates encourages the user to structure the description of their work in a way that matches their needs. Cycles of template design and optimisation lead to structures with strong similarities, and occasional important differences, with designed ontologies and vocabularies. The challenge for the future lies in determining how best to guide the actions of users towards the use of appropriate controlled vocabularies and structured descriptions, while using their choices to inform the design of these systems. The potential for Google Wave to provide a framework to enable local structuring as well as integration with global structured descriptions will be explored. About Cameron Neylon: Dr. Neylon is the author of the "Science in the Open" blog (http://blog.openwetware.org/scienceintheopen/). He is a proponent of open science, saying that "[m]any people find this scary. Some, perhaps a growing number, find it tremendously exciting" (http://blog.openwetware.org/scienceintheopen/about/). He works at the the Science and Technology Facilities Council (http://www.scitech.ac.uk/), the UK’s major provider and supporter of large scale academic research facilities, including synchrotrons, neutron sources, and high powered lasers. He is also an Academic Editor on PLoS ONE, and has contributed to PLoS Conversations (http://everyone.plos.org/2009/07/29/introducing-plos-conversations-user-generated-videos/), among many other things.

October 6, 2009 2pm, CLT 701
Task Modelling with HOPS
Anke Dittmar

Task modelling is a "traditional" technique in human-computer interaction (HCI). It is assumed that people use mental knowledge structures to perform tasks and that these structures can be elicited (or "induced") and then be represented as cognitive task models. Task models are used for analysing, evaluating and designing interactive systems. The talk gives a short overview of prevalent task modelling concepts and how they are applied in HCI. However, though current approaches claim to support a user-centred design they are often criticized for explaining actual working practices in a poor way. Some of the limitations are discussed in more detail. HOPS (Higher-Order Processes Specifications) is introduced as a general specification formalism to describe interaction. Its applicability to more advanced task modelling ideas is demonstrated.

September 24, 2009 3pm, 701
NetFPGA: An open platform for high-speed data processing
Andrew Moore

The NetFPGA is an open, multipurpose platform that provides a complete development tool chain for researchers and students to implement their ideas easily and build prototypes of high-speed, hardware-accelerated data processing systems. The NetFPGA card includes all of the logic resources, memory, and Gigabit Ethernet & SATA interfaces necessary to build a complete switch, router, and/or network security device. The NetFPGA is a low-cost reconfigurable hardware platform optimized for high-speed networking. Because the entire data path is implemented in hardware, the system can support back-to-back packets at full Gigabit line rates and has a processing latency measured in only a few clock cycles.

July 23, 2009 14.00 , Room 701, Claremont Tower
An Introduction to IDMAPS: University Data Flows & Personalisation
Sunil Rodger

IDMAPS (Institutional Data Management for Personalisation and Syndication) is a JISC-funded project currently being undertaken by ISS in Newcastle University. It aims to solve a common problem affecting large organisations: how to best manage and standardise data flows between the myriad of different systems which underpin its day-to-day operations. This will improve the reliability and quality of the data upon which these systems depend. It will also allow "Web 2.0"-style mashups to be developed, delivering a level of personalised content to end users which would currently be impossible to achieve.

July 2, 2009 2pm, Room 701, Claremont Tower
Structuring Specifications with Modes
Dr Fernando Luís Dotti

Several safety-critical systems, such as avionic, transportation and space systems, use the notion of operation modes. Operation modes are useful structuring units that facilitate design, specially if used with state-based formal methods. In this work we discuss abstractions for the specification of modal systems, modal systems refinement, and their relation to the specification method Event-B. The use of modes for fault tolerant systems is briefly discussed and a case study presented. Dr Fernando Luís Dotti is from the Catholic University of Rio Grande do Sul, Brazil, and is now on sabbatical with Newcastle University.

June 25, 2009 2pm, Room 701
The development of an integrated research Management System - MyImpact
Jill Golightly, Head of University Research Office

MyImpact is being developed as a central system to replace (or enhance) MyProfiles, to act as a source of publication information (and will be used to support the next RAE), with e-prints, an e-repository, and maybe other things that might be covered in this talk. CS is very likely to be using MyImpact to replace the School's local database of publications, with all staff being required to use it for all of their publication data. We are looking at how to transfer the existing publication database into the new system, but this talk will give us a chance to find out more about MyImpact and what it is going to provide, and to provide feedback on what is being proposed/developed.

June 4, 2009 2pm , Room 701, Claremont Tower
Structured Occurrence Nets
Brian Randell (joint work with Maciej Koutny)

This talk introduces the concept of a `structured occurrence net', which is based on that of an `occurrence net', a well-established formalism for an abstract record that represents causality and concurrency information concerning a single execution of a system. Structured occurrence nets consist of multiple occurrence nets, associated together by means of various types of relationship, and are intended for recording either the actual behaviour of complex systems as they communicate and evolve, or evidence that is being gathered and analysed concerning their alleged past behaviour. The talk discuss how structured occurrence nets could form a basis for improved techniques of system failure prevention and analysis.

May 7, 2009 14.00, Room 701 Claremont Tower
Characterising brain disease through wave and network analysis of
Jennifer Simonotto

Our group¹s interest is to unravel dynamics and characteristics of neural networks; thus, our aim is to develop tools to characterize and quantify these properties. A variety of neuroscience problems centre around excitable media or networks that differ from Onormal.¹ In the case of retina disease, the wave properties and resultant dynamical networks of bursts across the developing mouse retina are different between wild-type and the Crx mouse (a gene knockout animal model for retinitis pigmentosa). In the case of epilepsy, there are high frequency events (ripples and fast ripples) that occur only in epileptic tissue; these events are useful for focal area identification. I will present an overview of excitable media neural systems, and talk about methods we use to extract and characterise retinal signals as well as epilepsy signals, including extracting wave characteristics, extracting and analysing dynamical networks and looking at frequency criteria for feature identification. In addition to algorithm implementation and data analysis, our group has a great need to speed up computation time in any way we can. A tool for automatic data partition and job submission on Condor or other high-performance computing environments was developed and tested.

April 30, 2009 14.00, Room 701, Claremont Tower
'Research: metrics, quality and management'
Dr Peter Andras

Research metrics is a current hot topic following the recent RAE and in preparation for the future metrics-oriented REF that will be used to allocate research funding to universities. In this talk we will discuss a range of research metrics that are used and considered as part of research assessment, the realtionship between such metrics and the quality of the research, and the possible uses of research metrics in research management. In particular we will discuss the distinction between 'normal' and 'revolutionary' science in the sense of Kuhn, and the use of metrics in the context of these two science practice variants.

March 19, 2009 14.00, Room 701 Claremont Tower
Benchmarking Dependability of a System Biology Application
Dr Anatoliy Gorbenko

In this talk we report our practical experience in benchmarking a System Biology Web Service, and investigate instability of its performance and the delays induced by the communication medium. We discuss the results of a statistical data analysis and discuss the causes affecting the Web Service performance. The uncertainty discovered in Web Services operations reduces the overall dependability of Service-Oriented Architecture and require specific resilience techniques. >

December 4, 2008 13.00, CLT 701
A low-cost attack on a Microsoft captcha
Ahmad El-Ahmad

CAPTCHA is now almost a standard security technology. The most widely deployed CAPTCHAs are text-based schemes, which typically require users to solve a text recognition task. The state of the art of CAPTCHA design suggests that such text-based schemes should rely on segmentation resistance to provide security guarantee, as individual character recognition after segmentation can be solved with a high success rate by standard methods such as neural networks. In this paper, we present new character segmentation techniques of general value to attack a number of text CAPTCHAs, including the schemes designed and deployed by Microsoft, Yahoo and Google. In particular, the Microsoft CAPTCHA has been deployed since 2002 at many of their online services including Hotmail, MSN and Windows Live. Designed to be segmentation-resistant, this scheme has been studied and tuned by its designers over the years. However, our simple attack has achieved a segmentation success rate of higher than 90% against this scheme. It took on average ~80 ms for the attack to completely segment a challenge on an ordinary desktop computer. As a result, we estimate that this CAPTCHA could be instantly broken by a malicious bot with an overall (segmentation and then recognition) success rate of more than 60%. On the contrary, the design goal was that automated attacks should not achieve a success rate of higher than 0.01%. For the first time, this paper shows that CAPTCHAs that are carefully designed to be segmentation-resistant are vulnerable to novel but simple attacks.

December 1, 2008 1pm, Claremont Tower Room 701
Innovating Integrated Healthcare through Semantic Interoperability
Ziv Ofek, CTO – dbMotion

Innovating Integrated Healthcare through Semantic Interoperability Over the last few years, we have witnessed numerous initiatives by healthcare organizations to share medical information, using various models. Few of these initiatives, if any, have borne fruit in a wide operational scope and several have ended in failure, after years of effort and substantial financial investment. Experience shows that a significant proportion of the existing models for sharing information (Messaging, Interface Engines, ETL, etc) do not provide a suitable foundational solution for the various challenges that the medical world faces today: improving quality of treatment, reducing costs and medical errors, providing tools to monitor and control medical services, and so on. The reason for this lies in the fact that this solution is not a simple one and comprises many diverse issues: technologies, ethics, business concerns, information ownership, medical confidentiality, standards, data vocabulary, patient identification, maintenance, support, and so forth. The lack of success and past experience have led to a new outlook known as Interoperability: the ability to share services between different factors (doctors, systems, processes) in a manner that enables them to use the services in real time and to incorporate them in their various work environments and processes (Web browsers, Order Entry processes, medical files, etc). In my speech, I will describe this new outlook and will briefly showcase technology that provides an advanced and encompassing solution to the many challenges. In addition, I will describe Semantic Interoperability which provides a broad and complete foundation for managing the coming years’ challenges from a multidisciplinary perspective of both the customers (CIO, CMO, CEO, CFO, Quality Managers, etc) and the challenges (integration of medical information at the treatment location, continuity of medical care, management of medical quality, medical research, decision-making support, etc).

October 22, 2008 2pm, 701 Claremont Tower
A Rule-based Notation to Specify Executable Electronic Contracts
Massimo Strano

We presents a notation to specify executable electronic contracts to monitor compliance and/or enforcement of business-to-business interactions. A notable feature is that the notation takes into account the distributed nature of the underlying computations by paying due attention to timing and message validity constraints as well as the impact of exceptions/failures encountered during business interactions.

October 16, 2008 14.00, Room 701, Claremont Tower
Formal Specification and Analysis of Distributed Systems with Graph Grammars
Fernando Luis Dotti

The design of distributed systems, assuring the required functional properties, can be a complex task. Formal methods are frequently employed, however they often are non trivial to use. Due to its graphical notation, allowing intuitive constructions, and simple concepts we have experienced the use of Graph Grammars as specification formalism for distributed systems. A restricted form of Graph Grammar focused on object-based distributed systems, called Object-Based Graph Grammars (OBGG), has been defined and used for the specification of several distributed algorithms. An existing translation from OBGG to PROMELA allows one to analyse OBGG models through model-checking using the SPIN tool. In an additional development, we have shown the suitability of this specification formalism to the analysis of distributed systems in the presence of some types of faults. More specifically, we have used existing concepts which allows one to, starting from a system model, derive a model representing the behavior of the same system in the presence of a specific fault. The obtained model can be analysed using the same methods and tools. In case the original system supposedly supported some fault-tolerance, analysis could be carried out to check the effectiveness of the mechanisms.

October 8, 2008 14.00, Room 701 Claremont Tower
Software Engineering Then and Now
Brian Randell

"Software Engineering Then and Now" by Brian Randell This year has been the 40th Anniversary of the NATO Conference which, through its provocative title "Software Engineering", is now viewed as having given rise to this "discipline". In consequence I have found myself giving invited presentations at various conferences about my recollections of the 1968 conference, and my views on how software engineering has progressed in the ensuing years. This SRG talk is based on the presentations I gave at ICSE in Liepzig and COMPSAC in Turku.

October 7, 2008 2pm, Room 701 Claremont Tower
An Integrated Knowledge Base for European Dependability Research
Hugh Glaser

A large integrated knowledge base, incorporating the entirety of the DBLP and Citeseer CS publications listings, the CORDIS and NSF project listings for the EU and US respectively, and the entire list of articles in the RISKS Digest, plus information from all the members, has been constructed for the ReSIST (Resilience for Survivability in IST) Network of Excellence on Resilience. This knowledge base, an example of the so-called Semantic Web, gathers together and integrates over 60 million items of information (of the form "subject-predicate-object") about people, publications, projects, research areas, etc. One of the principle applications running on this knowledge base is RKBexplorer - which provides a multi-facetted browser window via which users can explore for example how different people are related (e.g. through co-authorship of papers, or co-membership of projects), which publications are are most closely related to which others (e.g. by authorship, citations, journal), etc. This talk will include a demonstration of the RKBexplorer, and discuss how such design and implementation problems as the fact that different information sources often use somewhat different versions of a given person's name, have different ways of representing information, etc., have been successfully addressed. (Given the generic nature of the information sources used, RKBexplorer should be of interest to computer science researchers generally, even)

September 10, 2008 1pm, 701 Claremont Tower
The parameter approximation of Phase-type distributions for modelling and fitting the telecommunication traffic data
Dr Kyungsup Kim

The behaviour of traffic sources has been studied in order to support the effective use of telecommunication. The aim of this presentation is to find proper models to describe strange network traffic properties such as heavy tailed and self-similar, and efficient algorithms to fit the observed data. We use the Markovian processes to model the traffic properties. We survey phase-type distribution and Markovian arrival processes in aspect of how to use model in traffic source and how to fit to measured data. Generally phase-type representation has a lot of redundant parameters. So the model for fitting should be restricted to some special subsets with the minimal number of parameters like hyper Erlang or Coxian distribution, unicyclic distribution etc in order to reduce the complexity of computation and modelling. We note that the general Phase-type distribution can be transformed into a unicyclic canonical representation with the minimal of parameters. We present the EM (expe! ctation maximization) fitting method to a unicyclic representation form to describe the empirical measured traffic. Finally, we discuss on the properties and limits of Markovian fitting for the non-Markovian behavior like telecommunication traffic. Dr Kyungsup Kim, a visitor in the School from Korea.

July 31, 2008 2pm, 701 Claremont Tower
Power and Performance?
Joris Slegers

In this talk I will present some ideas on balancing power and performance in systems of servers. We will assume these servers can be powered down which introduces power savings, but at the same time deteriorates performance. Modelling these issues will allow us to make informed, good and perhaps even optimal decisions on when to power down or power up servers. The talk will be very preliminary since it is my intention to submit a proposal for an EPSRC post-doc fellowship on this topic. I will outline my ideas on what I want to model and roughly how. The floor will then be opened for all comments and feedback. This should greatly improve my proposal and thus my chances of getting the fellowship, which in turn means our department will get 3 years of free research.

July 24, 2008 2pm, 701 Claremont Tower
Trends in Software - Middleware and Applications for the Web. The CommunityOne/JavaOne Perspective.
Maciej Machulak

In this talk I will present the current trends in software that have been presented and discussed at the CommunityOne 2008 and JavaOne 2008 conferences. I will talk about current advancements in the area of Web Scale Computing, Web Applications, Enterprise Integration, SOA and ROA and many more. I will discuss the role of Java in those areas and how Java plans to compete with Very High Level Languages (e.g. Ruby). The talk will cover a lot of different topics in not too much details - "a little about a lot instead of a lot about a little". References to presentations from both conferences will be provided for every topic that I'm going to present.

April 24, 2008 2pm, Room 701 Claremont Tower
Semantic Integration in the COntext INterchange (COIN) Project and Relationship to the Semantic Web
Dr. Stuart Madnick

Many online services access a large number of autonomous data sources and at the same time need to meet different user requirements. It is essential for these services to achieve semantic interoperability and integration among these information exchange entities. A serious problem is the existence of heterogeneous contexts, whereby each SOURCE of information and potential RECEIVER of that information may operate with a different context, leading to large-scale semantic heterogeneity. A context is the collection of implicit assumptions about the context definition (i.e., meaning) and context characteristics (i.e., quality) of the information. As a simple example, whereas most US universities grade on a 4.0 scale, MIT uses a 5.0 scale – posing a problem if one is comparing student grades. Another typical example is extraction of price information from the Web: but is the price in Dollars or Euros (If dollars, is it US dollars or Hong Kong dollars), does it include taxes, does it include shipping, etc. – and does that match the receiver’s assumptions? In this talk, examples of important context challenges will be presented and the critical role of metadata, in the form of context knowledge, will be discussed. The Context Interchange (COIN) approach, inspired by similar goals of the Semantic Web, provides a robust solution. COIN is much more flexible and scalable compared with conventional approaches. With a given ontology, the number of conversions in COIN is quadratic to the semantic aspect that has the largest number of distinctions. These semantic aspects are modeled as modifiers in a conceptual ontology; in most cases the number of conversions is linear with the number of modifiers, which is significantly smaller than traditional hard-wiring middleware approach where the number of conversion programs is quadratic to the number of sources and data receivers. In an example scenario, the COIN approach needs only 5 component conversions to be defined while traditional approaches could require over 100,000 conversion programs. COIN achieves this scalability by automatically composing all the comprehensive conversions from a small number of declaratively defined component conversions. SOME SELECTED REFERENCES C.H. Goh, S. Bressan, S. Madnick, and M. Siegel, “Context Interchange: New Features and Formalisms for the Intelligent Integration of Information,” ACM Transactions on Office Information Systems, July 1999. [SWP #3941, CISL #97-03; http://web.mit.edu/smadnick/www/papers/J043.pdf] H. Zhu and S. Madnick, “A Lightweight Ontology Approach to Scalable Interoperability,” Proceedings of VLDB Workshop on Ontologies-based techniques for DataBases and Information Systems (ODBIS 2006), Seoul, Korea, September 11, 2006, pp. 45-54, [SWP #4621-06, CISL 2006-06, http://ssrn.com/abstract=926605]; also to appear in LNCS 4623, Springer Computer Science, 2007.

April 24, 2008 14:00-15:00, Room 701, Claremont Tower
Semantic Integration In The Context Interchange (COIN) Project And Relationship To The Semantic Web
Dr. Stuart Madnick

Many online services access a large number of autonomous data sources and at the same time need to meet different user requirements. It is essential for these services to achieve semantic interoperability and integration among these information exchange entities. A serious problem is the existence of heterogeneous contexts, whereby each SOURCE of information and potential RECEIVER of that information may operate with a different context, leading to large-scale semantic heterogeneity. A context is the collection of implicit assumptions about the context definition (i.e., meaning) and context characteristics (i.e., quality) of the information. As a simple example, whereas most US universities grade on a 4.0 scale, MIT uses a 5.0 scale - posing a problem if one is comparing student grades. Another typical example is extraction of price information from the Web: but is the price in Dollars or Euros (If dollars, is it US dollars or Hong Kong dollars), does it include taxes, does it include shipping, etc. - and does that match the receiver's assumptions? In this talk, examples of important context challenges will be presented and the critical role of metadata, in the form of context knowledge, will be discussed. The Context Interchange (COIN) approach, inspired by similar goals of the Semantic Web, provides a robust solution. COIN is much more flexible and scalable compared with conventional approaches. With a given ontology, the number of conversions in COIN is quadratic to the semantic aspect that has the largest number of distinctions. These semantic aspects are modeled as modifiers in a conceptual ontology; in most cases the number of conversions is linear with the number of modifiers, which is significantly smaller than traditional hard-wiring middleware approach where the number of conversion programs is quadratic to the number of sources and data receivers. In an example scenario, the COIN approach needs only 5 component conversions to be defined while traditional approaches could require over 100,000 conversion programs. COIN achieves this scalability by automatically composing all the comprehensive conversions from a small number of declaratively defined component conversions.

February 14, 2008 2pm, Room 701 Claremont Tower
Dimensions of Dynamic Virtual Organisations
Jeremy Bryans

Dynamic Virtual Organisations (or Dynamic Coalitions) arise in commercial, governmental, political and military environments. They are loosely-coupled groups of co-operating but autonomous agents, that share information and resources in pursuit of a common goal, which may be to take advantage of a business opportunity or respond to an acute humanitarian crisis. The DCWorkbench is a proof-of-concept tool to demonstrate the analysis of dynamic coalition policies and structures. It was designed and built by a team in Newcastle last year. A specific scenario, based on crisis management in a military context, was developed to exercise the workbench, and a case study based on observing a domain expert was carried out. In this talk I will present the DCWorkbench and discuss the formal model which underpins it. I will also discuss the case study and give a demonstration of the DCWorkbench.

January 17, 2008 2pm, Room 701 Claremont Tower
DEPLOY integrated project
Alexander Romanovsky

In this talk I will introduce the new FP7 DEPLOY IP which starts on February 2008. DEPLOY stands for Industrial deployment of advanced system engineering methods for high productivity and dependability. Its overall aim is to make major advances in engineering methods for dependable systems through the deployment of formal engineering methods. Partners: Newcastle U (Coordinator), Aabo Akademi (Finland), Bosch (Germany), Cetic (Belgium), Clearsy (France), ETH (Switzerland), Dusseldorf U. (Germany), Nokia (Finland), SAP (Germany), Siemens (France), SPace Systems (Finland), Systerel (France) and Southampton U. (UK). This project builds on the successful RODIN STREP http://rodin.cs.ncl.ac.uk/, which we have just completed and focuses on deploying its results in the real industrial setting in 5 major EU industrial sectors: automotive, rail transportation, space systems, business information and pervasive telecoms.

January 15, 2008 1pm, Room 701
Using inherent service redundancy
Dr. Anatoliy Gorbenko

Abstract: The presentation is devoted to means and solution ensuring dependability in Service-Oriented Architectures (SOAs). The SOA paradigm is a further step in the evolution of the well-known component-based system development with Off-the-Shelf components. The main motivation for this work is the fact that ensuring and assessing dependability of complex service-oriented systems is complicated when these systems are dynamically built or when their components (i.e. Web Services) are dynamically replaced by the new ones with the same (or similar) functionality but unknown dependability characteristics. The lack of sufficient evidence about the characteristics of the communication medium, components used in the composition and their possible dependencies makes it extremely difficult to achieve and predict SOA dependability which can vary over a wide range in a random manner. Therefore, users cannot be confident in availability, trustworthiness, reasonable response time and others dependability characteristics. Dealing with such uncertainty, partially coming from the SOA nature, is the main challenge. This uncertainty should be treated as the threat (similar and in addition to the common faults, errors and failures). The presentation discusses fault-tolerantce mechanisms and solutions for building dependable service-oriented systems out of undependable Web Service components with changeable functional sets and uncertain dependability characteristics, making use of natural redundancy and diversity inherent to such systems.

December 11, 2007 1pm, Room 701
Inaugural Lecture
Atau Tanaka

In his seminar for Computing Science, incoming Chair of Digital Media, Atau Tanaka, will present some elements from his recent inaugural lecture in Culture Lab. He will retrace the history of computing in sound and image production, and look at the place of technology in creative practice. By taking pairs of personalities as example, he explores the rich history of interchange between art and science. We will see how the physicality of a musician is translated by digital technology, then transposed to give viscerality visual media and proprioceptive awareness to urban environments.

September 20, 2007 2pm, G21/22 Devonshire Building
Learning at digital tabletops
Ahmed Kharrufa

Collaboration around tables occurs during meetings, learning, brainstorming, and other group activities. My research addresses the support of small group collaboration around computer enhanced tabletops. In this talk, I will first introduce the relatively new concept of digital tabletops and explain why groupware built for digital tabletops differ from that built for traditional single display settings and the challenges faced when designing groupware for such large and horizontal surfaces. Then I will present some of the main points to be investigated, such as the use of paper and annotation on paper when collaborating around tables; and the support of digital tabletops for collaborative learning. I will conclude with a summary of the observational studies that I have conducted on collaboration around traditional tables and present the main findings.

September 6, 2007 2pm, 701 Claremont Tower
A Domain Specific Language for Dynamic Interest Management with Virtual Environments
Sam Aaron

(This talk is given preparation for the upcoming O'Reilly Ruby on Rails Conference. Therefore, a weighted focus is placed on both the implementation technologies and methods (in particular the language Ruby, and the framework Rails) in addition to the results of the research.) A Brief Introduction to Interest Management Interest management is emerging as an important field in computing. Given the potentially huge sets of material contemporary information and communications technologies make available to us, how do we create smaller, more easily consumed, sets of things we are interested in? This problem is brought into focus when we consider that there are currently more blog posts available than you could read in a lifetime. The presentation will introduce the results of Sam Aaron's PhD research into creating a domain specific language for representing interests, or more abstractly subsets of information. The primary focus of the research was on applying this system within virtual environments, but the underlying principles make the ideas equally applicable to the filtering of RSS feeds, emails, spatial information, etc. From Java to Rails As part of the research, Sam migrated from a Java/VRML implementation of his ideas to a hybrid consisting of a VRML browser, a Java CLI, and a set of Ruby libraries, all controlled via a Rails application. The presentation will cover the motivations for this change, the process of migration, and the benefits gained. The research gives examples of the real-world application of a variety of Ruby/Rails tools and libraries, including Rspec, ActiveRecord, open3, database migrations, and the use of Domain Specific Languages.

July 5, 2007 2pm, 701 Claremont Tower
Resilience-explicit Web Services Based upon the WS-Mediator Architecture
Yuhui (Jeff) Chen

Services-oriented architecture (SOA) is rapidly becoming a critical technology in e-Science and e-Commerce application development.Dependability of SOA has become an active area of research in recent years. We propose a novel architectural solution for improving dependability of Web Services. This distributed solution, called WS-Mediator, implements an overlay architecture combined with a number of fault tolerance solutions and resilience explicit computing to ensure Web Services dependability. The globally distributed architecture of the WS-Mediator collects dependability-related metadata from the end-user's perspective, analyses them and acts upon them to tolerate faults using dynamic reconfiguration of the traditional fault-tolerance techniques. This approach also introduces service redundancy based upon resilient-explicit dynamic decisions to improve dependability of Web Services without having to improve the dependabiliy of individual Web Services. We have implemented a Java WS-Mediator framework based on the WS-Mediator concept, which can be easily integrated into implementation of Java Web Services applications. We report the results of the extensive experiments conducted to evaluate the applicability of this approach.

June 21, 2007 2pm, 701 Claremont Tower
Mitigating Provider Uncertainty in the Establishment of Service Provisioning Agreements
Christopher Smith

The inherent uncertainty of open, distributed and multi-party systems can make it infeasible to establish mutually beneficial relationships between constituent entities. Service provision is one such relationship, in which a service provider supplies a service, with an associated quality level, to a consumer. The service provider experiences uncertainty in his ability to deliver a quality level due to factors relating to system behaviour, such as load fluctuations and failures, and due to the inherent uncertainty of statistical estimation. Inability of the provider to effectively quantify and reason over these uncertainties can result in errors in the estimation of the deliverable quality level; leading to financial losses, inefficient resource usage and consumer distrust. We demonstrate the negative consequences of such errors, provide a generic approach for quantification of uncertainty using a sampling based monitoring policy under cost constraints, and extend common economic utility models to provide explicit reasoning over this uncertainty during the negotiation of service provisioning agreements.

June 14, 2007 2pm, 701 Claremont Tower
Consensus When Coverage Cannot Be Complete
Khaled Alekeish

In a loss- and disconnection-prone network, such as a mobile ad-hoc network (manet), complete coverage requires that a message be transmitted until all operative nodes acknowledge reception. This is prohibitively expensive when small, mobile wireless devices collaborate by forming a manet. This paper presents a solution to the consensus problem, essential to support user collaboration, using a new class of broadcasts that intentionally sacrifice full coverage for savings in memory and bandwidth. However, when such broadcasts are used, consensus amongst n devices cannot be guaranteed if at most f of them can crash (unnoticeably) and if n ≤ 3f. We present a protocol for n > 3f and evaluate its performance through simulations.

June 7, 2007 2pm, G21/22 Devonshire Building
Dependable Software via Automatic Verification
Wei-Ngan Chin

Despite their popularity and importance, pointer-based programs remain a major challenge for program verification. We propose an automated verification system that is concise, precise and expressive for ensuring the safety of pointer-based programs. Our approach is based on separation logic and uses user-definable shape predicates to allow programmers to describe a wide range of data structures with their associated properties. To support automatic verification, we design a new entailment checking procedure that can handle well-founded inductive predicates using unfold/fold reasoning. To improve expressivity, we support a non-deterministic set of states for proof search, intersection types for methods and coercion rules for related shape predicates. We have proven the soundness and termination of our verification system, and have built a working system.

Biographical Data

Wei-Ngan Chin received his BSc and MSc in Computer Science from the University of Manchester and a PhD in Computing from Imperial College, London. He is presently an Associate Professor in the Department of Computer Science, National University of Singapore. His research interests are in programming languages and software engineering. He has worked on various program analyses and verification techniques that are aimed at clarity, reliability and reusability. His recent research topics include software verification, memory resource analysis, and OO genericity. He currently leads several projects, including an A*STAR-funded project on "A Constructive Framework for Dependable Software".

May 31, 2007 2pm, 701 Claremont Tower
Concurrency Semantics of Membrane Systems
Maciej Koutny

Membrane systems are a computational model inspired by the way living cells are divided by membranes into compartments where chemical reactions may take place. They share a number of features with Globally Asynchronous Locally Synchronous systems (GALS). The talk will consider synchrony and asynchrony between executed reactions in the computations of such systems using Petri nets and their processes as a formal behavioural model. The non-sequential semantics of the resulting nets is formalised through processes based on occurrence nets. Such processes provide a convenient tool for analysing synchrony and asynchrony in the executions of membrane systems and shed light on the causal relationships between the reactions taking place.

May 17, 2007 2pm, 701 Claremont Tower
Digital Curation or Digital Data? The impact of Services and Federation.
Phil Lord

The CARMEN project is aiming to store, distribute and compute over data from the neurosciences; currently, most electrophysiological is only used within the lab that produced it, in contrast to many types of biological data which are widely shared and made openly accessible to the community. Part of the work of the CARMEN project will be to address the curatorial needs of the neurosciences community. In particular, we need to provide provenance for the data we hold, offer explicit statements about the retention policies for the data and inform the user about license or other usage restrictions on the data. The CARMEN project is being built on a novel computing infrastructure, however, which will produce significant and specific challenges for the curatorial process. In this talk I will describe some of these challenges, and provide two solutions: first, as well as curating the data we need to curate the computational services that have operated on that data; second, for curation policies to be useful within this environment, they must be expressed in a computationally amenable form.

May 10, 2007 1pm, 701 Claremont Tower
Dynamic allocation of resources in Grid computing
Joris Slegers

Recent developments in grid computing allow the hosting of services on clusters of computers. The execution of submitted jobs is completely opaque to the customer. This enables the provider to freely (re)configure his available resources and thus increase his efficiency. In this talk we will discuss efficient policies for assigning resources when the arrival of requests is a bursty process. Firstly we consider a static allocation, i.e. without allowing for dynamic reallocation of servers. We show both the optimal solution and a closed formula approximation that is easy to use and asymptotically exact for heavy traffic. And secondly we consider dynamic allocation of resources, possibly in the presence of significant reconfiguration time and cost. We introduce some heuristic policies for this dynamic allocation. We also compare their results to the optimal dynamic allocation -which is very computationally intensive to obtain- in a few simple cases. This work in progress is done as part of the joint DOPCHE-project in Newcastle and Warwick.

March 8, 2007 2pm, 701 Claremont Tower
Drug Design by Autonomous Software Agents
David Leahy

As part of the Science City initiative the university, through the Business School, is appointing a number of "Professors of Business Practice" whose expertise lies in entrepreneurial aspects of Science and Technology. David Leahy, a former Head of Physical Sciences at Zeneca Pharmaceuticals and subsequently founder of the biotechnology company Cyprotex PLC has been appointed to one of these posts. He is a computational chemist and that he will be keen to explore collaboration here, possibly leading to the establishment of a research group.

His talk will concentrate on the CS aspects of his work (technical detail on the agent architecture and methodology), and should be of interest to those working in e-science/distributed systems, bioinformatics, semantic web, and machine learning research areas.

February 22, 2007 1pm, 701 Claremont Tower
Yet Another Coordination Language
Alexei Iliasov

The presentation will discuss experience of developing a new coordination language for mobile agent systems which overcomes some limitations of Linda and Pubsub although retaining a level of backward compatibility. A combination of operational semantics, theorem proving and model checking techniques was used to design and analyse the properties of the new language.

December 14, 2006 1pm, G21/22 Devonshire Building
The Qurator workbench: Capturing and Processing the User Perspective on Data Quality
Paolo Missier

Data-intensive e-science applications often rely on third-party data found in public repositories, whose quality is largely unknown. Although scientists are aware that this uncertainty may lead to incorrect scientific conclusions, in the absence of a quantitative characterization of data quality properties they find it difficult to formulate precise data acceptability criteria.

The definition of such criteria, however, is often a knowledge-intensive experimental process that involves the scientists' expertise. The Qurator project, an EPSRC-funded collaboration between the School of Computer Science at Manchester and the School of Computing in Aberdeen, has developed an information quality management workbench that supports data experts in the specification, rapid deployment and testing of personal quality hypotheses for specific types of data. This is achieved by providing a declarative model and language for the definition of users' hypotheses, called "quality views", and a compilation of views into executable components that can be embedded in a user's data processing application. Quality views are described in terms of an extensible semantic model for Information Quality.

This talk presents the model and shows how the Qurator workbench compiles quality views into Taverna Scufl models, effectively creating reusable quality sub-workflows that can be integrated into a host workflow during a deployment step. An application to an in silico proteomics workflow from the Imprint project is presented.

Presented by: Paolo Missier, Research Associate School of Computer Science University of Manchester

November 16, 2006 1pm, 1.19 Devonshire Building
The application of Stochastic Petri Nets to Systems Biology
Oliver Shaw

Stochastic Petri nets (SPN's) have been successfully applied to systems biology. Recent studies show SPN's are well suited to biological modelling at the level of a single cell. Unfortunately, SPN models of biochemical systems suffer from two major drawbacks; computational expense, and unknown kinetic parameters. This talk proposes tools and techniques to explore these problems. One such tool, NASTY, allows simulation jobs to be distributed over a large number of CPU's. The compute resource of NASTY allows the exploration of an existing model's parameter space using computationally intensive methods. First a genetic algorithm is utilised to parameterise the network, then sensitivity analysis is applied in an attempt to reduce the search space.

November 2, 2006 1pm, 701 Claremont Tower
Gene Network Modelling
Jennifer Hallinan

Computational models of gene networks have been intensively studied since the 1960s, and have yielded insights into the global, emergent behviour of large, complex networks of interacting agents. In this talk I cover the main ways in which computational models have been used in my own and other research groups, particularly in combination with evolutionary computation, to study issues such as neutral evolution, the interaction of self-organization and selection, and the evolution of modularity. Recently, high-throughput "omics" experiments have generated large amounts of interaction data of a number of different kinds, making it feasible to construct large, integrated functional networks based on biological data. I will present some recent work done at Newcastle University on these networks.

October 26, 2006 1pm, 701 Claremont Tower
ATV: An Efficient Method for Constructing a Certification Path
Omar Batarfi

Constructing a certification path is the mainly method of validating a certificate at a given time. The established path will consist of all the valid certificates that reside between the target certificate, a public key to be validated, and the trust anchor1 certificate. No standardization has been issued for constructing certification paths which may make the process complex and subject to error; therefore, any optimization will be significant. This talk will present a method for constructing a certification path in an efficient manner which is called ATV.

October 19, 2006 1pm, G21/22 Devonshire Building
Transaction Manager Failover: A Case Study Using JBOSS Application Server
Achmad Iman Kistijantoro

Three-tier middleware architecture is commonly used for hosting large-scale distributed applications. Typically the application is decomposed into three layers: front-end, middle tier and back-end. On this architecture, availability measures, such as replication, can be introduced in each tier in an application specific manner. In this talk we will discuss, for the case of Enterprise Java Bean components and JBoss application server, how replication for availability can be supported to tolerate application server/transaction manager failures. Replicating the state associated with the progression of a transaction (i.e., which phase of two-phase commit is enacted and the transactional resources involved) provides an opportunity to continue a transaction using a backup transaction manager if the transaction manager of the primary fails. We will also describe the technical issues involved and show how a solution can be engineered.

October 12, 2006 1pm, 701 Claremont Tower
Opacity and Information Flow in Voting Systems
Thea Peacock

Opacity is a relatively new definition for secure information flow. I first define opacity, and relate it to some existing security properties such as non-interference and non-deducibility. This establishes the distinct character of opacity, and indicates that it has a high degree of flexibility. I then describe a case study of voting systems, the aim of which was to test this flexibility, and investigate how opacity might be used in a practical application. Prêt à Voter is the main exemplar, with the Chaum, Neff and FOO schemes providing comparisons as appropriate. Three different analyses were carried out: information flow, systems-based and coercion-resistance. We find that opacity can form the basis for succinct expressions of coercion-resistance in different voting systems, and that it is a useful tool for security analysis.

October 5, 2006 1pm, G21/22 Devonshire Building
Finding trust paths using P2P search algorithms: a simulation study
Emerson Ribeiro de Mello

Distributed security models based on webs of trust are a good choice for large scale environments because they eliminate single points of failure and performance bottlenecks. However, decentralisation requires finding trust paths between, for instance, a client and a service provider, which is sometimes difficult. Since decentralized peer-to-peer networks share some characteristics with webs of trust, P2P search algorithms can be exploited for finding trust paths. In this work we present simulations that apply search algorithms used in descentralized P2P networks for finding paths in webs of trust, and provide an analysis of their performance and effectiveness. While previous work is concerned with searching resources provided by P2P nodes, in our work we are interested in the connections between nodes.

August 31, 2006 1pm, 701 Claremont Tower
Towards an automated service trading
Pablo Fernandez and Manuel Resinas

Automating the creation and management of SLAs in electronic commerce scenarios brings many advantages, such as increasing the speed in the contracting process or allowing providers to deploy an automated provision of services based on those SLAs. We define service trading as the process of locating, selecting, negotiating, and creating SLAs. This process can be applied to a variety of scenarios and, hence, their requirements are also very different. In this talk we shall introduce and motivate an abstract architecture for service trading and we detail the design of a software framework based on that architecture

August 29, 2006 1pm, 701 Claremont Tower
Performance Evaluation of CYCLON - An Inexpensive Membership Management Algorithm for Unstructured P2P Overlays
Francois Bonnet

The talk will examine the performance of CYCLON, a membership management algorithm designed for unstructured P2P overlays. It will present two models to analyse Cyclon performance, in terms of convergence speed and quality of the obtained overlay. The algorithm itself will be modelled using two discrete time Markov chains and generating functions will be used to find stationnary distributions.The upper bound established on the convergence speed of Cyclon shows that it is a fast mixing algorithm. (This work was carried out as a part of Francois masters thesis with Frédéric Tronel IRISA, Rennes, France, and appears in the DISC 2006.)

July 20, 2006 1pm, 701 Claremont Tower
Managing Missed Interactions in Distributed Virtual Environments
Simon Parkin

In a distributed virtual environment interest management schemes attempt to ensure virtual world objects exchange messages only if they are interacting. This provides an opportunity to allow such systems to scale to support many hundreds, even thousands, of virtual world objects by minimising the sending and processing of unnecessary messages. However, arriving at the optimum configuration for the interest management scheme is a challenging and complex problem - the customary approach of configuring parameters in an ad-hoc manner is in no way guaranteed to yield optimum system performance. In this talk I shall discuss the role of interest management schemes within virtual worlds, focusing on how their performance can be evaluated and optimised with a suite of simulation tools that we have developed.

June 22, 2006 1pm, 701 Claremont Tower
“Modelling job allocation where service duration is unknown” or “Why wasting time can be productive”
Nigel Thomas

In this talk a novel job allocation scheme in distributed systems (TAG) is modelled using the Markovian process algebra PEPA. This scheme requires no prior knowledge of job size and has been shown to be more efficient than shortest queue, round robin and random allocation when the job size distribution is heavy tailed and the load is not high. In this talk the job size distribution is assumed to be of a phase-type (hyper-exponential) and the queues are bounded. Numerical results are derived and compared with those derived from models employing random allocation and the shortest queue strategy. It is shown that TAG can perform well for a range of performance metrics. A new extension to the original TAG algorithm is proposed which makes further improvements in efficiency.

June 15, 2006 1pm, 701 Claremont Tower
Computational neuroanatomy and its applications
Tom Binzegger

The talk will discuss the structure of a generic neuronal circuit (cortical micro-circuit) which is central to the information processing abilities of mammals. This circuit is of considerable complexity, involving tens of thousands of neurons and millions of connections. I will show how it is possible to model the circuit structure in fine detail and briefly discuss the potential relevance of structure for designing efficient neural network implementations in the context of distributed computing.

June 13, 2006 1pm, 1.18 Claremont Tower
Empirical and Analytical Evaluation of Systems with Multiple Unreliable Servers
Jennie Palmer

We construct, analyze and solve models of systems where a number of servers offer services to an incoming stream of demands. Each server goes through alternating periods of being {em operative} and {em inoperative}. The objective is to evaluate and optimize performance and cost metrics. A large real-life data set containing information about server breakdowns is analyzed first. The results indicate that the durations of the operative periods are not distributed exponentially. However, hyperexponential distributions are found to be a good fit for the observed data. A model based on these distributions is then formulated, and is solved exactly using the method of spectral expansion. A simple approximation which is accurate for heavily loaded systems is also proposed. The results of a number of numerical experiments are reported.

June 1, 2006 1pm, 701 Claremont Tower
Signal-On-Fail: A Cost-effective Approach to Circumvent the Impossibility on Fault-tolerant Total Order
Qurat-ul-Ain Inayat

Fault-tolerant total ordering of client requests is a fundamental requirement for State Machine Replication. Yet, when replicas communicate over an asynchronous network such as the Internet, it cannot be achieved through a deterministic solution that can be guaranteed to terminate. This talk presents a novel approach to deal with this impossibility and exposes the performance gains obtained thereof. In our approach, processes with signal-on-fail semantics are built by (i) having a subset of Byzantine-prone processes paired to check each other’s protocol execution, and (ii) assuming that paired processes do not fail simultaneously. By dynamically invoking the construction of signal-on-fail process, two coordinator-based total-order protocols are developed and their performance evaluated using a LAN-based implementation. Failure-free order latencies and fail-over latencies are measured; the former are shown to be smaller compared to a protocol (by Liskov And Castro) which is generally regarded to perform exceedingly well in the best-case scenarios.

May 25, 2006 1pm, 701 Claremont Tower
Electronic Transfer of Prescriptions
Bob Sugden

This talk discusses the socio-technical issues encountered during the evaluation of three pilot implementations to enable electronic transmission of prescriptions (ETP) in England in 2002. Following the pilots, a national system for ETP is being progressively implemented in England during 2005. The evaluation observed a number of technical issues, which caused difficulties observed during the evaluation, and there were also some specifically social issues observed. This talk will focus on the boundary where the social and technical issues overlap. In particular, the pilots exhibited a structural approach to support for business processes, in comparison to the infrastructural support offered by current systems.

PROFILE:

Bob Sugdens research interests encompass the interrelationship of system requirements, design, organisational culture and business processes. He rejoined CSR in late 2003, after spending eight years working at the Sowerby Centre for Health Informatics at Newcastle (SCHIN), latterly as Director of Informatics. Whilst at SCHIN, he co-ordinated the Department of Health Prodigy project providing decision support for clinicians in primary care, working with a multi-disciplinary team. He was responsible for all technical development on Prodigy Release 1, and contributed to the design and evaluation of later research prototypes, which included collaborative work with the Medical Informatics group at Stanford (Protégé). Latterly at SCHIN he co-ordinated the evaluation of the Department of Health pilots for electronic transmission of prescriptions. Whilst at CSR from 1993-1995, he co-ordinated a project which studied development lifecycles and change control processes, with particular reference to the origins and management of requirements instability,including emergent requirements and compositionality problems. Prior to this he worked in industry, the NHS and local government, including directorships in a number of companies. He has worked as a project manager, requirements analyst, system designer and software engineer. He is a Chartered Engineer and a Fellow of the British Computer Society.

May 18, 2006 1pm, 701 Claremont Tower
Distributed Query Processing on P2P Systems
John Colquhoun

Today’s systems employing relational database management systems are based on the traditional client-server architecture. In this paper, we consider an alternative based on the P2P architecture. At present, P2P systems are mainly used for file-sharing applications, but this paper investigates the design of a P2P database, based on the popular BitTorrent file-sharing protocol. The peers cache the results of queries, and use them to satisfy the requests of other peers, so reducing the load on the server and producing more scalable systems. This paper will introduce both the concepts and the research issues relating to P2P database systems, with the differences between P2P file-sharing (BitTorrent) and a P2P database being highlighted.

In the system being investigated, a database table is initially located on one machine known as a seed. When other peers receive data from this machine, they can make it available to others. A central component, known as the Tracker, stores a record of which parts of the table each peer has. A peer requesting data receives from the Tracker a list of peers that could satisfy the query.

A simulator has been written to simulate BitTorrent and has been extended to explore the P2P database system. Various parameters, such as the upload rate of peers and system workload, can be altered and various statistics, including the number of I/O’s performed and the query response time, can be measured. This system has been tested with a sample database and the paper presents results analysing its behaviour. Future work will include experiments with more complex queries and sample workloads from real client-server databases.

May 4, 2006 1pm, 701 Claremont Tower
Service-Oriented Computing Contest
Aad van Moorsel

Ill report on a service-oriented computing contest a team of us (Chris, Emerson, Jim, Philipp, Savas, Aad) is participating in. We are building a service-oriented proof-of-concept for a real-estate application, using technologies like SSDL, formal correctness proofs and web services security. Ill also touch on some of the non-technical aspects of participating in this contest.

April 27, 2006 1pm, G21/22 Devonshire Building
The Architecture of a Repository System for Software Component Reuse.
Basem Y Alkazemi

Software component reuse is of obvious importance to the software engineering process, and is increasing in prominence due to the highly successful open-source code movement. While open-source code is, by its very nature, available for re-use, facilities for finding components that fits an engineer’s needs are somewhat limited. Repositories for open-source code are available on the world-wide web, and work well if you know what you are looking for. However, such repositories are more problematical for unstructured searching or browsing, or for assisting in the reuse of part of an open-source code system that has been deposited in the repository. This talk presents a repository for open-source code that aims to further promote the identification and reuse of software components.

April 20, 2006 1pm, G21/22 Devonshire Building
E2R - Implementing complex SAP modules for real... warts and all!
Paul Hopkins

Presenter:- Graham Parker - SAP BASIS Team Leader

Between March 2005 and October 2005, the University ran an intense business/technical project called "Enquiries to Registration" which re-engineered the whole process of recruiting postgraduate students across the University. The project implemented satandard SAP modules of Portal, Customer Relationship Management, Workflow, Business Warehouse, Document Scanning/Management, a major development of the Campus Management module and a host of other developments.

To support this bewildering array of new modules, it was necessary to design, acquire, implement and configure more than 25 SUN /Intel servers running Solaris, Linux and Windows 2003. Certain servers were clustered in to High Availability combinations, several new landscapes (Development, Test/QA and Production) were created. This is now one of the most complex technical solutions in any HE Institution.

Graham Parker is the team leader of the SAP BASIS team in Newcastle University and was intimately involved in all stages of the design, implementation, configuration and on-going settle down of this environment. The presentation will focus primarily on the technical aspects of the new SAP technical environment and provide a useful insight in to the "real-world" difficulties of trying to implement complex technical solutions in a very short space of time.

Prior to the SRG meeting, attendees might be interested in visiting www.ncl.ac.uk/iss/e2r/ to see some "movies" which show aspects of the "end-user" view of the E2R system. Note that there are no technical specifications available on this web-page!

April 6, 2006 1pm, 701 Claremont Tower
A Wide Area Network Emulator for CORBA Applications
Mohammad Alsaeed

It is usually infeasible to test distributed systems and protocols for dependability, fault-tolerance and performance in the real operation environment. This is attributed to the cost of setting up the hosting distributed environment and the difficulty of controlling such a dynamic environment. The factors are too many and are very dynamic. In this talk, Ill present the Wide Area Network Emulator for CORBA Applications: a general purpose software tool that emulates a distributed networked environment for CORBA middleware and applications and gives the user full control over the emulation environment in addition to the ability to inject network-related faults.

March 30, 2006 1pm, 701 Claremont Tower
Trust Relationships and Security in SOA
Emerson Ribeiro de Mello

At present the amount of services on the Internet are increasing significantly. Usually, each service has its own security and administrative policies. In this kind of environment, users need to manages their own personal information, for each service that they will interact with. Thus, users need to supply the same kind of information many times. With the objective to improve the business, service providers are integrating its applications, that result in more facilities to their users. However, new security challenges are associated with this new environment. This presentation shows a model for identity management and to establish trust relationships. The idea is to increase the amount of negotiations between clients and service providers, in order to assist both sides.

March 23, 2006 1pm, G21/22 Devonshire Building
Communication Patterns in Flexible Problem-solving Systems
Chris Holt

This is intended to be an informal discussion considering the question of what we can learn from human behaviour as applied to communities of artificial entities, e.g. expert system and neural net hybrids. Since such communities are likely to be insecure, each entity should keep a record of trust and likely accuracy to determine which of its friends should be listened to for any given question. This is not Byzantine generals, however; its closer to zero-information proof-style confidence levels. The balance between broadcast, narrowcast and one-to-one conversations is considered.

March 16, 2006 1pm, G21/22 Devonshire Building
Interaction and space
Chris Kray

In this talk, I will give an overview over my research, which is concerned with interaction and space. I will report on issues in and solutions for situated interaction on spatial topics in the context of a mobile tourist guide. Topics covered will include communicating spatial concepts and coping with location information of varying quality. Furthermore, I will discuss the concept of relative positioning and its application in ubiquitous computing. Here, I will present a system combining location sensing hardware and a distributed infrastructure for spatial interaction on planar surfaces. Finally, I will present one of my current research projects, which deals with pervasive navigation support for pedestrians using networks of situated displays. Ill point out some interesting research question in that area and discuss some initial ways of dealing with them.

March 9, 2006 1pm, 1.19 Devonshire Building
Performance Modeling and Evaluation of E-Business Systems
Giovanna Ferrari

In this talk we analyze and evaluate a general model of an E-business system, comprising an Application Server and a Database engine. Incoming jobs, once admitted for processing, compete for the available resources. Database accesses are of different types and are subject to time-outs. We have provided an efficient approximate solution, which accuracy is evaluated by comparing the model estimates with those obtained from simulations. We show here the results of a series of numerical and simulation experiments that examine the effect of several controllable parameters on the performance of the system.

March 2, 2006 1pm, 1.19 Devonshire Building
ReSIST Everything
Tom Anderson

Dependability, resilience, security and trustworthiness may not immediately spring to mind when you recall the life and works of Oscar Wilde, but the SRG title above is the best quote I could find (pace Daleks, Cybermen and the Borg). This SRG provides an opportunity to consider the very much wider set of opportunities that the ReSIST Network of Excellence can offer. The ReSIST NoE project has already hastily begun, involves 18 Eurpoean partners with research interests in Resilience, and aims: to integrate research in resilience; to build and scale-up resilience technologies (with emphasis on architecture and implementation, algorithms and mechanisms, socio-technical systems, verification methods and tools, evaluation methods and tools); to promote training and dissemination. We (Newcastle) have explicit opportunities for engagement in all of these areas, and in addition we have some modest funding for work on resilience integration. The short talk will try to outline the very broad research area covered and indicate how anyone interested could get involved. Ideally this will be followed by questions, and discussion of the answers - the presenter may well be uncertain and will welcome your assistance.

February 23, 2006 1pm, 1.19 Devonshire Building
Mona Lisa Overdrive or Driving Miss Daisy: Building Responsive Distributed Virtual Environments.
Graham Morgan

In recent years the commercialisation of distributed virtual environments by the digital games industry has provided easy access to virtual worlds for anyone with an Internet connection. Thousands of individuals are now participating in “far away galaxies”, “lands of myth and legend” and “cities of heroes and villains”. Even though this appears exciting, such virtual worlds only appeal to a minority of the game playing public. This is not due to the graphics and sound quality of such games (they are on a par with most games), nor is it due to the story lines provided by the game developers (they provide more in-depth, involving, stories). So what is it? In this talk we describe the problems hindering the development of the online games market and indicate how work at Newcastle is trying to solve such problems.

February 14, 2006 1pm, 1.19 Devonshire Building
A Proposal for Records in B
Neil Evans

Our current work in the RODIN project involves the redevelopment of an existing VDM specification using Event B. Although the aim is not to construct a translation from VDM to Event B, we propose the introduction of records in Event B (akin to those of VDM). In this talk I will explain how records can be used in the construction of Event B specifications and, more importantly, how these records are amenable to refinement.

February 2, 2006 1pm, 701 Claremont Tower
A mediator system for improving the dependability of Web Services
Yuhui Chen (Jeff)

Web Services bring unprecedented flexibility into e-commerce and e-science system architecture. It is becoming a key factor in the success of many e-Science or e-Commercial projects. As a prominent role in the integration of distributed computing infrastructure, Web Services need to provide high Quality of Services. However service-oriented architectures are inherently unreliable due to different internal and external errors and faults. Users may even experience service failures entirely caused by network problems. Most of current approaches concentrated on adopting fault tolerance mechanism to tolerate internal errors within an individual Web Service or the integration of a group of Web Services. In this talk, I will briefly present a novel solution for improving the dependability of Web Services. This solution is based on a concept of Web Services mediator, which provides efficient mechanism to tolerate external failures, especially communication faults between users and the Web Services, which result in undependable services delivered to Web Services users. Our approach can improve the dependability of pre-existing Web Services without modification on them and benefit Web Services users by providing dependable services to them.

December 20, 2005 14:00-16:30, Devonshire 1.19
A Tutorial on the Semantic Web
Phil Lord

December 13, 2005 13.00 - 14.00, G21/22 Devonshire Building
Self-healing software systems
Matthias Rohr

Self Healing is a concept to cope with the increasing complexity and dependability requirements of mission-critical component-based software systems. It is inspired by the autonomous healing mechanisms in biology. The goals of Self Healing and those of Fault Tolerance are similar, but differences can be found in the strategies. From the viewpoint of Software Engineering, we support this young research domain with missing definitions and a basic taxonomy to provide navigation points for further research. Our current research addresses the specification of Self Healing models based on architectural descriptions.

December 12, 2005 13.00 - 14.00, G21/22 Devonshire Building
Failure Models for Intrusion Tolerance
Timo Warns

Intrusion tolerance combines the fields of fault tolerance and security to prevent adverse consequences in the presence of intrusions. Such a context yields several challenges for the application of fault tolerance approaches. Fault-tolerant algorithms were developed with the threshold model under the assumption of independent failures. However, this assumption is not valid in the presence of adversaries anymore. Within the talk, we briefly present current failure models that cover dependent failures and, therefore, provide an improved assumption coverage. Furthermore, we show that they are strictly stronger than the common threshold model and equivalent among each other wrt. the solvability of problems. Finally, we introduce a new model for directed dependent failures and show that this model is strictly stronger than the models known so far.

December 1, 2005 13.00 - 14.00, 1.19 Devonshire Building
Robustness and recovery in biological systems – New approaches for Grid computing?
Marcus Kaiser

Biological systems are remarkably robust towards damage. Disabling genes has often little effect on the survival of an organism. Similarly, the human neural system can in some cases compensate for the removal of half of the brain. After describing the organization of brain fibre networks, I will present examples of recovery in neural and metabolic systems. I will show how structural features of the network might enable or enhance recovery. Subsequently, I will discuss how these features could be applied to computer networks. In particular, I will outline a future research project on faster recovery after node failure in the domain of Grid computing.

November 10, 2005 13:00, Claremont Tower 701
Dynasoar
Chris Fowler

This talk will discuss work in progress in the Dynasoar project which is investigating an architecture for dynamically deploying Web Services, on-demand, over a grid or the Internet. This is an approach to grid computing in which there are no jobs, but only services. Rather than scheduling jobs, Dynasoar automatically deploys a service on an available host if no existing deployments can meet the computational requirements. The architecture makes a clear separation between Web Service Providers, who offer services to consumers, and Host Providers, who offer computational resources on which services can be deployed. The talk will describe the architecture, a prototype implementation, and the new usage models that it supports.

March 24, 2005 1.30pm, T519
Dynamic Routing Among Several Intermittently Available Servers
Simon Martin

March 24, 2005 1.05pm, T519
Optimal Tree Structures for Large Service Networks
Jennie Palmer

February 23, 2005 1.05pm, T518
Efficient Group Communication in Multi-Hop Ad-Hoc Networks with Crash Failures
Einar Vollset

February 10, 2005 1.05pm, T120
Once Upon a Time, in a Land Far, Far Away or Superscalar Computers: Some Revelations
Brian Randell

November 16, 2004 1.10pm, T519
Bacillus Subtilis Protein-Protein Interaction Network Analysis
Olusola Idowu

October 7, 2004 1.05pm, T519
Design and Evaluation of a QoS-Adaptive System for Reliable Multicasting
Antonio di Ferdinando

October 4, 2004 1.05pm, T519
Network Analysis of Proteomes
Peter Andras

September 6, 2004 1.30pm, T519
Sorry, the Computer Ate Your Vote, Have a Nice Day!
Peter Ryan

August 20, 2004 1.30pm, T519
Error Analysis of a Security-Oriented User Interface
Roy Maxion (Carnegie Mellon University, USA)

July 20, 2004 2.00pm, T519
Using Unification for Opacity Properties
Laurent Mazare (Verimag, Grenoble)

June 22, 2004 1.10pm, T519
Evolution of Java
Robert Stroud

March 10, 2004 1.05pm, T518
A Defense-Centric Attack Taxonomy
Roy Maxion (Carnegie Mellon University, USA)

February 17, 2004 1.00pm, T518
The Autograph Protocol: A Reliable Broadcast Protocol for Mobile Ad-hoc Networks
Einar Vollset

February 11, 2004 12.30pm, T518
Post OGSI: The Latest Developments in the Web/Grid
Savas Parastatidis

October 15, 2003 1.10pm,
An Overview of Security Research at Carnegie Mellon University<92>s Dependable Systems Laboratory
Roy Maxion (Carnegie Mellon University, USA)

September 2, 2003 1.00pm,
Proving the Correctness of an Asynchronous Communication Mechanism using Rely-Guarantee
Neil Henderson

July 23, 2003 1.00pm,
An Introduction to Web/Grid Service Technologies and Specifications
Savas Parastatidis

July 9, 2003 1 pm, T518
Network-Aware Application Adaptation for Mobile Hosts
Benjamin Atkin

Applications which communicate extensively over a network can be highly sensitive to network performance. This is especially true in a wireless network, where mobile hosts must cope with large variations in available bandwidth. Explicitly adapting to network conditions can help mask this variability. For instance, a client-server application can defer inessential work and otherwise reduce communication when the quality of connection to the server is poor. In this talk I will describe an approach to application adaptation which differs from the usual mode-based technique. I will also discuss two implementations of modeless adaptation which I have worked on: ATP, a library for bandwidth-aware communication, and MFS, a cache manager for a distributed file system.

June 4, 2003 1.00pm,
From Cash Tolerance to Authenticated Byzantine Tolerance: A Structured Approach, the Cost and Benefits
Dimane Mpoeleng

March 5, 2003 1.05pm, T518
Dealing with Complexity
Peter Andras

Complexity is a top favourite buzz word in current science media. Many times its meaning is not very clear, and many times it is a source of diverging speculations due to its vague definitions. Here we propose a more rigorous approach to the topic of complexity. We discuss what makes systems complex, how can measure this complexity, what is the link between complexity and chaos, and what are the advantages and disadvantages of complexity in the context of information processing systems. We will cover several kinds of information processing systems, including biological systems, social systems and technical and abstract information processing systems.

January 2, 2002 ,
CoRSA - Constraint-based Requirements and Safety Analyis
Kevin Hollingworth

It is now widely accepted that the development of critical embedded systems should be subjected to safety analysis. However, unlike for conventional critical systems, current approaches for the safety analysis of computer-based systems are of limited effectiveness. In this presentation a novel approach for modelling and safety analysis of embedded systems will be introduced. This approach is based on establishing a relationship between the analysis of safety properties of a state transition model and the resolution of a constraint satisfaction problem. In particular the presentation will: introduce a process model for CoRSA; describe a template notation for modelling entities of a physical environment or controller, based on expressing state transition models as constraints; establish relationships between the analysis of these models and the resolution of constraint satisfaction problems.

January 1, 2002 2pm, CLT701
High quality P2P service provisioning with decentralized trust management
Vu Le Hung

Trust management is essential in fostering cooperation and high quality service provisioning in several peer-to-peer (P2P) applications. Among those applications are customer-to-customer (C2C) trading sites or markets of services implemented on top of centralized systems, P2P, and online social networks. Under these application contexts, existing work does not really address the heterogeneity of the problem setting in practice. This heterogeneity includes the different approaches employed by the participants to evaluate trustworthiness of their partners, the diversity in contextual factors that influence service provisioning quality, as well as the variety of possible behavioral patterns. This talk presents the design and usage of appropriate computational trust models to enforce cooperation and ensure high quality P2P service provisioning, considering the above heterogeneity issues. First, I propose a graphical probabilistic framework for peers to model and evaluate trustworthiness of the others in a highly heterogeneous setting. The framework targets many important issues: the multi-dimensionality of trust, the reliability of different rating sources, and the personalized modeling and computation of trust on a participant based on the quality of services it provides. Next, an analysis on the effective usage of computational trust models in environments where participants exhibit various behaviors, e.g., honest, rational, and malicious, will be presented. I provide theoretical results showing under which conditions cooperation emerges when using trust learning models with given accuracy and how cooperation can still be sustained while reducing cost and accuracy of those models. As a another contribution, I designed and implemented a general prototyping and simulation framework for reputation-based trust systems. The developed simulator can be used for many purposes, such as to enable empirical simulation to discover new phenomena or to test various hypotheses in complex settings. Two potential applications of computational trust models are then discussed: (1) the selection and ranking of (Web) services based on quality ratings from reputable users, and (2) the use of a trust model to choose reliable delegates in a key recovery scenario in distributed social networks. Finally, I will identify a number of various issues in building next-generation reputation-based trust systems as a future direction of my research. Note: This talk comprises the result of my PhD research at EPFL over recent years in collaboration with my colleagues from several institutions.

January 1, 2002 14:00 - 15:00, TBC
Web and Distributed Software Development Risks Management: WeDRisk Approach
Ayad Keshlaf and Kamarul Abdul Basit

In spite of a variety of software risk management approaches, the software industry is still suffering from associated risks. Web and distributed software development is an example, where there are specific challenges and risk areas which need to be addressed, considered and managed. In this paper we present a list of potential web and distributed risks that we have identified based on their challenges and characteristics. We survey a number of software risk management approaches and identify their weaknesses and strengths for managing web and distributed development risks. Examples of weaknesses which we identify include the treatment of cultural issues, geographic location, and process and product perspectives. The identified strengths are quite general and only few of them are targeted to web and distributed developments. Following the review of strengths and weaknesses we present an approach called WeDRisk which we propose in order to tackle the weaknesses of the existing approaches, and to accommodate the continuously evolving challenges to web and distributed software development. WeDRisk tries to cover some aspects and perspectives which have not been covered up to now. WeDRisk is a part of PhD research at Newcastle University (UK) to develop an improved approach to measure and control web and distributed development risks.

January 1, 2002 14:00 - 15:00, Rm. 701, Claremont Tower
Seminar: Social Inclusion through the Digital Economy –a Computer Science Perspective
Paul Watson, Patrick Olivier, Aad van Moorsel

Poor health, disability, family breakdown, poverty and unemployment are just some of the reasons why people of all ages may become marginalised from society and may lack the skills, confidence or opportunities to access and benefit from digital technologies that have the potential to transform their lives. The Digital Economy Research Hub aims to tackle social exclusion by making it easier for people to access the life-changing benefits offered by digital technologies. The £12 million Hub is based at Newcastle University and is a key element of the Research Councils UK Digital Economy research programme. The Hub, in which the University of Dundee will also play a key role, will address four fields where digital technologies and the building of a truly inclusive digital economy could deliver major social benefits: Connected Home & Community; Accessibility; Inclusive Transport Services; and Creative Industries. The researchers will be working closely with people who are affected by social exclusion. The team will work with a group of 3000 volunteers from the Dundee and Newcastle areas, including people from a range of age groups and with a variety of disabilities. In the North East, a large panel of older people recruited and supported by Years Ahead, the Regional Forum on Ageing which is a key partner in the Hub, will be involved. In this seminar we will give an overview of the project, before focussing on the computer science challenges. One of the aims is to interest more researchers in becoming involved in tackling some of the major research challenges of the project. To encourage this, we will be launching a PhD studentship competition, allowing CS academic staff who are not investigators on the project to supervise a PhD on a related topic of their choosing.

January 1, 2002 ,
Coordinated Exception Handling in Distributed Object Systems: from Model to System Implementation
Jie Xu

Exception handling in distributed and concurrent programs is a difficult task though it is often necessary. In many cases traditional exception mechanisms for sequential programs are no longer appropriate. One major difficulty is that the process of handling an exception may need to involve multiple concurrent components when they are cooperating in pursuit of a global goal. Another complication is that several exceptions may be raised concurrently in different nodes of a distributed environment. Existing proposals and actual concurrent languages either ignore these difficulties or only cope with a limited form of them. In this talk, I'll report major results of the three-year joint investigation of coordinated exception handling, conducted together with Alexander Romanovsky and Brian Randell. We attempt to provide a general solution, developed especially for distributed object systems, starting with from conceptual model, together with algorithmic mechanisms (and their correctness proofs) for coordinating concurrent components and resolving multiple exceptions, through to an actual system implementation. An industrial production cell is chosen as a case study to demonstrate the usefulness of the proposed model and algorithms. The supporting system and the resolution mechanism are implemented in distributed Ada 95 and examined through several performance-related experiments.

January 1, 2002 ,
Anonymous and confidential communications with PDA
Carlos Molina

Personal Digital Assistants (PDAs) have the potential to be the most numerous computers next century. They are cheap enough (currently about $500) to be affordable by everybody and powerful enough to be used as datebooks, notepads, word processors, and as mobile communicators for e--mail, accessing remote services (data bases and Web pages) and performing commercial transactions. However, the acceptance of PDAs as handy communicators depends heavily on how easy it is for a lay person to purchase or borrow a PDA, turn it on and put it to work. Equally important is the guarantee that the change to PDAs will not compromise the rights of confidentiality and anonymity that users are used to. A model based on the public phone box service operated by metal coins is presented as a possible solution to hide the identity of the PDA user. The use of public and secret keys encryption mechanisms guarantees confidentiality.

November 30, 1999 ,

November 30, 1999 , Room 701, Claremont Tower

• Colloquia
• Systems Research Groups
• ASL
• Security
• University Events
• Events RSS
• Events Atom
• Twitter