Research Overview
My research interests concentrate on the theoretical foundations and effective engineering of Enterprise Systems: large-scale, distributed, business critical systems that extend across and between organisations. This has led to research contributions in Cloud computing (University of Melbourne); open, scalable infrastructure for the National Health Service (Oxford University); service-oriented architectures, Web service choreography (School of Informatics, University of Edinburgh); and optimising data movement in Terabyte scale repositories (National e-Science Centre). Complementing my academic experience, I have worked on projects at two industrial research laboratories, extended transactions at Hewlett Packard and formal methods at BAE Systems. My future research agenda addresses the challenges of engineering autonomic and on-demand Cloud computing technologies.
I have the following current research interests:
- Service Oriented Architectures
- Formal methods for service composition: Process calculus
- Workflow: orchestration and choreography languages
- Cloud computing
- Grid computing and its application to automating large-scale science, e-Science
- Health-informatics
- Peer-to-peer systems
- Semantic Web technologies
Research Projects
My systems research agenda has been pursued through the following academic and industrial projects:
1) Cloud Computing
Research Fellow, University of Melbourne: October 2009 - Current
In order to meet the increasing data storage and computing demands, applications are beginning to rely on cloud computing services: large-scale out-sourced data and compute resources made available to end-users and applications. Clouds are primarily optimised for very specific functionality, which means that data-intensive applications must increasingly interact with and utilise multiple specialised stand-alone clouds. My research at the University of Melbourne focuses on infrastructure to assist engineers in building and optimising applications which are constructed from multiple interoperating cloud services. This work is partnered with Microsoft.
2) Scalable Infrastructure for the National Health Service
Research Associate, University of Oxford: August 2008 - August 2009
At the University of Oxford I worked on a EPSRC/MRC grand challenge health-informatics project, partnered with the National Health Service (NHS), the University of Edinburgh and Imperial College, London. The premise: the sheer quantity and complexity of medical information, even within a single speciality, is beyond the power of one person to comprehend. However, the specific pieces of information most relevant to a particular clinical decision will typically be scattered over a wide range of databases, applications, journals and written notes. Centralisation of clinical knowledge is becoming rapidly less practical as the volume of data increases. My research at Oxford focused on designing, building and verifying open distributed systems to address the data explosion in medical informatics.
Key publications:
Adam Barker, Paolo Besana, David Robertson and Jon B. Weissman. The Benefits Of Service Choreography For Data-Intensive Computing. In Proceedings of the 7th International Workshop on Challenges of Large Applications in Distributed Environments (CLADE'09), in conjunction with HPDC'09: The 18th International Symposium on High Performance Distributed Computing, pages 1-10. ACM, June 2009.
Paolo Besana, Vivek Patkar, Adam Barker, David Glasspool and David Robertson. Sharing Choreographies in OpenKnowledge: A Novel Approach to Interoperability. Journal of Software (JSW), Vol 4, No 8 (2009), pages 833-842, Oct 2009.
Paolo Besana and Adam Barker. An Executable Calculus for Service Choreography. In Proceedings of the 17th International Conference on Cooperative Information Systems (CoopIS 2009), volume 5870 of LNCS, pages 373-380. Springer, November 2009.
3) Orchestrating Data-Centric Workflows
Research Associate, National e-Science Centre, University of Edinburgh: March 2007 - August 2008
When orchestrating data-centric workflows centralised servers common to standard workflow systems (e.g. implementations of BPEL can become a bottleneck to performance. My research at the National e-Science Centre delivered the Circulate architecture, a light-weight hybrid workflow model which maintains the robustness and simplicity of centralised orchestration, but facilitates decentralised choreography by allowing services to exchange data directly with one another. An open-source, Web services based implementation serves as a live deployment platform. Performance evaluation conducted on the PlanetLab framework concludes that a substantial reduction in communication overhead results in a 2--4 fold performance benefit across common workflow patterns. As the complexity of a workflow grows, (i.e. workflow patterns are used in combination with one another) the advantage of using the hybrid architecture increases.
This post-doctoral research was conducted at the National e-Science Centre (NeSC) in collaboration with Jon Weissman from the University of Minnesota, Malcolm Atkinson and Jano van Hemert from the University of Edinburgh.
Key publications:
Adam Barker, Jon B. Weissman and Jano van Hemert. Eliminating the Middle Man: Peer-to-Peer Dataflow. In HPDC'08: Proceedings of the 17th International Symposium on High Performance Distributed Computing, pages 55-64. ACM, June 2008.
Adam Barker, Jon B. Weissman, and Jano van Hemert. Orchestrating Data-Centric Workflows. In Proceedings of the 8th IEEE International Symposium on Cluster Computing and the Grid (CCGrid), pages 210-217. IEEE Computer Society, May 2008.
Adam Barker and Jano van Hemert. Scientific Workflow: A Survey and Research Directions. In Roman Wyrzykowski et al., editors, Seventh International Conference on Parallel Processing and Applied Mathematics, Revised Selected Papers, volume 4967 of LNCS, pages 746-753. Springer, 2008.
4) Web Service Choreography
School of Informatics, University of Edinburgh
My doctoral research at the University of Edinburgh focused on the design and implementation of a decentralised service choreography language, with formal semantics. My thesis made four key contributions to knowledge, the first is the directly executable MultiAgent Service Choreography language (or MASC). Secondly, MASC allows peers to coordinate in open-systems, specifications of choreography are disseminated at runtime, this allows peers to coordinate without any hard-coded or prior knowledge of the interaction pattern. Furthermore, services do not have to be altered prior to enactment (as is the case with current solutions, e.g. the W3C Web Services Choreography Description Language (WS-CDL), this offers greater flexibility and is a less invasive solution. Thirdly, sections of the choreography specification can be spliced in at runtime (e.g. the concrete services to call), allowing peers to operate in scenarios where it is not possible to explicitly define the pattern of interaction at design-time. Finally a practical contribution, MASC is implemented as a Java-based Web service choreography toolkit.
My doctoral work was primarily supervised by David Robertson, Christopher Walton and Austin Tate from the Centre for Intelligent Systems and their Applications (CISA) in the School of Informatics, University of Edinburgh.
Key publications:
Adam Barker, Christopher D. Walton and David Robertson. Choreographing Web Services. IEEE Transactions on Services Computing, volume 2, number 2, pages 152-166, IEEE Computer Society, April-June 2009.
D. Robertson, C. Walton, A. Barker, P. Besana, Y. Chen-Burger, F. Hassan, D. Lambert, G. Li, J. McGinnis, N. Osman, A. Bundy, F. McNeill, F. van Harmelen, C. Sierra, F. Giunchiglia. Models of Interaction as a Grounding for Peer to Peer Knowledge Sharing. In E. Chang, T. Dillon, R. Meersman and K. Sycara editors, Advances in Web Semantics, vol 1, LNCS-IFIP, 2009.
C. Walton and Adam Barker. An Agent Based e-Science Experiment Builder. In Proceedings of The 1st International Workshop on Semantic Intelligent Middleware for the Web and the Grid, European Conference on Artificial Intelligence (ECAI), Valencia, Spain, August 2004.
Grid Computing: Supporting Large-Scale Science
I have been employed on a number of research projects on the UK e-Science programme focusing on the application of distributed computing techniques to solve problems in automating large-scale science. I attended the International Summer School on Grid Computing 2007
5) Gene Expression Studies in Early Human Development
Human embryonic material is extremely scarce world-wide and currently only two laboratories are licensed in the UK to collect and analyse human embryonic tissue. Dissemination of the results to the wider community is therefore vital for the progression of research. Developmental Gene Expression Map (DGEMap) was a EU-funded Design Study, which accelerated an integrated European approach to gene expression in early human development. Working closely with Susan Lindsay from the Institute of Human Genetics at Newcastle University and Richard Baldock from the MRC Human Genetic Unit at the University of Edinburgh, I took responsibility for the Informatics research and development which supports the integration of distributed gene expression data.
Key publication:
Adam Barker, Jano I. van Hemert, Richard A. Baldock and Malcolm P. Atkinson. An e-Infrastructure to Support Collaborative Embryo Research. In Proceedings of the 9th IEEE International Symposium on Cluster Computing and the Grid (CCGrid), pages 520-525. IEEE Computer Society, May 2009.
6) Virtual Observatories: Automating Astronomy
The concepts of workflow have recently been applied to automating large-scale science (or e-Science), coining the term scientific workflow. Business workflow tools look more like traditional programming languages, and are, in general pitched at the wrong level of abstraction for scientists to take advantage of. My doctoral research worked closely with Robert Mann, an astronomer from the Royal Observatory on a number of Virtual Observatory projects: AstroGrid and the Large Synoptic Survey Telescope (LSST). As a result, my doctoral thesis evolved a number of concrete use-case scenarios and identified a set of core requirements for scientific workflow systems. My open-source Multi Agent Service Choreography (MASC) framework was successfully applied directly to AstroGrid and LSST in order to demonstrate proof of concept and evolve the software.
Key publications:
Adam Barker and Robert G. Mann. Flexible Service Composition. In Proceedings of Cooperative Information Agents X, 10th International Workshop, CIA 2006, Edinburgh, September 11-13, 2006, volume 4149 of LNCS, pages 446-460. Springer, 2006.
Adam Barker and Robert G. Mann. Agent-Based Scientific Workflow Composition. In Proceedings of Astronomical Data Analysis Software and Systems (ADASS) XV, volume 351 of Astronomical Society of the Pacific Conference Series, pages 485-488, 2006.
Industrial Research and Engineering
Complementing my academic research, I have worked at two industrial research laboratories, focusing on engineering industrial strength systems which have immediate application.
7) Distributed Extended Transactions: Hewlett Packard
Traditional ACID (Atomicity, Consistency, Isolation, Durability) transactions work well in tightly coupled homogenous environments, where they are typically short lived and message delivery can be guaranteed. However these tightly coupled atomic semantics do not fit in well with the architecture of the Internet, where message delivery cannot be guaranteed and transactions may be long running. The OASIS Business Transaction Protocol (BTP) is an extended transaction model that allows coordination of resources which are exposed by multiple autonomous organisations. This model relaxes the traditional ACID properties and eradicates the exclusive locking of a resource by a transaction. A research project conducted at Hewlett Packard Arjuna, resulted in an implementation of the protocol and demonstrated that the CORBA Activity Service was a sufficiently generic framework to support this complex extended transaction model. This work was in collaboration with Mark Little, now CTO at JBoss.
8) Modelling Real-time Systems: BAE Systems
As a joint project between the University of Newcastle and British Aerospace Engineering (BAE Systems), my BSc thesis involved designing and implementing a modelling tool for analysing real-time formal models, used to design aviation software. Implementation involved the use of formal methods (VDM-SL) and Java. This work was supervised by John Fitzgerald at Newcastle University and Stephen Paynter at BAE Systems, Bristol.