Dr. Horacio González-Vélez

Associate Professor and Head of the Cloud Competency Centre
National College of Ireland

Elastic Algorithmic Skeletons: Enhancing Structured Parallel Programming for Clouds and GPUs

This talk presents an overview on how to efficiently program structured parallel applications for heterogeneous architectures (e.g. clouds or GPUs) by introducing elasticity through autonomic resource awareness.

Widely considered as the logical successor of the internet, cloud computing entails the exchange of computer and data resources across global networks. It constitutes a new value-added paradigm for network computing where higher efficiency, massive scalability, and speed rely on effective software development. Furthermore, GPUs and other non-traditional accelerators have helped to enhance parallel application performance at a significant software development cost. Nevertheless, it has become evident that improved programming models are needed.

This talk reports on research endeavours which incorporate structural information into a parallel application by employing algorithmic skeletons. Skeletons abstract commonly-used patterns of parallel computation, communication, and interaction. By means of resource awareness, skeletons can help structured parallel applications to adapt to the most suitable computing resources regardless of its architectural nature, whilst minimising the associated costs and time.

Biography of Horacio González-Vélez:

As associate professor and head of the Cloud Competency Centre at the National College of Ireland in Dublin, Horacio González-Vélez directs the NCI's cloud infrastructure, postgraduate programmes, and research with a clear emphasis on innovation, education inclusion, and industrial partnerships.

He spent over a decade working in engineering and product marketing for innovation-driven companies such as Silicon Graphics and Sun Microsystems. Award-winning lecturer and researcher, Horacio has also carried out applied research in parallel and distributed computing, funded by a number of public and industrial organisations including the European Commission FP6 & FP7, UK NESTA, the Scottish Funding Council, NVidia, Dell, and Microsoft.

He is an accredited expert on cloud computing for the BCS, the UK Chartered Institute for IT. He has been recognised with the European Commission ICT award for his efforts on scientific dissemination and the UK NESTA Crucible Fellowship for his inter-disciplinary research on computational science. He holds a PhD in Informatics from the University of Edinburgh.

Prof. Alfredo De Santis

University of Salerno, Italy

New Frontiers in Anomaly Detection: Staying Tuned with Emerging Security Threats

With the widespread development of the Internet and the ever increasing success of network-centric applications in substantially any sector of the everyday’s life, the need of detecting and preventing network-originated abuses in real time is more and more important, in order to guarantee appropriate protection and reaction capabilities against the huge amount of emerging security threats. Many solutions and tools are already available for checking the traffic flows and spotting anomalous usage patterns or malicious behaviors, but most of them are substantially unable to cope with unknown phenomena (e.g., 0-day attacks). In fact, the concept of normal or anomalous behavior is a very elusive one, since it depends on a large number of variable factors, often not immediately evident, associated to usual network activities and resource usage. Hence, a new generation of self-learning models that adaptively consider and understand the hidden relationships between these factors and the innermost dynamics underlying the involved networks and applications, is needed in order to effectively recognize security threats and react to them.

Accordingly the main focus of this keynote is outlining, by starting from a survey of the currently available approaches, the new direction for future developments in anomaly detection. This will be accomplished by considering the problem from several perspectives, ranging from the traditional analysis techniques, based on acquiring knowledge about normal and anomalous events through a statistical idealization of past observations, to newer approaches based on chaos theory and nonlinear analysis. The latter approaches can be more helpful in understanding the above phenomena by observing them simultaneously on multiple time scales, thanks to the projection of the observations in a multi-dimensional and more feature-rich space.

All these techniques are empowered and complemented by a host of methodologies originally born in the machine-learning framework and adapted for providing a deterministic interpretation of the complex traffic dynamics associated to normal and anomalous events.

Additional future research opportunities originate from the consideration that different properties and criteria could be applied for inferring more sophisticated traffic features, that can be used to obtain a deeper and more comprehensive understanding about traffic profiles associated to security-related events.

Biography of Alfredo De Santis:

TBA

Ms. Reka Balogh and Mr. Rob Porter

Cisco Systems

Software Defined networks

The rapid and evolving developments in computerized IT systems place ever increasing demands on customers and their IT departments to deliver: new ways of providing services, applications and innovating to remain competitive while addressing the increased complexity that characterizes IT systems.  

In meeting these challenges, which include: management complexity, sprawl of equipment and scaling issues, the nature of the challenges and potential solutions must be considered. For example, an IDC Report in 2011 identifies that: infrastructure to support these demands will grow 10 times its current size before 2020, however, the growth in trained staff is expected to be only 10% of this. These demands and growth demonstrate that a new approach is required.

One such approach which is gaining traction is: Software Defined Networking (SDN). This approach to the provision of network services is perceived by many to be the ideal solution given the rapid evolution in the client and user demands.  We will provide insight into what the industry means by SDN, the different approaches that are in the market, how it is influencing customers, research opportunities, hardware providers, job roles, skills requirements, new businesses and ultimately the impact it has on industry and commerce.

Biography of Reka Balogh:

Reka is a Data Centre sales specialist for Cisco Systems, where her area of focus is the UK and Ireland. Her specialisation has centred around the Data Centre industry for her entire professional career. She has extensive experience and knowledge in Data Centre requirements and specifying solutions to meet customers’ defined needs.
On a daily basis, Reka is involved in shaping customer futures, understanding their goals and objectives, describing industry trends and directions, and ultimately identifying key Cisco solutions to help customers realise these goals.  

Biography of Rob Porter:

Rob began his career at Cisco Systems in 2001 and has significantly influenced not only Ciscos’ partner community and customers, but also Cisco internally. Rob has a strong background in networking and software development; he has developed comprehensive technical knowledge of Data Centre technologies, including storage, virtualisation and servers.

Rob was heavily involved in developing the Cisco Network Academy Program for 6 years, during which he spent a significant amount of time training Cisco technical experts on a global basis. Rob has extensive industry knowledge and a detailed understanding of the challenges that customers' face in utilising Data Centre technologies. In his role he has developed significant expertise and experience in the design of ‘next generation’ Data Centers.