Program overview

	Monday (July 7)	Tuesday (July 8)	Wednesday (July 9)	Thrusday (July 10)
9:30-11:00	Erol Gelenbe (ICL) Fundamental Problems in Autonomic Communications	Secondo Montrucchio (Sheffield University) New services, socio-economic impacts and future perspectives	Fabrice Saffre (BT) Designing for "Self-*"	CASCADAS Demo Day (10:00-16:00)
	Coffee break	Coffee break	Coffee break
11:30-13:00	Stefano Crosta (CISCO) Realistic applications of AC for networking solutions	José Halloy (ULB) Self-organized decision-making in group living animals as an inspiration for designing autonomic systems	Peter Deussen (FOKUS) Supervision of Autonomic Systems
	Lunch	Lunch	Lunch
14:30-16:00	Workshop sessions	Roberto Battiti and Mauro Brunato (UNITN) on Self-preservation and reactive search	Workshop sessions

Jump to the detailed tutorial program.

Jump to the detailed workshop program.

Jump to the detailed CASCADAS Demo Day program.

Detailed tutorial program

Fundamental Problems in Autonomic Communications

Tutorial by: Erol Gelenbe (Imperial College London)

Abstract:
We will address some fundamental problems in autonomic communications such as "can one estabish communications in an unkown infinite network environment", and "what should one do at each decision step when one only has limited information". These issues will be addressed using mathematical models and also with practical examples that use the Cognitive Packet Network Test-Bed. This research has been supported by the FP6 CASCADAS Project.

Scheduled for: Monday, Time slot 1

Realistic applications of AC for networking solutions

Tutorial by: Stefano Crosta (CISCO Reserach)

Abstract:
Autonomic Computing principles seen in practical applications for networking solution: self-organization and self healing from routing to services. What are the concepts that made it into implementation in the industry, and what can AC research bring next?

Scheduled for: Monday, Time slot 2

New services, socio-economic impacts and future perspectives

Tutorial by: Secondo Montrucchio (Sheffield University), Paola Fantini (Politechnico di Milano, School of Management)

Abstract:
Internet is changing the world and the world is changing internet: new paradigms are emerging, which can be analysed through several dimensions of the technology trends, in close connection with economic and social progress.
A co-evolution is actually taking place, involving the extent of connectivity, the quality of human-computing interaction, the degree of embedded intelligence, the environmental pervasiveness, the security mechanisms enacted within the web, on one side, and the behaviour of people, companies and institutions, on the other side.
Early warnings of this co-evolution can already be seen in relation to the current spreading of the Web 2.0 technologies, by which the web users are evolving from simple web content users to connected social net-workers.
This development, enabled also by new accomplishments in nanotechnology, micro-electromechanics and semantics, is leading towards a possible empowered world: the "internet of People and Things", marked by an enhanced degree of knowledge and control, in which creativity will be exploited anytime and anywhere.
In the future, the human-machine interaction will become more and more "seamless", indicating with this term an invisible mediation, characterized by the absence of all that today seems obvious and that is involved in our communication network activities.
In this landscape, a main role is owned by the EU, which can support and drive this foreseen development in compliance with its main objectives of bringing benefits for Europe's citizens, businesses, industry and governments. In particular, the EC strategies aim at reducing the digital divide and social exclusion and at establishing a Single Market for both information society and media all over European Countries.

Scheduled for: Tuesday, Time slot 1

Self-organized decision-making in group living animals as an inspiration for designing autonomic systems

Tutorial by: José Halloy (Liberal University of Brussels)

Abstract:
Animals living in groups face the issue of making decision collectively. However, animals are limited, by natural constraints, in the type and amount of information they can gather and process. Nevertheless, group living animals produce remarkable collective behaviors. We will show, in some experimental cases, that these collective patterns are self-organized meaning that they do not require a highly structured society or some form of leadership and centralized information. Even when the individuals are all equal and base their actions on limited information, the group as a whole is capable of organizing thanks to stochasticity in individual behavior and appropriate non-linear social feedbacks between the members of the group and their environment. When these principles are used to design artificial distributed systems, they can produce the same kind of self-organized patterns and thus be used as a source of autonomy. We will discuss some examples of this ongoing research field in engineering.

Scheduled for: Tuesday, Time slot 2

Self-Preservation and Reactive Search

Tutorial by: Roberto Battiti and Mauro Brunato (University of Trento)

Abstract:
A distributed architecture based on autonomic elements and aggregated services can be the target of disruptive actions both from the outside (identity spoofing, denial of service) and from the inside (malicious and selfish participants). Preventing such attacks in a highly dynamic environment requires fast detecton and ready action. In this context, pattern recognition and learning techniques are extremely useful.
Reactive Search advocates the integration of sub-symbolic machine learning techniques into search heuristics for solving complex optimization problems. The word "reactive" hints at a ready response to events during the search through an internal online feedback loop for the self-tuning of critical parameters. Methodologies of interest for Reactive Search include machine learning and statistics, in particular reinforcement learning, active or query learning, neural networks, and meta-heuristics.

Scheduled for: Tuesday, Time slot 3

Designing for "Self-*"

Tutorial by: Fabrice Saffre (British Telecom)

Abstract:
With the complexity of network-based services and the diversity/dynamicity of hosting environments both increasing exponentially, managing the whole life-cycle of a distributed application (deployment, updates, migration, replication, load-balancing...) through centralised planning and control is fast becoming an impossible task. In this context, design principles exploiting the emergent properties of large ensembles of autonomous units represent an attractive and radical alternative to traditional approaches. In this tutorial, we will demonstrate the ability of fully decentralised systems to self-organise and exhibit desirable properties such as scalability, robustness, adaptability etc.
The key to the successful design of self-* systems is to pay careful attention to choosing and calibrating the local rules governing the behaviour of individual elements (service components, device managers, mobile robots...). These rules must be engineered to take full advantage of collective dynamics when executed in parallel by a large number of interacting units. A combination of modelling and simulation techniques allows us to understand complex dynamics, identify useful features and mitigate the effect of counter-productive ones. We will illustrate these methods across four classes of algorithms: self-aggregation, differentiation, synchronisation and collective decision-making.

Scheduled for: Wednesday, Time slot 1

Supervision of Autonomic Systems

Tutorial by: Peter Deussen (Fraunhofer Institut FOKUS)

Abstract:
The ongoing increase of complexity of service provisioning platforms in the converging Telco and Internet Service domain requires the development of novel approaches for self-management. Supervision, i.e. the continuous monitoring of a system (or a class of system executions, i.e. a service), and the selection or generation and application of appropriate reaction patterns if the system enters (or is in danger to enter) a problem situation or unwanted state, is the basic mechanism for activities related to failure management, accounting, security, and performance evaluation and optimisation.
In the current cloud of approaches in the "Autonomics" domain, two main streams can be identified: The "cognitive approach" where supervision is performed using some model structure (comprising a system's self-model and a model of its environment) for decision making, and the "self-organisation approach", where supervision is part of the behaviour of a system which emerges from the interaction of a (probably large) number of entities equipped with relatively simple rule systems and local knowledge of the states of the neighboured entities.
In this talk we introduce control loops as basic structure of supervision activities and discuss this notion by means of several examples. We then discuss a principle framework for a cognitive supervision approach based on model abstraction and model composition. Moreover, we introduce an autonomic configuration mechanism for supervision enhancements based on self-organisation algorithms. Our thesis is that the cognitive and the self-organisation approach are complementary in the sense that each approach has the capability to support each other.

Scheduled for: Wednesday, Time slot 2

Detailed workshop program

Self-Organizing E-Business Systems

Paper by: Alfons Salden

Abstract:
In the future e-society the need for sustaining business collaboration will grow ever more rapidly. Unfortunately the proposed e-business architectures and infrastructures are still too immature to partly automate the choreography and orchestration of services required in e-business collaboration. Collaborating companies have to seamlessly interact and communicate in order to create and provide as an enterprise differentiated products and service portfolios to their customers and to ensure their competitive edge on the markets. Innovation through collaboration asks for very highly sophisticated development support during creation and integration of such portfolios across e-business networks. We present a collective intelligent agent system architecture and infrastructure that can sustain both choreography and orchestration across all layers of an evolving e-business network. Such autonomous intelligent agent systems define and make operational business categorization systems - they enable indexing, query interfacing, retrieval, association, (re-) generation, (re-)structuring and (re-)organization of enterprise cycles. Furthermore, they can define and make operational natural anticipatory and selection of attention (pre-) schemes that can support and/or automate e-business collaboration.

Scheduled for: Monday, Time slot 3

Intelligent Flooding for Inter-car Communication

Paper by: Zsolt Pocsaji and Robert Szabo

Abstract:
In this paper we present a design and architecture for intelligent car-to-car message-distribution to enhance traffic safety. The proposed algorithm is part of a middleware utilizing the global positioning service (GPS) and WiFi based ad-hoc communication. The middleware enables traffic related user applications to be developed easily above the core messaging functions. The designed architecture has been prototyped and some measurements were performed to functionally and performance wise validate the system.

Scheduled for: Monday, Time slot 3

Distributed Service Framework for Digital Cities

Paper by: Laura Ferrari, Antonio Manzalini, Corrado Moiso and Iacopo Carreras

Abstract:
In this paper, we explore the main advantages deriving from the application of a Distributed Service Framework to future ICT and Telco realms. We do believe that this is a necessary step to be explored in order to keep the pace of Telco Network and Service platforms emerging trends. We present the architectural choices developed in the European project BIONETS, and provide an example of how such frameworks can be applied to a Digital City application scenario.

Scheduled for: Monday, Time slot 3

A Knowledge Network Toolkit for Autonomic Componentware

Paper by: Matthias Baumgarten

Abstract:
With the dawn of smart world infrastructures on a global scale, the need for fully autonomous operating systems and services has never been more urgent. A key aspect for such systems is the availability of relevant contextual information so that they can autonomously configure, adapt and optimize their behavior towards changing conditions. This is known as context or situation awareness, which is of fundamental importance for successful autonomic computing and services. However, acquiring relevant information from the real, virtual or operational environment is only the first step to facilitate such contextual awareness. Additional processing is required to pre-organize, correlate or simply reformat such data so that they are readily accessible as well as usable by a multitude of services. Within a distributed environment, this translates directly into a diverse network of knowledge, where individual views correspond to specific contexts and situations that are ultimately understandable yet manageable in real time. This paper, discusses a Knowledge Network (KN) approach that has been developed as part of an Autonomous Componentware toolkit, called ACE-Toolkit (Autonomic Communication Elements). The reference based KN framework provides the means for the ACE-Toolkit and its components to acquire a higher degree of contextual awareness where knowledge from various sources can be utilized efficiently at various levels of granularity. On the other hand the KN components themselves are realized as ACE's thus taking full advantage of the autonomic features offered by the toolkit.

Scheduled for: Wednesday, Time slot 3

Purposive agents for self-organized Network Management Systems

Paper by: Pasquale Donadio, Antonio Cimmino and Bela Berde

Abstract:
Purposive agents are emerging as feasible and attractive mediators for facilitating Grid services in a heterogeneous Grid environment. For improving the efficiency of resource usage in a dynamic networking environment, an adequate multi-agent based computational paradigm is adopted, and an effective network management model based on the multi-level agent organization is proposed. Regarding operational details, we present a purposive infrastructure of autonomous systems, based on agent-mediated Grid services in a network management application domain. The composite purposive environment is used to define individual mediator agents, but also to model, identify, derive, organize, and coordinate multi-agent systems in the dynamic Grid environment. We have, finally, designed and implemented a prototype of a Network Management System that validates our thesis.

Scheduled for: Wednesday, Time slot 3

Mapping Ecology to Autonomic Communication Systems

Paper by: Beatriz Otero, Pere Barlet, Salvatore Spadaro and Josep Solé-Pareta

Abstract:
IBM presented (in 2001) the idea of autonomic computing: many different ways of interacting. Self-governing components can simplify configuration, healing, optimization and protection of IT systems (thus hiding complexities to human operators). Today, Autonomic Technology also refers to the self-managing characteristics of resources as such the capacity of hiding completely its complexity to operators and users. Systems make decisions using high-level policies from operators. They will constantly check and optimize their status and automatically adapt themselves to changing conditions. Autonomic technologies may represent promising solutions for the evolution of Future Internet, ICT and Telecommunications. This paper addresses the problem of designing future service frameworks based on autonomic technologies and leveraging bio-inspired laws, algorithms and patterns naturally existing in the ecology of living entities. We first define an overall architectural model and then the autonomic abstraction. Next, we study the principles and capabilities of autonomic communication systems based on ecological patterns and propose examples of how to map these laws into the architecture. This study is illustrated with two use cases that can be easily prototyped to test the feasibility of the model.

Scheduled for: Wednesday, Time slot 3

Detailed CASCADAS Demo Day program

Scheduled for: Thursday, 10AM - 4PM

During the demonstration day the main achievements of the European funded project CASCADAS will be presented. The project is part of the call "Situated and Autonomic Communication" of the Future and Emerging Technology (FET) Programme financed by the European 6th Research Framework (IST).

The main goal of CASCADAS is to develop an autonomic component-based framework to enable composition, execution and deployment of innovative services capable of flexing and coping with unpredictable environments by dynamically self-adapting to situation evolutions.
To find out more please read our white paper describing the research done in the project: Bringing Autonomic Services to Life.

Program:

• 10H-13H: Demonstration session
• 13H-16PH: Lunch followed by open discussions

Read more...