Network of Excellence in Internet Science

Towards a Theory of Internet Science (JRA1)

About this Working Group

This activity is about the development of a multidisciplinary scientific approach to the understanding of Internet networks that is envisaged in Internet Science. Despite the rise of so-called “network science” in the last decade, a range of underdefined issues and unresolved fundamental differences in discipline-bound approaches limits the applicability of this network science across different disciplines, in particular in the social sciences and humanities. The Internet involves humans, societal groups, institutions and ideas; these entities, and the actual or potential links within and among them, are every bit as important as the technological layer (networks of machines and other ‘things’, and the communication infrastructure itself) through which they connect. A theory of networks should thus take into account academic insights produced in disciplines dealing with different aspects of human behaviour (e.g. sociology, anthropology, economics, political sciences etc.), along with so-called ‘hard sciences’ (e.g. mathematics, physics) and information and communication-based disciplines (e.g. information theory, performance analysis, etc.) if it is to understand the evolution and behaviour of networks considering both “forward” (human to network) and “backward” (network to human) interactions. A key element in the Network Science approach that will be taken into consideration is that, by observing similarities, differences and interactions among different types of "network", we can discover fundamental rules and principles applicable to a large class of networks regardless of particular technology, protocol, social background, economic situation, etc. But, along this dimension of increased generalization, it is unavoidable that we lose the fine details of context.

In other words, we aim for a continuous “conceptual oscillation” between extracting generic elements and highlighting their role in a better understanding of concrete human behaviour patterns. To give an example of how to operationalize such an approach on the methodological level, for instance we envision using pattern recognizing techniques (on aggregated level) as indicators for in-depth, context-rich analysis (on a micro-level). An example is the study of resilience classes (the aggregated behaviour of a network in response to a challenge) in the context of the technical, economic/policy, sociological and legal conditions pertaining at the time (the concrete, micro level). Another example is to combine elements from social network analyses (heterogeneity of agents and how their social and behavioural attributes influence their network-role) with structural and statistical analysis as done by complex networks theory and statistical physics. Moreover, the foundations of the Internet as a technical infrastructure are often not taken into account when analysing (social) networks supported by this infrastructure. Our aim, and first objective, is to bridge the different disciplinary perspectives on the functionalities, evolution and applications of the Internet, seen as a socio-technical system. The main objective is to understand how technological, policy, economic and social elements interact with each other and condition the evolution of Internet systems, and thereby contribute to creating the conditions for further innovative services and functionalities to appear. Second, it is important to discover fundamental models of network graphs (where networks are to be intended as virtual as well as physical) to understand and explain why they have developed their current structure, activity and impacts, and how they will evolve in the future. When looking at the evolution of the Internet both as a technological and social artefact, research at the level of web developments (sometimes called “Web Science”) and also of patterns of connectivity and/or traffic graphs, can help understanding the global picture and the broader research framework. A central tenet of the interdisciplinary investigation in the Network of Excellence in Internet Science is the set of linkages between microscopic (local) interactions and large-scale behaviour, such as how user selfishness and sociability, physical constraints on network deployment and operation, and fundamental theoretical constraints (including the limited memory, cognitive capacity and ‘rationality’ of network users) influence network structure, sustainability and evolution. Third, a key topic of Internet Science regards the cooperation for the production, distribution, and consumption of “information”. In the abstract, information holds particular properties of universality and infinite reusability, in contrast to ordinary goods and services. It also offers different types of externality – the value of information for one party may increase or decrease if others know it, and beliefs and expectations may be as important as ‘hard’ information. Cooperation in networks can result in emergence of global distributed intelligence from local interactions.

At the broader level, this "collective intelligence" can expand the current levels of human cognition and give rise to new forms of social organisation and knowledge creation. In computing and network sciences, this can even be used to control the network itself (e.g. to reduce congestion by spreading the traffic), and constitutes the base for building a self-organising and resilient network capable of efficient recovery from (accidental or deliberate) network failures, and detecting, isolating and/or correcting nodes whose behaviour causes problems in the network. Understanding this cooperative intelligence (e.g. whether it resembles the intelligence of individuals in the network) is central to Internet Science in association with grid and high performance computing. One often-overlooked feature is the cost of networking. In complex systems, the cost of network activities has to be balanced against the benefits of distributed actions. On the other hand, to be able to estimate “costs” that participants are willing to pay, we need to add a layer of meaning to the notion of “information”. Thus, in this project we address syntactic, semantic, social and economic aspects of information. Conversely, in the backward dimension, the operation of economic, societal, etc. systems in the Internet provides a huge set of empirical data, which can be seen as an experimental platform.

The capacities, standards, and so forth arising in the ‘technical Internet’ change the costs and benefits in comparison to off-line environments, and the beliefs on the uptake and application of these capabilities. In turn, this affects manufacturing practices, collaboration paradigms, marketing, and the creation of intangible financial assets. The above activities will be immersed in a reflexive theoretical framework characterised by the interplay between two deeper dimensions. The first dimension is concerned with questions of ontology and epistemology, i.e. the starting assumptions about what the Internet is made of and how knowledge in and about the Internet is constructed in the different disciplines comprising Internet Science. The second is concerned with questions of social and economic value, i.e. what are the fundamental human and societal drivers for Internet developments. Whereas the first set of questions will address the challenge of communicating not just across different disciplines but across different epistemologies, the second set of questions will draw from the field of economic anthropology and other social sciences to seek the right balance between the ‘base’ or ‘commons’, and the society and market in the context of the online knowledge economy and of the emerging cultural forms and social organizations of the Future Internet. The ultimate goal of this Internet Science NoE will be a better understanding of the complex nature of Internet networks, services and applications, and of their design based on desirable social, economic or environmental objectives.

JRA1 Deliverables

Deliverable 1.1 What challenges for Internet Science?

Deliverable 1.3: Graph-theoretic challenges in network science

Deliverable 1.4.1: JRA1 Workshop Organization

Deliverable 1.7.1: 1st Report on integration and excellence building in JRA1

Deliverable 1.2: Roadmap for an economic theory for information networks

Deliverable 1.4.2: Report on the Interdisciplinary Workshop on Decision Making

Deliverable 1.5: Roadmap on developing the fundamental basis of autonomous and dynamic complex networks

Deliverable 1.6: Roadmap for collective network intelligence

Deliverable 1.7.2: 2nd report on integration and excellence building in Theory of Internet Science

Deliverable 1.7.3: 3rd report on integration and excellence building in Internet as Critical Infrastructure