This work deals with the use of the Network Robustness Simulator (NRS) to assess and compare the robustness of a set of networks. The NRS is able to compute a number of graph metrics which are put together in a single value $R^*$ by using the principal component analysis and generating heat maps as a visual result. The main features of the attack models are explained as main tool to analyze the robustness in a dynamic way.

Currently NRS implements random, targeted, epidemic and cascade failures. The main characteristics of these models are addressed.

Interconnection among telecommunication networks and other critical infrastructures (CIs) usually is provided through nodes that are spatially close, generating a geographical interdependency. As result of this interdependency, in previous works are shown that for a given radius (r) the maximum network's robustness to cascading failures is achieved, despite the high deployment cost associated to the large number of interlinks. In this letter, an enhanced interconnection model to limit the number of interlinks for geographical interdependent networks is proposed. The proposed interconnection model considers the maximum number of interlinks that a node can handle and the nodes' risk to failures or attacks. Thus, our model yields promising results to maintain an acceptable level in network robustness under cascading failures by reducing deployment cost associated to decrease the number of interlinks.

During the talk a brief technical explanation about how the Network Robustness Simulator (NRS) has been implemented is introduced. Then a demonstration is done, showing how the NRS automatically executes a set of simulations and how the visualizer is able to deal with both single network and in interdependent networks backgrounds.

Reliability analyses of water distribution networks conducted on geometric-, hydraulic- and economic grounds all reflect the complexity of network resilience that is still difficult to generalize or classify, as well as it is impossible to assess without full integration of all three aspects. Practical assessments of hydraulic reliability usually consider the demand loss from consecutive failure of single pipe ending in a form of network resilience index with the value between 0 and 1, which is representing a kind of averaged consequences. This concept can further be expanded by analyzing the effects of increased investment and/or operational costs on the reliability improvement. It is mostly favourable due to its simplicity and having powerful computational tools able to quickly calculate the index of even very large and complex networks. Nevertheless, the average index is potentially hiding the weak points in the network as well as the condition of the network expressed by mechanical reliability i.e. the failure probability may inadequately be captured.

The presentation discusses two commonly applied indices and the improvement suggested by the visual representation of the network reliability, through so called Hydraulic Reliability Diagram. The concussion point the improvements but also spell the need for more rigorous (mathematical) verification of the established hypotheses.

SURFnet has been operating its own WDM transport network for more than 15 years. Recently, SURFnet has started to deploy a new photonic transport layer. Deployment of this network will be dominantly brown-field with active services. This imposes challenges on the migration strategy in order to guarantee availability and reliability of the transport services to depending platforms and customers. In this presentation SURFnet will touch on the strategy and measures taken to make the migration robust.

Power grids form an integral part of a city's infrastructure. With evolving technologies like smart grids, it becomes more important to assess the system's potential to respond quicker to failures and volatile loads. A bottom up approach involving evaluation of component level risk assessment and scaling up to a system level risk assessment was chosen. Transformers form an integral part of the infrastructure of any electrical power system. With a growing concern for asset management strategies and predictive maintenance, it becomes highly essential to study the reliability of the components and estimate their lifetime when subjected to operation. The objective of the research is to predict the scale of the risk subjected on the transformer under continuous operation with the influence of varying temperature and load conditions.

Transport and logistics systems are the blood vessels of our economy, catering for primary mobility needs of people and firms. In this talk I explore robustness issues in these networks, focusing on system level impact analysis. I introduce the main impact mechanisms of disturbances, including common models used to quantify these impacts. I describe some recent cases of empirical, quantitative analysis of robustness of complex networks. Finally I propose a list of research issues that require co-operation between scientists specialised in complexity analysis and in transportation system analysis.

Large scale emergencies such as a large scale flooding or a devastating storm can severely impact Critical Infrastructures (CI). Since a subsequent disruption or even destruction of CI may increase the severity of an emergency, it is important for emergency management organizations to include an impact assessment of CI in their risk management, emergency planning, and emergency preparation.

In this presentation, I will discuss a methodology for emergency management organizations to assess the impact of CI disruptions. The methodology takes into account common cause failures affecting multiple CI concurrently, cascading effects as well as the potential disruptive impact on emergency management operations. The methodology builds on earlier CI assessment approaches at the national level and has been tested in a case study.

The continuous availability of critical (information) infrastructure is crucial for the well-being of modern society. The ongoing digitalization of critical infrastructure systems puts their continuity at risk by introducing vulnerabilities. Many critical infrastructures consist of a supply chain of multiple actors that, together, bring about a critical process (such as energy supply, drinking water, communications, financial services or transport). The cybersecurity challenges and solutions for newly introduced vulnerabilities can be addressed at the level of individual actors in a critical supply chain, at the supply chain or network level – through a collaborative approach – , or a combination of both. This presentation focuses on typical cybersecurity challenges in critical supply chains and discusses how to identify the right level to implement effective and feasible solutions.

We investigate whether it is possible to exploit the existence of a largest connected component of the sexual contact network of men who have sex with men with syphilis in San Francisco to design new notification strategies that can potentially contribute to reducing the number of cases. We develop a model capable of incorporating different types of notification strategies and compare the corresponding incidence of syphilis. Through simulations on weighted networks constructed from real data, we show that notifying the community of the infection state of few central nodes appears to be the most effective approach, balancing the cost of notification and the reduction of syphilis incidence. We also analyze the robustness of the considered notification strategies against the addition of links.

Kemeny's constant and its relation to a robustness metric, called the effective graph resistance, is found in literature for regular graphs. Based on Moore-Penrose pseudo-inverse of the Laplacian of a network, we prove a closed-form formula for the Kemeny's constant and derive the relation between Kemeny's constant and the effective graph resistance for a general connected, undirected graph. Physical meaning of the Kemeny’s constant is presented.

Determining a set of "important" nodes in a network constitutes a basic endeavour in network science. Inspired by electrical flows in a resistor network, we study the diagonal element $Q_{jj}^{\dagger}$ of the pseudo-inverse matrix $Q^{\dagger}$ of the weighted Laplacian matrix of the graph $G$. We propose a new graph metric that complements the effective graph resistance $R_{G}$ and that specifies the heterogeneity of the nodal spreading capacity in a graph. We also highlight an exact relation between "structure" and "function" of a network, based on the effective resistance matrix of a graph.