Individual security, contagion, and network design

Abstract

Individuals derive benefits from their connections, but these may expose them to external threats. Agents therefore invest in security to protect themselves. What are the network architectures that maximize collective welfare? We propose a model to explore the tension between connectivity and exposure to an external threat when security choices are decentralized. We find that both over-investment and underinvestment in security are possible, and that optimal network architectures depend on the prevailing source of inefficiencies. Social welfare may be maximized in sparse connected networks when under-investment pressures are present, and fragmented networks when over-investment pressures prevail.

Decentralized Defence of a (Directed) Network Structure

“…This is the reason why, in general, under random attack, decentralized and centralized defence allocations are likely to not coincide (Proposition 4). , Acemoglu et al (2016), Goyal & Vigier (2014), and Cerdeiro et al (2017) are among the closest papers to ours. They study a sequential game in which a designer moves first and chooses a defence allocation, and in a second stage the adversary chooses how to allocate attack-resources across the nodes.…”

“…More recently, new contributions have analyzed these games in a network setting (Acemoglu et al 2016, , Goyal & Vigier (2014) and Cerdeiro et al (2017)). In these models, payoff of the players is generally tied to a network structure which connects some of them.…”

“…They study a sequential game in which a designer moves first and chooses a defence allocation, and in a second stage the adversary chooses how to allocate attack-resources across the nodes. In Cerdeiro et al (2017), the authors also discuss how the designer could optimally design the network in order to solve possible inefficiencies arising when security choices are decentralized. In particular, the authors show that decentralised security choices could lead to both over and under-investment in security.…”

Solving Security of Cyber-Physical Systems with Multi-Objective Optimization Algorithms

“…Cerdeir et al presented a partial characterization of the network to analyze welfare under equilibrium and then induced full protection [6]. Wu et al studied the security detection in fusion-based methods.…”

Individual Security and Network Design with Malicious Nodes

“…One of the key findings is that the externalities resulting from security decisions made by selfish agents lead to, potentially significant, inefficiencies. This motivates research on methods for improving information security, such as insurance [BS10] and network design [CDG14,CDG17]. We study the problem of network design for interdependent security in the setup where a strategic adversary collaborates with some nodes in order to disrupt the network.…”