University of Twente Student Theses

Login

Network games, information spread and endogenous security

de Witte, B. (2015) Network games, information spread and endogenous security.

[img]
Preview
PDF
3MB
Abstract:A digital security breach where confidential information is being obtained often does not only influence the agent whose system is being infiltrated, also other agents socially or digitally connected to the damaged system can be affected. This leads to externalities in which the actual security of an agent does not only depend on his own investments in security, it also depends on investments by other agents. As no agent would invest to protect information of others, misaligned incentives may appear in this setting. It has been argued that these incentives provoke under-investments, which in turn make the network prone to fail. Recently this presumption is challenged however by the introduction of an intelligent adversary who chooses an optimal trade-off between expected damage and precision of the attack. In this research we build on the impact of an intelligent adversary by combing the strategic attack model with a new model for information spread. We show that agents tend to compete for security under the strategic attack as an increase in investments can cause the adversary to attack someone else. When dependencies among agents are low, because the network is not very dense or because the probability that information is shared is small, agents even invest more in security than they would in the social optimum. When dependencies increase these over-investments prevail due to a second order force originating from the adversary. In this situation the adversary chooses a more precise attack as the expected gain compensates the increased costs for a more precise attack. Nevertheless, when dependencies continue to increase and consequently it becomes meaningless to discourage an attack, at some point over-investments pass on to under-investments. We show that this point is reached earlier in a more dense network. However, when the network consists of several components where the intelligent adversary can strategically decide which component is attacked, investments are always higher in a more dense network.
Item Type:Essay (Master)
Clients:
Centraal Planbureau, Den Haag
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:31 mathematics, 83 economics
Programme:Applied Mathematics MSc (60348)
Link to this item:http://purl.utwente.nl/essays/66944
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page