 |
Center for Nonlinear Studies |
The Internet is a world of danger: organizations are under constant attack, computers remain vulnerable to infection, a dark force is rising (e.g. embodied by botnets), and governments start to enter this game with almost unpredictable potential consequences. While implicitly understood, the mechanisms of (in)security remain poorly understood at the Internet scale, and it is not clear what is the global state of (in)security of the Internet at a given time. In the talk I construct an immune-system like model, that incorporates all basic ingredients of insecurity dynamics (vulnerability discovery and prediction, time to patch release, delay in patching installation by users) and make reasonable assumptions for the growth of the threat after a vulnerability has appeared. I also introduce three possible kinds of security responses and build plausible (in)security scenarios. Although this model needs further empirical validation, it clearly depicts the mechanisms leading to sharp dynamics in the evolution of Internet (in)security. It shows that, while ubiquitous, cyber-threats have generally relatively short time windows to develop and operate in order to perform massive attacks, but they can easily exploit the long tail for parasiting. These results quantify the utilities and opportunity costs that have led to the birth and rise of the black market for vulnerabilities/stolen data - the cornerstone of today's cybercrime. More broadly, it shows how people may take advantage of a "windfall effect" triggered by weaknesses in policies (legal, software, etc) and the potential economic costs for society.
Host: Alexander Gutfraind