Positive feedback and the madness of crowds
Hilarie K. Orman, Richard Schroeppel · 1996
This paper discusses the necessity of bringing security measures to very complex systems, and some ideas for how to do so.It has long been the hope that we would never reach such a situation -that security technology would eventually succeed in harnessing software technology and bringing it to heel.This has not happened, and it is now time to expand security technology to cope with the seemingly intractable problems of huge heterogeneous systems.The consideration of paradigms from other fields may be useful.This paper hopes to motivate research into a fundamental and difficult problem that affects Internet security today: unpredicted or unintended dynamic behavior that can degrade or destroy the ability of the network to deliver messages.We will argue that the source of the problem is hidden variables that adversely affect composition of network systems at any service level: physical, link, network, transport, application, and even the end-user.