Abstract

Networks are a crucial and integral aspect of modern enterprises, with the productivity and fate of many enterprises significantly intertwined with the performance and reliability of their networks. Yet despite their importance, these networks remain highly susceptible to poor performance, failures, and outages.

I posit that the fragility of enterprise networks is largely a byproduct of the complexity involved in managing conflicting global policies. In this talk, I argue that only by understanding this complexity can we harden these networks. In my research, I have developed simple yet highly effective models of enterprise networks and used these models as first order principles in designing practical frameworks that improve the performance and reliability of these networks. In the first half of this talk, I will present a set of novel models that capture the difficulty of implementing high level network objectives within the network. In the second half of the talk, I will focus on using the frameworks arising out of these models to mitigate complexity by informing the design of more efficient and reliable services and by enabling large enterprises to debug their network’s configuration.

Together these studies illustrate how eliminating complexity can lead to manageable networks, thus reducing misconfigurations and promoting innovative changes. Understanding complexity has several benefits: in the short run, knowledge of complexity leads to the development of mechanisms that foster better performance and higher resiliency; and in the long run, the dialogue created by understanding complexity will shape the design of future networking interfaces and abstractions.