Building Secure and Reliable Systems

Chapter 2. Understanding Adversaries

By Heather Adkins and David Huska

with Jen Barnason‎

In August 1986, Clifford Stoll, a systems administrator at Lawrence Livermore Laboratory, stumbled upon a seemingly benign accounting error that led to a 10-month search for someone stealing government secrets from the United States.¹ Largely considered to be the first public example of its kind, Stoll spearheaded an investigation that laid bare the specific tactics, techniques, and procedures (TTPs) the adversary used to achieve their goals. Through careful study, the investigation team was able to construct a picture of how the attacker targeted and siphoned data out of protected systems. Many system designers have incorporated lessons that arose from Stoll’s seminal article describing the team’s efforts, “Stalking the Wily Hacker.”

In March 2012, Google responded to an unusual power outage at one of its Belgian datacenters that ultimately led to local data corruption. Investigation revealed that a cat had damaged a nearby external power supply, triggering a series of cascading failures in the building’s power systems. By studying how complex systems fail in similar ways, Google has been able to adopt resilient design practices when suspending, burying, and submerging cables around the world.

Understanding a system’s adversaries is critical to building resilience and survivability for a wide variety of catastrophes. In the reliability context, adversaries usually operate with benign intent ...