Open source software security

The New School of Information Security

30 November -0001

As promised here is my full review of The New School of Information Security by Adam Shostack and Andrew Stewart:

The New School of Information Security is one of the most timely and radical books on computer and information security that I've ever read. Adam Shostack and Andrew Stewart help to stimulate a significant paradigm shift that has been brewing in the infosec sphere for some time. With solid evidence and well grounded arguments Shostack and Stewart advocate for a new, and much needed, approach to information security: the New School.

Chapter 1 begins with a quick look at some prominent problems in the information security landscape today. By looking at spam, malware, identity theft, and computer breaches the authors provide a rough sketch of the current infosec landscape. Given the apparent failure of current approaches to security in the face of these threats the authors rhetorically pose the question of simply starting over and building a new approach from scratch before providing the opening sketch of their New School. The authors advocate the need for a new approach to computer security, the New School. The New School is described as quantifiable, "putting our ideas and beliefs through tests designed to draw out their flaws and limitations." This concept of metrics and empiricism is a common thread throughout the book.

Chapter 2 describes the "scene," or the state of the computer security industry today. By applying some elementary game theory the authors sketch out some of the dilemmas facing information security today. Then they delve into some of the historic origins of modern computer security. They point out that much of the computer security "conventional wisdom" has grown out of the military's needs for computer security and how that foundation isn't necessarily the best. They also explore the influence of hackers and crackers on the evolution of the industry. Finally they explore the relationship of capitalism and money to the field, including the driving factors of making money and how these have shaped the development of security today. The authors point out that while many good things have come from these various influences, they have also produced some unfortunate side effects that don't necessarily have to be taken for granted. The chapter goes on to examine the economy of the security industry, including the idea of "best practices" (which the authors very roundly decry) as well as turnkey solutions. The authors also point out the difficulty in measuring security products given the lack of objective test data produced in the sector. The chapter concludes with the though that "without proper use of objective data to test our ideas, we can't tell if we are mistaken or misguided in our judgement." They provide further evidence that the industry as a whole isn't often guided by any sort of quantifiable data (thus removing the 'science' from computer science) and that all too often "conventional wisdom" is misguided and sometimes blatantly wrong because it lacks a solid empirical foundation.

Chapter 3 looks at some of the underpinnings of gathering solid scientific evidence with which to test the ideas of the New School. Without good evidence, they point out, it is nearly impossible to make accurate decisions. The authors point out the problems with much of the evidence used to support common claims in computer security, including surveys, and show the bias present in much of the survey data used to justify security decision making. The chapter goes on to lament the lack of an objective trade press in the industry and then delves into the vulnerability discovery lifecycle that drives much of computer security. The authors examine how vulnerabilities are discovered, how vendors often ignore flaws in their products in their rush to market, and the fact that there are sometimes problems with using vulnerability reports as solid metrics for security. The chapter then goes on to examine how data about security can be collected, either by hobbyists or individuals. Ultimately, the authors lament the fact that much of the data collected about security isn't shared with the community and thus it becomes nearly impossible to make better decisions. The lack of objective, available data makes it extremely difficult for us to draw reliable conclusions based on trends or quantify the current state of security.

Chapter 4 looks at security breaches and specifically argues for the benefits of breach notification as one of the best ways to produce quantifiable metrics in security. The authors point out that breach notification rarely has long term consequences to a companies stock price or customer loyalty and the benefit of breach data would be invaluable to researchers. The authors argue that breach notification is a key component to the outlook of the New School. In joining the New School organizations have to learn "to focus on observation and objective measurement." They argue that only by doing so can we move information security from an art to a science. They say that while "it is true that computer security consists of a fog of moving parts...complex problems do get solved. Investigators bring a broad set of analytic techniques ranging from explanatory complex economic models." At this point in the book the authors begin to introduce another key component of the New School, that is the need for integration of other fields of study into computer security. The authors argue that by utilizing approaches and theories developed in the fields of psychology, economics, sociology, and other academic areas our understanding of information security can be broadened and greatly enhanced. They always come back to ideas of empiricism, however, stating that "the core aspect of scientific research - the ability to gather objective data against which to test hypotheses - has been largely missing from information security." The authors emphasize that not only does data need to be collected, it must also be shared in order to aid in our understanding of the data.

Chapter 5 begins to draw upon outside fields of academia to enhance the New School. This chapter begins by introducing several economic models and explaining how they influence information security. While economic approaches to security are nothing new (risk mitigation, calculations of value and exposure equaling risk, etc.) the New School argues that "because computers are inevitably employed within a larger world, information security as a discipline must embrace lessons from a far wider field." The authors argue that economic models don't only have to be applied at a macro level to computer security, but can also be applied to more compartmentalized security problems (such as getting users to select good passwords). They also examine the success potential of certain security products based on economic analysis. The chapter goes on to discuss how lessons from psychology can be incorporated into our security decision making and to help us understand computer security more fully. Finally the chapter draws on lessons from sociology and shows how they too can inform our understanding of security.

Chapter 6 focuses on spending. The chapter is devoted to examining how organizations spend their money on information security and why. Like the earlier chapters, this one applies the New School approach to attempt to analyze spending habits and challenges many of the foundational logic that supports common security spending plans. The chapter draws on lessons from economics and psychology to examine the patterns of spending and suggests some ways in which we can improve our spending on security. Ultimately the authors argue that we understand the factors that should influence spending and focus our efforts on the most quantifiably effective expenditures of money.

Chapter 7, or Life in the New School, discusses many of the challenges facing the New School. These range from the lack of quality data to the dearth of a standardized security vocabulary. This chapter mainly points out the challenges that lie ahead and the many ways that a new approach can help overcome them.

Chapter 8 is a blanket call to join the New School along with instructions for how to begin. The authors argue that New School proponents should collect good data, analyze that data and seek new perspectives. They point out that the New School draws from a diverse body of academic knowledge and advocates synthesizing work from other academic area into the New School approach. Ultimately the New School challenges us to change how we think about information security. Not only should we question the "conventional wisdom" we take for granted, but we should also seek out new hypothesis and ways to test them in order to expand our understanding of computer security as a whole.

The book is an easy read and make quite an impression. Shostack and Stewart lead the charge towards a more empirical approach to computer security. The field has matured enough that we should begin treating it seriously, and in order to do so we need to be able to speak authoritatively about issues. The voodoo of conventional wisdom is no longer good enough when making recommendations as experts. We need to be able to point to solid evidence to justify security strategies and implementations. We also need to be able to look at quantifiable data when evaluating new products and tools. Ultimately I see the field moving in this direction and I give kudos to Shostack and Steward for issuing this clarion call to an industry that will hopefully take their message to heart.