Preventing Piracy, Reverse Engineering, and Tampering

P ersonal privacy, national security, and other fundamental values hinge on the ability to protect data from unauthorized access. As computing becomes pervasive, concerns about data protection have taken on new urgency. For example, obtaining unauthorized access to someone's medical history once would have required physically breaking into a doc-tor's office, searching through one or more filing cabinets , and extracting the patient's folder. Today, it is often possible to obtain such records by breaking into the doctor's computer from a remote location. What makes securing digital data difficult is that it is rarely static—rather, data is manipulated by software, often in a networked environment. For example, parts of a hospital's medical database may be accessible to queries by the police department. In many cases, to steal or manipulate data an attacker need not take over the host computer but instead only has to defeat the software programs responsible for protecting the data. Software itself is a form of data and as such is vulnerable to theft and misuse. Given the enormous investment of time, money, and intellectual capital in software development, piracy has long been— and continues to be—a major threat to the software industry. In its most recent study, which tracked 26 popular business applications in 85 countries, the Business Software Alliance (www.bsa.org) reported that the global economic impact of pirated software totaled nearly US$11 billion in 2001. 1 The problem, however, extends well beyond that of software piracy. Software is increasingly being distributed as mobile code in architecture-independent formats. Most such formats are essentially equivalent to source code, which makes the software susceptible to decompilation and reverse engineering. Malicious parties can steal the intellectual property associated with such code with relative ease. Clearly, there is a strong need for developing more efficient and effective mechanisms to protect software. Unfortunately, none of the major approaches currently used by software developers and vendors provide adequate protection, especially on today's open computing platforms. However, three promising techniques under development— tamper proofing, obfuscation, and watermarking— offer hope for the future. As Figure 1 shows, from a software vendor's point of view, a program executes in either a trusted or untrusted environment. We use the term " program " for simplicity—software protection mechanisms can and should be applied not only to complete programs, but also to program components, such as reusable class libraries. The typical end user's computer is inherently untrustworthy because …