MIT researchers have developed Metior, a groundbreaking framework that enables quantitative evaluation of obfuscation schemes used to safeguard against side-channel attacks. By converting information flow into mathematical variables, Metior empowers engineers to gain deeper insights into the performance of security measures. This knowledge aids in the selection of effective strategies during the chip design process.
Metior, MIT’s new system, assesses the likelihood of an attacker compromising a specific security scheme to pilfer confidential information.
Researchers have devised a system to scrutinize how hackers can circumvent particular cybersecurity methods, with the goal of determining the extent to which an attacker can extract secret information from a computer program.
Observing a computer program’s behavior, such as its memory access patterns, clever hackers can acquire sensitive data like passwords. Completely blocking such “side-channel attacks” is often computationally expensive and impractical for real-world systems. Instead, engineers employ obfuscation schemes to restrict, though not eliminate, attackers’ ability to gather secret information.
To facilitate a better understanding of obfuscation schemes’ effectiveness, MIT researchers created Metior, a framework that quantitatively evaluates the amount of information an attacker can extract from a victim program protected by an obfuscation scheme.
Metior, known as a comprehensive model, enables users to study the impact of different victim programs, attacker strategies, and obfuscation scheme configurations on the leakage of sensitive information. Engineers involved in microprocessor development can leverage this framework to evaluate the effectiveness of various security schemes and identify the most promising architecture at the early stages of chip design.
“Metior helps us realize that we should not view these security schemes in isolation. Analyzing the effectiveness of an obfuscation scheme for a single victim may be tempting, but it fails to provide a comprehensive understanding of why these attacks work. Taking a higher-level perspective offers a holistic picture of the underlying dynamics,” explains Peter Deutsch, a graduate student and lead author of an open-access paper on Metior.
The team of authors includes Peter Deutsch, Weon Taek Na, Thomas Bourgeat, Joel Emer, and Mengjia Yan. Yan, the senior author, is an assistant professor at MIT and a member of the Computer Science and Artificial Intelligence Laboratory (CSAIL). The research was recently presented at the International Symposium on Computer Architecture.
Unveiling the effectiveness of obfuscation
While numerous obfuscation schemes exist, popular approaches typically introduce randomization into a victim’s behavior to make it more challenging for attackers to extract secrets. For example, an obfuscation scheme might involve a program accessing additional areas of computer memory, confusing attackers. Other schemes adjust the frequency of memory or shared resource access, making it difficult for attackers to detect clear patterns.
Despite making it harder for attackers to succeed, these approaches still result in some leakage of information from the victim. Hence, Yan and her team sought to determine the extent of such leakage.
Previously, they developed CaSA, a tool to quantify information leakage from a specific type of obfuscation scheme. However, with Metior, they set more ambitious goals: creating a unified model capable of analyzing any obfuscation scheme, including those not yet developed.
To achieve this, Metior was designed to map the flow of information through an obfuscation scheme into random variables. For instance, the model converts the shared structure access by a victim and an attacker, such as memory, into a mathematical representation.
Once Metior establishes the mathematical framework, it employs techniques from information theory to ascertain the extent to which an attacker can learn information from the victim. With these pieces in place, Metior quantifies the attacker’s likelihood of successfully guessing the victim’s secret information.
“We transform all the intricate elements of this microarchitectural side-channel into a mathematical problem. This enables us to explore various strategies and gain a better understanding of how minor adjustments can enhance defenses against information leakage,” elaborates Deutsch.
Revealing surprising insights
The researchers conducted three case studies using Metior to compare attack strategies and analyze information leakage from state-of-the-art obfuscation schemes. Through their evaluations, Metior identified interesting behaviors that were previously poorly understood.
For instance, prior analysis indicated that a sophisticated side-channel attack known as probabilistic prime and probe was successful due to its preliminary profiling of a victim system to understand its defenses.
Using Metior, the researchers demonstrated that this advanced attack performs no better than a simple, generic attack. Additionally, it exploits different victim behaviors than previously assumed.
Moving forward, the researchers aim to further enhance Metior to enable more efficient analysis of complex obfuscation schemes. They also plan to investigate additional obfuscation schemes and victim program types while conducting detailed analyses of the most popular defense mechanisms.
Ultimately, the researchers envision that their work will inspire the development of microarchitectural security evaluation methodologies that can be applied early in the chip design process.
“Microprocessor development is an incredibly costly and complex endeavor, and design resources are scarce. Having an effective method to evaluate the value of a security feature before committing to microprocessor development is crucial. Metior offers a general solution for achieving this,” emphasizes Joel Emer, an MIT professor of the practice in computer science and electrical engineering.
Reference: Metior: A Comprehensive Model to Evaluate Obfuscating Side-Channel Defense Schemes
This research receives funding from the National Science Foundation, the Air Force Office of Scientific Research, Intel, and the MIT RSC Research Fund.
Table of Contents
Frequently Asked Questions (FAQs) about side-channel attacks
What is Metior?
Metior is a framework developed by MIT researchers for the quantitative evaluation of obfuscation schemes used in chip design to protect against side-channel attacks. It allows engineers to better understand the effectiveness of security measures and aids in the selection of strategies.
What are side-channel attacks?
Side-channel attacks are techniques used by hackers to obtain secret information by observing a computer program’s behavior, such as memory access patterns. These attacks aim to extract sensitive data, such as passwords, by exploiting subtle information leaks rather than directly attacking the encryption or authentication mechanisms.
How does Metior help in chip design?
Metior provides a comprehensive model to evaluate the effectiveness of different obfuscation schemes in limiting information leakage from victim programs. It allows engineers to study the impact of various victim programs, attacker strategies, and obfuscation scheme configurations, aiding in the development of more secure microprocessors.
Why use obfuscation schemes instead of blocking side-channel attacks completely?
Blocking side-channel attacks completely can be computationally expensive and impractical for real-world systems. Obfuscation schemes seek to limit the attacker’s ability to gather secret information by introducing randomization or altering access patterns, making it harder for them to decipher the information. These schemes provide a practical compromise between security and computational feasibility.
What are the future plans for Metior?
The researchers aim to enhance Metior to analyze more complex obfuscation schemes efficiently. They also plan to investigate additional obfuscation schemes, victim program types, and popular defense mechanisms. The long-term goal is to inspire the development of microarchitectural security evaluation methodologies for early assessment in the chip design process.
More about side-channel attacks
- MIT News: “MIT’s Cybersecurity Metior: A Secret Weapon Against Side-Channel Attacks”
- Metior: A Comprehensive Model to Evaluate Obfuscating Side-Channel Defense Schemes
- International Symposium on Computer Architecture (ISCA)
- MIT Computer Science and Artificial Intelligence Laboratory (CSAIL)
- MIT Electrical Engineering and Computer Science (EECS)
- MIT Research Support Committee (RSC) Research Fund
- National Science Foundation (NSF)
- Air Force Office of Scientific Research
- Intel Corporation
