CRIME (Compression Ratio Info-leak Made Easy), disclosed in 2012, exploited TLS-layer compression — an optional feature that could shrink encrypted traffic — to infer secret data (like session cookies) by observing how the compressed size of a response changed based on guessed content.
The fix
The response was straightforward compared to some other TLS vulnerabilities: disable TLS-layer compression entirely, which browsers and servers did quickly following disclosure. This is also part of why TLS 1.3 removed compression from the protocol specification outright, rather than leaving it as an optional, occasionally-misused feature.
Why compression and encryption make an unusually bad combination
CRIME exploited a subtle but general property: when data is compressed before encryption, and an attacker can influence part of what gets compressed alongside a secret, the resulting compressed size can reveal information about the secret through how well the attacker's guess overlapped with it. This is a broader lesson in applied cryptography — compression and encryption interacting is a recurring source of information leaks, and it resurfaced later in a related attack called BREACH that applied similar logic to HTTP-level compression.
Why disabling compression was an unusually clean fix
Unlike some TLS vulnerabilities that require careful, narrowly scoped patches to avoid breaking legitimate functionality, TLS-layer compression turned out to provide only marginal real-world bandwidth benefit for most traffic, since most web content is already compressed at the HTTP layer or is inherently high-entropy. This made disabling it outright a low-cost, high-value fix rather than a difficult trade-off.
Who discovered CRIME and how it built on earlier research
CRIME was disclosed by security researchers Juliano Rizzo and Thai Duong, the same pair responsible for demonstrating BEAST the year before — their continued, focused research into TLS 1.0-era weaknesses across consecutive years is part of why this period is often treated as a particularly concentrated stretch of foundational TLS vulnerability research.
The related BREACH attack that emerged afterward
A related follow-up attack called BREACH, disclosed in 2013, applied a similar compression-based leak technique but targeted HTTP-level compression (like gzip) rather than TLS-level compression specifically — meaning disabling TLS compression alone, the fix for CRIME, didn't fully close every variant of this general class of attack, since HTTP-level compression required its own separate mitigation guidance.
The lasting rule of thumb it left behind
CRIME's core lesson, don't compress attacker-influenced data together with secrets before encrypting, has become a standing rule of thumb cited well beyond TLS itself, informing secure design guidance for other protocols and application-layer systems that combine compression and encryption in similar ways.
Why CRIME's relatively quick, clean resolution stands out in this history
Compared to several other vulnerabilities covered in this section that required years of gradual, staged industry response, CRIME's resolution, disable TLS compression outright, was unusually fast and complete precisely because the feature being removed provided so little real offsetting benefit. This made CRIME something of an outlier: a serious vulnerability whose fix required essentially no difficult trade-off discussion, which is part of why it doesn't dominate broader TLS history discussions the way some of its contemporaries do despite being conceptually significant.
Its enduring value as a teaching example
CRIME continues to appear in security training materials specifically because it's unusually easy to explain conceptually, compress secrets alongside attacker-controlled data and the size leaks information, while still representing genuine, subtle applied cryptography — a rare combination of accessibility and technical depth that makes it a favorite example for introducing side-channel attack concepts to newcomers.
CRIME's fast, clean resolution is the exception rather than the rule among the vulnerabilities covered throughout this history section.