Timing attack

As explained in Hacking Cryptography

A timing attack is a side-channel attack that infers secret information from differences in the time required to perform computations.

A timing attack infers secret information from differences in computation time. An implementation can therefore leak data even when its outputs appear correct: how long an operation takes becomes an observable side channel. section 10.4

A classic failure is an ordinary comparison that stops at the first unequal byte. Matching prefixes take longer, allowing repeated measurements to reveal information about the secret. Constant-time comparison instead accumulates XOR differences across every byte and tests the final accumulator. section 10.4.1

Comparison and control-flow leaks

A constant-time comparison can return zero immediately when lengths differ, then OR each x[i] ^ y[i] into an accumulator before testing whether it is zero. The same caution applies to square-and-multiply algorithms, branching, and indexing based on secrets. Modern speculative execution can make secret-dependent control flow observable without an oscilloscope. section 10.4.1 section 10.5

Reduce timing leakage with direct engineering rules:

• Use specialized constant-time comparison functions such as OpenSSL CRYPTO_memcmp, Python hmac.compare_digest, Node.js crypto.timingSafeEqual, Java MessageDigest.isEqual, or Rust's subtle crate. section 10.4.1
• Avoid branching and iterating on sensitive values; prefer constant-time implementations for cryptographic code. section 10.5 section 10.9

Figure 10.4 Looking at runner C’s finish time of 2 seconds, we can roughly guess that they fell around the second hurdle.

Use the hurdle-race analogy to see how a finish time can reveal where an unseen delay occurred. section 10.4.1

How Hacking Cryptography covers it

How deep it goes

Explains early-exit comparisons, XOR accumulation, secret-dependent branching, and speculative execution, then stops short of a complete attack implementation.

Examples

Uses memcmp, constant-time byte comparison, cryptographic library APIs, and branch-prediction benchmarks to expose timing differences.

What it assumes

Assumes familiarity with comparisons and control flow while grounding the security risk in executable implementation patterns.

What you can do after

Diagnose timing leakage and apply constant-time comparison APIs while avoiding secret-dependent branches and indexing.

How it connects

Connects timing attacks with side-channel analysis, constant-time programming, branch prediction, speculative execution, and cryptographic implementation.

liveBookmanning.com

Read Hacking Cryptography on liveBook Get the book at manning.com

Related concepts

• constant-time comparisonaddressed by · A constant-time comparison is an equality-checking technique whose execution pattern does not reveal information through the position or content of mismatches.
• side-channel attackkind of · A side-channel attack infers secret information from unintended physical or operational leakage produced by a system’s implementation rather than its intended outputs.
• constant-time programmingaddressed by · Constant-time programming is the practice of designing algorithms whose execution behavior does not depend on secret data or varying input values.
• branching on secret valuesuses · Branching on secret values is control flow whose selected path depends on private or sensitive data.
• speculative executionuses · Speculative execution is a processor technique that performs predicted work before its need or correctness is confirmed, discarding results if the prediction is wrong.
• meanmeasures · A mean is a measure of central tendency, calculated as the arithmetic average of values or, for a random variable, as the probability-weighted average of its possible values.

This explainer was generated by Manning's knowledge graph from the text of Hacking Cryptography. The section links point to the parts of the book it draws on.