Timing-sensitive analysis in python

Time consistency is critical in many fields, especially in sensitive applications like cryptography. The FaCT DSL is used to define functions with predictable execution times to prevent timing leaks. In this notebook, functions have been created to simulate varying computational complexity (e.g., sorting, searching).

In the code snippet below the timing-sensitive sorting function is set and measures how long it takes to execute, with an optional delay.

def timing_sensitive_sort(size, delay=False): arr = np.random.rand(size) if delay: time.sleep(0.1) # Simulated delay to mimic a timing inconsistency return np.sort(arr) # Measure execution time with and without delay def measure_timing_sensitive_sort(size, delay=False): start = time.time() timing_sensitive_sort(size, delay) return time.time() - start

Run to view results

This code creates a function to sort a random array, with an optional delay to simulate timing inconsistencies, and measures how long it takes to execute.

Size

# Define function to collect results with dynamic input sizes def timing_experiment(sizes, delays=[False, True]): results = [] for size in sizes: for delay in delays: times = [measure_timing_sensitive_sort(size, delay) for _ in range(10)] results.append((size, delay, np.mean(times), np.std(times))) return results # Interactive input for sizes range size_range = np.arange(100, 2000, 200) # Default range, but replaceable results = timing_experiment(size_range)

Run to view results

The plot shows the consistency of execution times across different input sizes without any intentional delay, allowing to observe how input size affects timing stability.

plt.figure(figsize=(12, 6)) sns.boxplot(x=[result[0] for result in results if not result[1]], y=[result[2] for result in results if not result[1]]) plt.xlabel('Input size') plt.ylabel('Execution time (s)') plt.title('Execution time consistency without delays') plt.show()

Run to view results

.css-15w88e5{color:var(--chakra-colors-fg-neutral-primary);font-weight:inherit;letter-spacing:-0.09px;}Timing-sensitive analysis in python

Timing-sensitive analysis in python