ADI Leak Detection
A system for ADI Leak Detection memory leaks from software applications using data collected from memory usage of the software application over a period of time. A model based on the memory usage data is generated to determine a trend in memory usage by the software application. The trend is used to determine whether the software application has a memory leak (processing block 308).
The memory leak detection method includes a leak testing process that performs a series of adaptive tests on the product under test to determine whether the product has a gross leak, an insignificant leak, or a marginal leak during each of the adaptive tests. The product under test is classified as being acceptable if the measured leakage flowrate during each adaptive test is less than the predicted leakage flowrate.
The Science Behind ADI Leak Detection Technology
During a test period, leak detection sensor 18 generates a signal indicative of an actual leakage flowrate to leak detection processor 16. The actual leakage flowrate is compared with the calculated adaptive leakage flowrate at the end of the test period to determine the difference between the actual leakage flowrate and the predicted leakage flowrate. The difference is referred to as the “Leak Diff.”
If the difference is greater than Lmin, then the product is classified as having a gross leak. Otherwise, the product is classified as having an insignificant leak.
The leak detection method is configured to automatically detect memory leaks by generating a model from memory usage data collected from the software application over a period of time. The model is based on a linear regression analysis of the memory usage data. The resulting function is then used by leak determination engine 206 to determine whether the software application has a leak.