Automated Performance Testing Based on Active Deep Learning

TL;DR

This paper introduces ACTA, an active learning-based black-box performance testing method that efficiently generates tests to identify performance issues without needing extensive historical data.

Contribution

It presents a novel automated test generation approach using active learning and generative adversarial networks for black-box performance testing.

Findings

ACTA performs comparably to random testing and other machine learning methods.

It effectively generates tests addressing specified performance conditions.

The method reduces the need for large historical test datasets.

Abstract

Generating tests that can reveal performance issues in large and complex software systems within a reasonable amount of time is a challenging task. On one hand, there are numerous combinations of input data values to explore. On the other hand, we have a limited test budget to execute tests. What makes this task even more difficult is the lack of access to source code and the internal details of these systems. In this paper, we present an automated test generation method called ACTA for black-box performance testing. ACTA is based on active learning, which means that it does not require a large set of historical test data to learn about the performance characteristics of the system under test. Instead, it dynamically chooses the tests to execute using uncertainty sampling. ACTA relies on a conditional variant of generative adversarial networks,and facilitates specifying performance requirements in terms of conditions and generating tests that address those conditions.We have evaluated ACTA on a benchmark web application, and the experimental results indicate that this method is comparable with random testing, and two other machine learning methods,i.e. PerfXRL and DN.