Test Code Generation for Telecom Software Systems using Two-Stage Generative Model

TL;DR

The paper tackles the challenge of testing complex telecom software across multi-vendor Open RAN environments by proposing a two-stage generative framework. It first creates privacy-preserving synthetic time-series inputs from field-trial data using time-series diffusion and generative models, then generates test scripts via a code-focused LLM conditioned on natural-language test descriptions. Experimental results on public and telecom-specific data show the synthetic TS data preserves distributional characteristics and that code-generation models, especially CodeLLama, can produce usable test scripts with favorable content metrics. The work demonstrates a promising end-to-end approach for automated telecom software testing, with potential impact on faster release cycles, reduced field-trial costs, and improved cross-vendor interoperability.

Abstract

In recent years, the evolution of Telecom towards achieving intelligent, autonomous, and open networks has led to an increasingly complex Telecom Software system, supporting various heterogeneous deployment scenarios, with multi-standard and multi-vendor support. As a result, it becomes a challenge for large-scale Telecom software companies to develop and test software for all deployment scenarios. To address these challenges, we propose a framework for Automated Test Generation for large-scale Telecom Software systems. We begin by generating Test Case Input data for test scenarios observed using a time-series Generative model trained on historical Telecom Network data during field trials. Additionally, the time-series Generative model helps in preserving the privacy of Telecom data. The generated time-series software performance data are then utilized with test descriptions written in natural language to generate Test Script using the Generative Large Language Model. Our comprehensive experiments on public datasets and Telecom datasets obtained from operational Telecom Networks demonstrate that the framework can effectively generate comprehensive test case data input and useful test code.

Figures (5)

• Figure 1: Prompt Encoding.
• Figure 2: Main Components of the Framework.
• Figure 3: SSSDS4 t-SNE plot on telecom data.
• Figure 4: timeGAN t-SNE plot on telecom data.
• Figure 5: Example of generated code using CodeLLama and Mistral on Telecom samples.