Optimizing Network Automation with pyATS: Multi-Threaded Execution in MPLS-L3VPN Environments

As modern network infrastructures grow in scale and complexity, the limitations of manual configuration models have become increasingly evident, needing a shift towards automation to ensure operational coherence and scalability. This research and proof-of-concept scenario investigates the effectiveness of Cisco's Python Automated Test Systems, also known as pyATS, specifically its Genie library, designed to automate MPLS Layer 3 VPN (L3VPN) infrastructures. A Python-centric Infrastructure-as-Code (IaC) approach was used using Jinja2 templates to reuse device-specific configurations. The proposed architecture was deployed in an emulated GNS3 environment with a simplified core MPLS, and its performance was benchmarked against Netmiko and Ansible, two widely adopted IT automation tools. Key metrics including per-core CPU utilization, RAM consumption, and configuration deployment time were monitored using multithreaded scripts executed from a centralized SDN controller. Results indicate that Netmiko minimizes CPU and memory usage but shows high variability; on the other hand, Ansible achieves the fastest and most consistent deployments at cost-effective with higher CPU loads, while pyATS offers a balanced tradeoff between resource efficiency and operational stability. The findings underscore the importance of aligning the selection of automation IT tools with operational priorities and highlight the practical value of pyATS in structured network validation within the NetDevOps and IaC paradigms.