Low-latency network architectures such as Cloud RAN have put forward higher requirements on the performance and reliability of remote management of servers in data centers, where millisecond-level response times are needed. As the core component for implementing out-of-band management of servers, the baseboard management controller (BMC) plays a key role in ensuring system availability. However, while the Redfish protocol offers a modern standard for server management, there is still a lack of a standard benchmark framework for systematic evaluation because there are significant differences in the implementation methods and performance of BMCs across manufacturers.
This thesis proposes a modular BMC benchmark framework based on the Redfish protocol, covering both key performance metrics (such as API response time and firmware update duration) and reliability checking cases (such as system recovery and robustness under stress). The framework is developed in Python, supports compatibility testing of BMCs from multiple manufacturers, and it can automatically generate test reports in structured JSON format. In the evaluation of two BMCs, the framework shows good reusability and comparison capabilities, revealing the trade-offs between different products in terms of performance and stability, and provides support for the selection of BMC solutions in data centers.
