The student will design and implement a benchmark orchestration for
containerized workloads on Linux, Windows, and macOS, using the TOPSAIL
framework. The goal is not only to measure performance but also to provide a
systematic testing tool for real-world scenarios, suitable for integration into
CI pipelines to monitor the impact of PR changes on performance.

The orchestrator will:

• automate and coordinate various test scenarios for Podman and Docker,

• evaluate real-world workloads beyond synthetic benchmarks (including re-using/re-implementation of existing benchmarks),

• address challenges such as reproducibility, system variability, and cross-platform compatibility,

• provide structured methods to analyze performance trends across versions and configurations (including before/after comparisons).

The thesis will also include a survey and operationalization of the key metrics commonly used to evaluate containerized vs. VM-based workloads, such as:

• throughput and latency (including percentiles like p95/p99),

• CPU, memory, I/O, and storage utilization/overhead (IOPS, latency),

• stability, jitter, and determinism of repeated runs,

• orchestration- and isolation-related overhead,

The final submission will include: architecture design of the orchestrator, a functional prototype, a set of test scenarios, measurement methodology, and evaluation results with recommendations for practical adoption.