The transformation of innovative research ideas to production systems is highly dependent on the capability of performing realistic and... Show moreThe transformation of innovative research ideas to production systems is highly dependent on the capability of performing realistic and reproducible network experiments. Simulation testbeds o↵er scalability, reproducibility but lack fidelity due to model abstraction and simplification, while physical testbeds o↵er high fidelity but lack reproducibility and often technically challenging and economically infeasible to perform large-scale experiments. In this work, we present a hybrid testbed consisting of container-based network emulation and physical devices to advocate high fidelity and reproducible networking experiments. In particular, the testbed integrates network emulators (Mininet) [5], a distributed control environment (ONOS) [1], physical switches (Pica8) and end-hosts (Raspberry Pi and commodity servers). The testbed (1) o↵ers functional fidelity through unmodified code execution on an emulated network, (2) supports large-scale network experiments using lightweight OS-level virtualization techniques and capable of running across distributed physical machines, (3) provides the topology flexibility, and (4) enhances the repeatability and reproducibility of network experiments. We validate the fidelity of the hybrid testbed through extensive experiments under di↵erent network conditions (e.g., varying topology and traffic pattern), and compare the results with the benchmark data collected on physical devices. M.S. in Computer Science, May 2017 Show less
