Learning-based congestion controllers offer better adaptability compared to traditional heuristics. However, the unreliability of learning techniques can cause learning-based controllers to behave poorly, creating a need for formal guarantees. While methods for formally verifying learned congestion controllers exist, these methods offer binary feedback that cannot optimize the controller toward better behavior. We improve this state-of-the-art via Canopy, a new property-driven framework that integrates learning with formal reasoning in the learning loop. Canopy uses novel quantitative certification with an abstract interpreter to guide the training process, rewarding models, and evaluating robust and safe model performance on worst-case inputs. Our evaluation demonstrates that unlike state-of-the-art learned controllers, Canopy-trained controllers provide both adaptability and worst-case reliability across a range of network conditions.

Software-defined networks (SDNs) only logically centralize the control plane. In reality, SDN controllers are distributed entities, which may exhibit different behavior on event triggers. We identify several classes of faults that afflict an SDN controller cluster and demonstrate them on two enterprise SDN controllers, ONOS and OpenDaylight. We present JURY, a system to validate controller activities in a clustered SDN deployment, involving topological and forwarding state, without imposing any restrictions on the controller behavior. Our evaluation shows that JURY requires minimal changes to the SDN controllers for deployment, and is capable of validating controller actions in near real time with low performance overheads.

With majority of the world's data and computation handled by cloud-based systems, cloud management stacks such as Apache's CloudStack, VMware's vSphere and OpenStack have become an increasingly important component in cloud software. However, like every other complex distributed system, these cloud stacks are susceptible to faults, whose root cause is often hard to diagnose. We present HANSEL, a system that leverages non-intrusive network monitoring to expedite root cause analysis of such faults manifesting in OpenStack operations. HANSEL is fast and accurate, and precise even under conditions of stress.