Abstract: Embodiments herein selectively transmit probes between network nodes to estimate metrics associated with end-to-end paths through the network. In one embodiment, the network provider identifies key nodes based on a performance heuristic, such as the number of paths that run through each nodes, the amount of traffic that flows through the nodes, and the like. In addition to transmitting probes between the key nodes, supplemental probes can be transmitted from each customer edge device (e.g., a customer endpoint) to the key nodes in the provider's network. The data generated by probing between the key nodes and by probing between the customer edge devices and the key nodes can be combined to estimate metrics associated with end-to-end routes through the network.

Abstract: Techniques for simulation of orchestration and automation are disclosed. These techniques include simulating a distributed orchestration system using a simulation topology including a plurality of queues used in place of processing elements in the distributed orchestration system. The techniques further include identifying a plurality of characteristics of the simulation topology, after the simulation, and modifying the distributed orchestration system based on the plurality of characteristics.

Abstract: Techniques for automated analysis and classification of network configurations are provided. Resource provisioning code used to provision resources in a network is received, and a set of constraints is generated based on the resource provisioning code, where the set of constraints define valid network configurations. A configuration tree for the network is received, and a set of candidates is generated for the network based at least in part on the configuration tree. Upon determining that a first candidate of the set of candidates does not fail any constraint in the set of constraints, the first candidate is recorded as a manual intent. The manual intent is integrated into an automated configuration system for the network.

Abstract: Techniques for automated analysis and classification of network configurations are provided. Resource provisioning code used to provision resources in a network is received, and a set of constraints is generated based on the resource provisioning code, where the set of constraints define valid network configurations. A configuration tree for the network is received, and a set of candidates is generated for the network based at least in part on the configuration tree. Upon determining that a first candidate of the set of candidates does not fail any constraint in the set of constraints, the first candidate is recorded as a manual intent. The manual intent is integrated into an automated configuration system for the network.