The Control Plane of the PORTOLAN Internet Topology Measurement System, Based on Smartphone Crowdsourcing: Architecture, Protocols and Implementation

The Internet actually connects around two billion users through a global communication infrastructure, emerged from an early nucleus of academic and government networks and constantly evolving. An accurate description of its structure, at multiple levels of abstraction, is important for several purposes, such as designing routing protocols, detecting network failures, or planning Internet Service Providers (ISPs) business relationships. As ISPs operate as commercial entities, they are reluctant to publicly reveal their network structure. As a consequence, in the last 10-15 years a number of measurement systems aiming to discover the Internet structure and properties have been deployed. However, despite significant efforts, the Internet structure has not yet been fully discovered, as measurements were carried out following a top-down approach, from the core of the Internet down to its edges, not being able of detecting its peripheral structure. The PORTOLAN Internet Topology Measurement System suggests a new approach in measuring the Internet, based on smartphone crowdsourcing, which will allow the discovery of the Internet peripheral structure, still in large part invisible to current measurement systems: a new bottom-up and bottomto-bottom measurement paradigm. This thesis contributions are the design and the implementation of a scalable and efficient control plane architecture for the PORTOLAN Internet Topology Measurement System. The proposed architecture provides an interface for specifying measurement tasks and coordinates the smartphones and the tasks execution performed by the smartphones themselves. Moreover, a measurement campaign to validate and evaluate the implemented framework has been conducted, and led to the discovery of previously unknown links at the AS-level of abstraction.