This master thesis presents an approach to apply purely passive, privacy-preserving network monitoring to a performance monitoring problem. Internet service providers and network operators need to know, which services their customers are consuming on which quality level. In addition, the customer must be enabled to relate the experienced quality to the agreed service level parameters. The application of passive network monitoring for such tasks is useful: First, it can evaluate the real traffic of the customers, without injecting additional traffic to the network and second, measurement results are exactly based on the services the customers are accessing. On the other hand, passive network monitoring always includes the risk of compromising the customers’ right for privacy and therefore often raises privacy concerns from the public. Nevertheless, recent privacy-preserving monitoring frameworks allow overcoming the “privacy-versus-utility” dilemma by decoupling data capturing, filtering and data processing in separate steps, always applying the required level of privacy. This work evaluates delay measurements based on passive monitoring and presents different possibilities to estimate the connection quality to remote Internet locations. The mechanism exploits passive RTT measurement results gathered from inter-domain traffic to remote Internet hosts and aggregates them. They system only reports reduced but relevant information to the operator, while discarding unnecessary information for privacy reasons. The remote locations represent aggregations on different administrative or geographic precision levels, like BGP prefixes, Autonomous Systems or countries.