Local Allocation of End-to-End Quality-of-Service in High-Speed Networks