Analysis of evolution scenarios for end-to-end quality of services provisioning in the internet

One of the main objectives in the next generation networks design is to propose customers a service of end-to-end connectivity with defined quality characteristics on-demand. This objective is challenging: a typical communication session in the Internet involves multiple networks managed by different providers and all networks along a path of a given session should forward guaranteed QoS traffic according to specific quality requirements. This paper analyzes the real Internet topology consisting of more than 30,000 Autonomous Systems and investigates probability to establish an end-to-end path between end-users in several scenarios. Received numerical results demonstrate that high probability of end-to-end connectivity can be achieved if end-user hosts are connected to large networks (with high connectivity degree). In other cases, probability to get end-to-end QoS path between end-nodes is relatively low unless a very significant number of Autonomous Systems in the Internet is involved to the provisioning process. The service with guaranteed end-to-end quality is unlikely to be delivered between any two end-points in the current Internet unless strong cooperation incentives exist for all parties in a communication chain.