A Traffic Engineering Algorithm for Provisioning Virtual Private Networks in the Enhanced Hose Model

A Virtual Private Network is a logical network established on top of a public packet switched network. To guarantee that quality of service requirements, specified by customers, can be met, the network service provider needs to reserve enough resources on the network and allocate/manage them in an optimal way. Traffic engineering algorithms can be used by the Network Service Provider to establish multiple Virtual Private Networks in an optimal way, while meeting customers' Quality of Service requirements. For delay sensitive network applications, it is critical to meet both bandwidth and delay requirements. In contrast to traditional Virtual Private Network Quality of Service models (customer-pipe model and hose model), which focused only on bandwidth requirements, a new model called the enhanced hose model has been proposed, which considers both bandwidth and delay requirements. However, to the best of our knowledge, thus far, traffic engineering problems associated with establishing multiple enhanced hose model Virtual Private Networks have not been investigated. In this paper, we proposed a novel Virtual Private Network traffic engineering algorithm, called the minimum bandwidth-delay cost tree algorithm to address these problems. According to experimental simulations conducted and reported in our paper, the minimum bandwidth-delay cost tree algorithm can indeed achieved better performance (lower rejection ratios) compared to previous algorithms.