Dynamic virtual optical networks supporting uncertain traffic demands [invited]

This paper focuses on the concept of dynamic virtual optical networks (VONs) as a key enabling technology to support the requirements of the future Internet. To address the unpredictable nature of services and their characteristics, dynamic and on-demand planning and replanning of VONs is required. This paper proposes the use of novel modeling approaches to specifically consider the uncertain nature of traffic requests. In addition, it studies mechanisms to evaluate and minimize the impact of this uncertainty on the availability and efficiency of the virtual resources, considering at the same time requirements such as service resilience and security. The effect of periodic VON replanning is investigated and the relevant trade-offs are identified. Our modeling results show that taking into consideration the dynamic and uncertain nature of services provides significant benefits regarding the ability of the infrastructure to support these services and the associated resource efficiency.