Search

The convergence between IT and optical network services is a fundamental step to support new emerging applications, typically distributed and with strict requirements in terms of performance and service reliability. The FP7 GEYSERS project has designed a new network architecture capable of joint and on-demand provisioning of ‘Optical Network + Any-IT’ resources for end-to-end service delivery. This service, referred to as Network + IT Provisioning Service (NIPS), is enabled through the cross-layer cooperation between the Service Middleware Layer (SML) and an enhanced Network Control Plane (NCP+) along the overall service lifecycle. This paper describes the GEYSERS NCP+ architecture that extends standard ASON/GMPLS and PCE solutions to operate over a virtual optical network infrastructure and offer on-demand provisioning of enhanced network transport services between virtual IT resources connected to the network edges. In the proposed architecture, some of the functionalities traditionally provided by the SML, like the selection of the IT end-points, are moved to the NCP+ side, allowing global optimization of the utilization of network and IT resources and of energy consumption. Simulation results are provided to evaluate the benefits of the combined computation of IT end-points and network paths, both in terms of reduced service blocking probability and utilization of the infrastructure resources. [ABSTRACT FROM PUBLISHER]

High-capacity networks based on optical technologies enable global delivery of high-performance network-based applications driven by Future Internet services. Due to the accelerated evolution of these applications, dynamic adaptability of the underlying optical infrastructure becomes crucial to support efficient data transport through the operators' networks. Optical network virtualization can be considered as a key technology for addressing this challenge. This paper presents a novel architecture enabled by optical network virtualization that adopts the concept of Infrastructure as a Service (IaaS). For the IaaS framework, three intelligent and dynamic composition mechanisms, employing Mixed Integer Linear Programming (MILP), K-Shortest Path (K-SP) and Random path routing algorithms are proposed. The performance of these algorithms in terms of the virtual optical network (VON) composition performance and the algorithm computational complexity is evaluated and compared in this study. The results show that MILP can provide optimal solutions for VON composition but it takes the longest time to execute, while the proposed Random algorithm can achieve the shortest running time. [ABSTRACT FROM PUBLISHER]