Resource Allocation Policies in Disaster Situations
Elastic Optical Network, Quality of Service, Service Degradation, Resource Scarcity, Disaster
Elastic optical networks (EON) allow the transmission of large volumes of data through multiple channels with different spectrum granularities. However, the resource availability of EON networks can be severely impacted by congestion, natural disasters or human- made attacks. To deal with these resource constraints, the network operator needs to make choices about which lightpath will be served by the network. The Decision-making process must take into account requirements of the different Classes of Service (CoS), as well as the possibility of service degradation, providing lower bandwidth than reques- ted or adjusting the time instant of establishing the ligthpath, to adapt the network provisioning when optical resources are insufficient. In this doctoral thesis, problems related to the scarcity of resources and survival in Elastic Optical Networks are addres- sed. Strategies that were at least as efficient as existing strategies in the literature were developed and validated: an algorithm that considers a proportional Quality of Service (QoS) model and information from higher layers to decide which lightpath to be degra- ded in provisioning process, aiming to reduce the impact the unavailability of resources in applications sensitive to delays and bandwidth; an algorithm that uses a hybrid multi- criteria decision-making technique to select requests to be provisioned; an algorithm for selecting lightpaths to be restored after disasters; an algorithm that aims to reduce spec- trum fragmentation in Elastic Optical Networks by allocating new requests on paths that produce less spectrum fragmentation after allocation. Some of the proposed algorithms have a multi-criteria decision approach that considers CoS, bandwidth, number of hops and completion time. Furthermore, service degradation is also considered for lightpaths that cannot be restored to full bandwidth.