Abstract
Immediate reservation (IR) and advanced reservation (AR) become two common traffic models for future space division multiplexing-elastic optical networks (SDM-EONs). However, it is crucial to reduce conflicts between different types of requests and to accommodate more requests with limited resources. In this paper, we propose a resource allocation scheme based on complete planning process (RA-CPP) for IR and AR in SDM-EONs, including request provisioning, resource planning, passive adjustment, and active adjustment. In the request provisioning, the starting time of requests is considered and the earlier transmitted requests have higher resource selection priority. Next, the spectrum blocks with minimal resource spacing between adjacent requests (RS-AR) are planned for requests to maximize the free resources of the network in the resource planning. In addition, for the new request that would be blocked, passive adjustment begins. After planning reserved resources for the new request, conflicting AR requests are re-provisioned. Finally, the active adjustment re-optimizes the resources by increasing the tightness of spectrum resources. Simulation results show that the proposed RA-CPP scheme has greater advantages in blocking probability and spectrum utilization compared with the benchmark schemes.
Highlights
With the rapid development of 5G mobile communications, massive customized high-traffic services proliferate, resulting in the unprecedented growth of network traffic [1]
To improve the success rate of Immediate reservation (IR) and advance reservation (AR) requests, this paper proposes a resource allocation scheme based on complete planning process (RA-CPP) in space division multiplexing (SDM)-elastic optical networks (EONs)
NETWORK MODEL We model the SDM-EONs as a graph G(V, E, C), where V is the set of nodes, E represents the set of bi-direction links between nodes, and C is the set of cores on each fiber link
Summary
With the rapid development of 5G mobile communications, massive customized high-traffic services proliferate, resulting in the unprecedented growth of network traffic [1]. Resources are provisioned for requests when they arrive at the network, and allocated at the starting time. In the request provisioning process, some studies have used a delayed allocation approach [11,12,13] for AR requests, in which resources are not provisioned immediately as soon as the requests arrive, but before the starting time This approach undoubtedly increases the flexibility of RCSA by exploiting the temporal nature of AR requests. The provisioned resources are directly allocated to services at starting time These spectrum blocks (SBs) planned in advance may fail to achieve resource optimization due to continuous AR reprovisioning.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
