Fragmentation-Aware Shared Path Protection Algorithm In Elastic Optical Networks

Elastic Optical Network(EON)adopts Orthogonal Frequency Division Multiplexing(OFDM)technology,which has many advantages such as flexible spectrum assignment and supporting multi-rate transmission.It is regarded as the direction of flexible and intelligent development for optical networks.Routing and Spectrum Assignment(RSA)is one of the key issues in EON.Due to the dynamic establishment and release of lightpaths,the idle spectrum resources on the link are discretized and spectrum fragmentation occurs.Therefore,it is the key issue for RSA in EON to reduce fragmentation and to improve spectrum utilization.The fiber links in EON carry a large number of service connections.Once a failure occurs,it will cause unpredictable losses.Therefore,it is particularly important to consider the survivability of the RSA.Shared Path Protection(SPP)allows protection resources to be shared and can reduce spectrum resource occupancy,which has attracted great attention from academia and industry.This paper investigates the RSA problem based on SPP in EON.Combining fragmentation factors,two effective algorithms are proposed,which are evaluated by simulation.Firstly,we introduce the background and development status of EON,analyze the RSA problem and survivability problem,mainly focusing on fragmentation-aware RSA algorithm and survivability RSA algorithm,and classify and summarize the main ideas of some typical RSA algorithms.Secondly,since the existing RSA algorithms based on SPP often only consider how to improve the utilization of shared resources,but ignore the problems of idle fragmentation and protection fragmentation,we propose an Adaptive Modulation and Fragmentation-Aware Shared Path Protection(AMFA-SPP)algorithm.The proposed algorithm uses the spectrum block bearing weight(SBBW)to measure the ability of connection establishment for different spectrum blocks,and prefers the paths including links with larger SBBW as candidate working paths.The working path is established by choosing the spectrum block with the minimum difference of SBBW before and after the assignment,which is beneficial to reduce idle fragmentation of spectrum.Then,the proposed algorithm selects the paths containing links with larger weight of both available spectrum block and reserved protection spectrum block as candidate protection paths.After that,the proposed algorithm prefers to use the reserved protection spectrum block with the smallest SBBW difference before and after the assignment to establish the protection path.When the establishment fails,the available spectrum blocks with the minimum difference of SBBW before and after the assignment are selected to establish the protection path.The proposal is helpful to improve the shared resource utilization while reducing idle and protection fragmentation.Thirdly,since the existing SPP algorithms that considers the probability of link failure have high spare capacity redundancy,and the probability of reliability calculation model is inaccurate,we propose a Fragmentation-aware Shared Path Protection with Differentiated Relaibility(FSPP-DR)algorithm.Based on the joint failure probability of service connection,we present a calculation model for reliability probability of SPP.On the basis of meeting service reliability requirements,the proposal first try to establish service connections by a single-path,and to establish service connections by SPP only when the establishment fails,which is helpful to reduce space capacity redundancy.In the stage of route selection,links with a small probability of failure and a larger SBBW are preferentially selected to increase the successful probability for path establishment.In the stage of spectrum assignment,the proposal preferentially selects the specturm block with minimum difference of SBBW before and after the assignment.Our algorithm can improve the utilization of shared resources and reduce idle fragmentation and protection fragmentation.Finally,the EON simulation platform is built by Visual C++to simulate and analyze the two proposed algorithms under the dynamic traffic model and to compare them with the existing algorithms in the literatures.The effectiveness of the algorithms is verified.