Flexgrid technology is now considered to be a promising solution for future high-speed network design. In this context, we need a tutorial that covers the key aspects of elastic optical networks. This tutorial paper starts with a brief introduction of the elastic optical network and its unique characteristics. The paper then moves to the architecture of the elastic optical network and its operation principle. To complete the discussion of network architecture, this paper focuses on the different node architectures, and compares their performance in terms of scalability and flexibility. Thereafter, this paper reviews and classifies routing and spectrum allocation (RSA) approaches including their pros and cons. Furthermore, various aspects, namely, fragmentation, modulation, quality-of-transmission, traffic grooming, survivability, energy saving, and networking cost related to RSA, are presented. Finally, the paper explores the experimental demonstrations that have tested the functionality of the elastic optical network, and follows that with the research challenges and open issues posed by flexible networks.

Routing and Spectrum Allocation in Elastic Optical Networks: A Tutorial

Bandwidth fragmentation, a serious issue in elastic optical networks (EONs), can be suppressed by proper management in order to enhance the accommodated traffic demands. In this context, we need an in-depth survey that covers bandwidth fragmentation problems and how to suppress them. This survey paper starts with the basic concept of EONs and their unique properties. This paper then moves to the fragmentation problem in EONs. We discuss and analyze the major conventional spectrum allocation policies in terms of the fragmentation effect in a network. The taxonomies of the fragmentation management approaches are presented along with different node architectures. Subsequently, this paper reviews state-of-the-art fragmentation management approaches. Next, we evaluate and analyze the major fragmentation management approaches in terms of blocking probability. Finally, we address the research challenges and open issues on fragmentation problem in EONs that should be addressed in future research.

Fragmentation Problems and Management Approaches in Elastic Optical Networks: A Survey

This paper proposes DeepRMSA, a deep reinforcement learning framework for routing, modulation and spectrum assignment (RMSA) in elastic optical networks (EONs). DeepRMSA learns the correct online RMSA policies by parameterizing the policies with deep neural networks (DNNs) that can sense complex EON states. The DNNs are trained with experiences of dynamic lightpath provisioning. We first modify the asynchronous advantage actor-critic algorithm and present an episode-based training mechanism for DeepRMSA, namely, DeepRMSA-EP. DeepRMSA-EP divides the dynamic provisioning process into multiple episodes (each containing the servicing of a fixed number of lightpath requests) and performs training by the end of each episode. The optimization target of DeepRMSA-EP at each step of servicing a request is to maximize the cumulative reward within the rest of the episode. Thus, we obviate the need for estimating the rewards related to unknown future states. To overcome the instability issue in the training of DeepRMSA-EP due to the oscillations of cumulative rewards, we further propose a window-based flexible training mechanism, i.e., DeepRMSA-FLX. DeepRMSA-FLX attempts to smooth out the oscillations by defining the optimization scope at each step as a sliding window, and ensuring that the cumulative rewards always include rewards from a fixed number of requests. Evaluations with the two sample topologies show that DeepRMSA-FLX can effectively stabilize the training while achieving blocking probability reductions of more than 20.3% and 14.3%, when compared with the baselines.

DeepRMSA: A Deep Reinforcement Learning Framework for Routing, Modulation and Spectrum Assignment in Elastic Optical Networks

Elastic optical networks (EONs) enable network operators to have agile bandwidth management in the optical layer. In this paper, we investigate dynamic and adaptive bandwidth defragmentation (DF) in EONs with time-varying traffic using connection reconfigurations. We consider how to design DF procedure in a systematic way, and study the problems that have not been fully explored so far. Basically, we divide the procedure design into four subproblems: “How to reconfigure?,” “ How to migrate traffic?,” “When to reconfigure?,” and “What to reconfigure?,” and solve them sequentially. For “How to reconfigure?,” we investigate the combination of routing and spectrum assignment (RSA) algorithms for DF, i.e., the RSA algorithm that the connections are originally served with and the algorithm that they are re-optimized with. For “ How to migrate traffic?,” we propose to construct a dependency graph to represent the relations among the selected connections and to use it to assist the best-effort traffic migration. A move-to-vacancy method is also proposed to further reduce the traffic disruptions. For “When to reconfigure?” and “ What to reconfigure?,” we propose intelligent timing selection and adaptive DF ratio selection methods to tackle the tradeoff between the bandwidth blocking probability (BBP) performance and operational complexity. Simulation results show that the algorithm with both methods implemented (DF-AT-AR) achieves better tradeoff between BBP performance and operational complexity, when compared with existing algorithms.

Dynamic and Adaptive Bandwidth Defragmentation in Spectrum-Sliced Elastic Optical Networks With Time-Varying Traffic

Spectrum fragmentation limits the efficiency of spectrum utilization in elastic optical networks (EONs). This paper studies how to take advantage of the centralized network control and management provided by software-defined EONs (SD-EONs) for realizing OpenFlow-assisted implementation of online defragmentation (DF). We first discuss the overall system design and OpenFlow protocol extensions to support efficient online DF and conduct DF experiments with routing and spectrum assignment (RSA) reconfigurations in a single-domain SD-EON. Then, we propose to realize fragmentation-aware RSA (FA-RSA) in multi-domain SD-EONs with the cooperation of multiple OpenFlow controllers. In order to provision inter-domain lightpaths with restricted domain visibility on intradomain resource utilization, we design and implement an inter-domain protocol to facilitate FA-RSA in multi-domain SD-EONs and demonstrate controlling the spectrum fragmentation on inter-domain links with FA-RSA. Our experimental results indicate that the OpenFlow-controlled DF systems perform well and can improve the performance of SD-EONs effectively.

OpenFlow-assisted online defragmentation in single-/multi-domain software-defined elastic optical networks [invited]

This paper investigates the spectrum fragmentation issue, which undermines the bandwidth efficiency in elastic optical networks. After categorizing the two-dimensional fragmentation problem as the fragmentation and misalignment subproblems, this paper proposes joint routing and spectrum assignment (RSA) algorithms to alleviate the spectral fragmentation in the lightpath provisioning process. The time complexity of the two proposed algorithms are analyzed in detail, and both algorithms can run in 0(kdnC log C) time, where k is the number of the shortest path in the routing algorithm, d is the maximum node degree in the network, n is the number of nodes in the network, and C is the link capacity expressed as the number of spectral slots. Simulation results indicate that the proposed fragmentation-aware (FA) RSA algorithm and the FA algorithm with congestion avoidance (CA) outperform the existing schemes in terms of blocking probability (BP) reduction. Compared with the benchmark K-shortest-path routing and first-fit assignment (KSP-FF) algorithm, the proposed FA and FA-CA algorithms can achieve a BP reduction of [100%, 4.43%] and [100%, 6.45%], respectively, according to the traffic load in a sample NSFNET topology.

Huan Zhang

Papers

Spectral and spatial 2D fragmentation-aware routing and spectrum assignment algorithms in elastic optical networks [invited]