Empowered by the optical orthogonal frequency-division multiplexing (O-OFDM) technology, flexible online service provisioning can be realized with dynamic routing, modulation, and spectrum assignment (RMSA). In this paper, we propose several online service provisioning algorithms that incorporate dynamic RMSA with a hybrid single-/multi-path routing (HSMR) scheme. We investigate two types of HSMR schemes, namely HSMR using online path computation (HSMR-OPC) and HSMR using fixed path sets (HSMR-FPS). Moreover, for HSMR-FPS, we analyze several path selection policies to optimize the design. We evaluate the proposed algorithms with numerical simulations using a Poisson traffic model and two mesh network topologies. The simulation results have demonstrated that the proposed HSMR schemes can effectively reduce the bandwidth blocking probability (BBP) of dynamic RMSA, as compared to two benchmark algorithms that use single-path routing and split spectrum. Our simulation results suggest that HSMR-OPC can achieve the lowest BBP among all HSMR schemes. This is attributed to the fact that HSMR-OPC optimizes routing paths for each request on the fly with considerations of both bandwidth utilizations and lengths of links. Our simulation results also indicate that the HSMR-FPS scheme that use the largest slots-over-square-of-hops first path-selection policy obtains the lowest BBP among all HSMR-FPS schemes. We then investigate the proposed algorithms' impacts on other network performance metrics, including network throughput and network bandwidth fragmentation ratio. To the best of our knowledge, this is the first attempt to consider dynamic RMSA based on both online path computation and offline path computation with various path selection policies for multipath provisioning in O-OFDM networks.

This article discusses the technologies for realizing highly efficient data migration and backup for big data applications in elastic optical inter-data-center (inter-DC) networks. We first describe the impacts of big data applications on underlying network infrastructure and introduce the concept of flexible-grid elastic optical inter-DC networks. Then we model the data migration in such networks as dynamic anycast and propose several efficient algorithms. Joint resource defragmentation is also discussed to further improve network performance. For efficient data backup, we leverage a mutual backup model and investigate how to avoid the prolonged negative impacts on DCs' normal operation by minimizing the DC backup window.

Based on the concept of infrastructure as a service, optical network virtualization can facilitate the sharing of physical infrastructure among different users and applications. In this paper, we design algorithms for both transparent and opaque virtual optical network embedding (VONE) over flexible-grid elastic optical networks. For transparent VONE, we first formulate an integer linear programming (ILP) model that leverages the all-or-nothing multi-commodity flow in graphs. Then, to consider the continuity and consecutiveness of substrate fiber links' (SFLs') optical spectra, we propose a layered-auxiliary-graph (LAG) approach that decomposes the physical infrastructure into several layered graphs according to the bandwidth requirement of a virtual optical network request. With LAG, we design two heuristic algorithms: one applies LAG to achieve integrated routing and spectrum assignment in link mapping (i.e., local resource capacity (LRC)-layered shortest-path routing LaSP), while the other realizes coordinated node and link mapping using LAG (i.e., layered local resource capacity(LaLRC)-LaSP). The simulation results from three different substrate topologies demonstrate that LaLRC-LaSP achieves better blocking performance than LRC-LaSP and an existing benchmark algorithm. For the opaque VONE, an ILP model is also formulated. We then design a LRC metric that considers the spectrum consecutiveness of SFLs. With this metric, a novel heuristic for opaque VONE, consecutiveness-aware LRC-K shortest-path-first fit (CaLRC-KSP-FF), is proposed. Simulation results show that compared with the existing algorithms, CaLRC-KSP-FF can reduce the request blocking probability significantly.

Optics InfoBase is the Optical Society's online library for flagship journals, partnered and copublished journals, and recent proceedings from OSA conferences.

Network function virtualization (NFV) is a promising technology to decouple the network functions from dedicated hardware elements, leading to the significant cost reduction in network service provisioning. As more and more users are trying to access their services wherever and whenever, we expect the NFV-related service function chains (SFCs) to be dynamic and adaptive, i.e., they can be readjusted to adapt to the service requests' dynamics for better user experience. In this paper, we study how to optimize SFC deployment and readjustment in the dynamic situation. Specifically, we try to jointly optimize the deployment of new users' SFCs and the readjustment of in-service users' SFCs while considering the trade-off between resource consumption and operational overhead. We first formulate an integer linear programming (ILP) model to solve the problem exactly. Then, to reduce the time complexity, we design a column generation (CG) model for the optimization. Simulation results show that the proposed CG-based algorithm can approximate the performance of the ILP and outperform an existing benchmark in terms of the profit from service provisioning.

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.