Search for: All records

In network-cloud ecosystems, large-scale failures affecting network carrier and datacenter (DC) infrastructures can severely disrupt cloud services. Post-disaster cloud service restoration requires cooperation among carriers and DC providers (DCPs) to minimize downtime. Such cooperation is challenging due to proprietary and regulatory policies, which limit access to confidential information (detailed topology, resource availability, etc.). Accordingly, we introduce a third-party entity, a provider-neutral exchange, which enables cooperation by sharing abstracted information. We formulate an optimization problem for DCP–carrier cooperation to maximize service restoration while minimizing restoration time and cost. We propose a scalable heuristic, demonstrating significant improvement in restoration efficiency with different topologies and failure scenarios. 

Cooperation among telecom carriers and datacenter providers (DCPs) is essential to ensure the resiliency of network-cloud ecosystems. To enable efficient cooperative recovery in case of traffic congestion or network failures, we introduce a novel, to our knowledge, multi-entity cooperation platform (MCP) for implementing cooperative recovery planning. The MCP is built over distributed ledger technology (DLT), which ensures decentralized and tamper-proof information exchange among stakeholders to achieve open and fair cooperation. We experimentally demonstrate a proof-of-concept DLT-based MCP on a testbed. We showcase a DCP–carrier cooperative planning process and the corresponding recovery in the data-plane, showing the possibility of multi-entity cooperation for quick recovery of network-cloud ecosystems. 

In the post-pandemic era, global working patterns have been reshaped, and the demand for online network services has increased significantly. Therefore, cross-data-center content migration has become a relevant problem to address, leading to higher attention in data backup/recovery planning. Beyond traditional pre-disaster content redundancy approaches, this work focuses on the challenge of rapid post-disaster content evacuation under the threat of cascading failures. In fact, due to the interdependence of data centers (DCs), inter-DC optical networks, and power grid networks, disasters may have a domino effect on these infrastructures, with their impact gradually expanding over time and space. In this paper, we propose two trajectory models that capture the dynamic evolution of cascading failures, and we propose a trajectory-based content evacuation (TCE) strategy that considers the spatiotemporal evolution of cascading failures to minimize content loss. Numerical results show that, when each DC needs to evacuate about 200 TB of massive content, TCE can reduce content loss by up to 25% compared to baseline strategies. 

Not AvailableNetwork operators tend to migrate multiband optical networks to wider bands by batch upgrade (i.e., with a pay-as-you-grow strategy affecting only a subset of links at a time). However, temporary interruptions of services routed along the fibers that require equipment upgrades can create significant service blocking, which can only be avoided with interim lightpath re-allocation during the upgrade period. To seamlessly upgrade the network from C+L to C+L+S bands, we propose a batch upgrade strategy to reduce the upgrade cost, and a band-selective re-allocation method during the upgrade period to minimize blocking probability (BP). Simulations on the US-24 topology demonstrate up to nearly 50% cost reduction and BP as low as 0.3%. 

Abstract Efficient network management in optical backbone networks is essential to manage continuous traffic growth. To accommodate this growth, network operators need to upgrade their infrastructure at appropriate times. Given the cost constraint of upgrading the entire network at once, upgrading the network periodically in multiple batches is a more pragmatic approach to meet the growing demands. While multi-period, batch-upgrade strategies to increase network capacity from the conventional C band to C+L bands have been proposed, they did not consider so far the possibility to re-provision existing traffic. In this work, we investigate how to selectively re-provision connections from C band to L band during a batch upgrade. This is to ensure greater availability of C-band resources which can help to delay network upgrade and hence reduce upgrade cost, while limiting the number of disrupted connections in the network. This study proposes two re-provisioning strategies, namely, Budget-Based (BB) and Margin-Aware (MA) re-provisioning, which rely on the Quality of Transmission (QoT) of lightpaths. These strategies leverage the knowledge of Generalized Signal-to-Noise Ratio (GSNR) to choose which lightpaths to re-provision. We compare these strategies with a baseline distance-based strategy that uses path length to select and re-provision lightpaths. We also incorporate Machine Learning techniques for QoT estimation of lightpaths to reduce the computational time required for optical-path feasibility check. Numerical results show that, compared to distance-based strategy, BB and MA strategies reduce disruption by about 22% and 27%, respectively, in representative network topologies. 

We propose a privacy-preserving strategy based on federated learning to localize soft failures in multi-carrier optical networks using a self-supervised approach on unlabeled data. Evaluations conducted on data from a testbed demonstrate the effectiveness of the proposed strategy. 

In network-cloud ecosystems, cooperation among different entities, for example, network carriers and datacenter providers (DCPs), is crucial to enhance resiliency, especially during large-scale failures or congestion. However, such cooperation is constrained by limited visibility of confidential information, for example, network topology, resource availability, and so on, of different entities owing to proprietary and regulatory policies. To facilitate cooperation, we present and discuss the role of a third-party entity, called provider neutral exchange (PNE), which acts as a broker/mediator and enables cooperation among multiple entities by sharing abstracted (instead of detailed) information of individual entities. We design novel cooperation strategies for post-disaster service restoration and categorize them as: multi-carrier cooperation and DCP-carrier cooperation. Results under different failure scenarios show benefits of cooperation in terms of service-restoration efficiency, restoration time, and restoration cost. 

Large-scale carrier networks are fundamental ICT infrastructures that support future 5G/6G services, and their resilience is a primary societal concern. Differently from single-carrier networks (in which one carrier owns multiple networks), in multi-carrier network ecosystems (in which the networks in the fields are operated by different carriers), cooperation among such different carriers is crucial to achieve resilience against large-scale failures. However, such cooperation is challenging since carriers may not disclose confidential information, e.g., detailed resource availability. In this study, we investigate how to perform carrier cooperative recovery in the case of large-scale failures/disasters. We propose two-stage carrier-carrier cooperative recovery planning by incorporating a coordinated scheduling for faster recovery. Through numerical evaluation, we confirm the potential benefit of carrier cooperation in terms of both recovery time and recovery cost reduction. 

Not AvailableEfficient network management in optical backbone networks is crucial for handling continuous traffic growth. In this work, we address the challenges of managing dynamic traffic in C- and C+L-band optical backbone networks while exploring application flexibility, namely the compressibility and delayability metrics. We propose a strategy, named Delay-Aware and Compression-Aware (DACA) provisioning algorithm, which reduces blocking probability, thereby increasing information-carrying capacity of the network compared to baseline strategies. 

Multi-band transmission is a promising solution for capacity enhancement in optical networks. We propose a novel strategy, named C to C+L Upgrade (CLU), to gradually upgrade links from C to C+L bands. We develop a Recurrent Neural Network (RNN)-based model to efficiently predict links for upgrade, based on network state and resource utilization, to reduce blocking and upgrade cost. Our results show that CLU outperforms baseline strategies (which do not employ predictive decisions) by upgrading fewer links at appropriate times.