Optical satellite networks, supported by optical inter-satellite links (OISLs), provide reliable and low-latency optical connectivity. However, periodic and predictable sun outage events significantly compromise OISL availability, leading to frequent OISL interruptions and reduced network reliability. Existing routing algorithms often overlook the regularity of sun outage-induced interrupts and their differentiated impacts on services, resulting in degraded service performance. To address this challenge, this paper proposes a sun outage-enhanced time discretization OISL model and introduces a sun outage link-aware routing (SOLR) algorithm. By incorporating joint awareness of sun outage patterns and service requirements, SOLR employs an adaptive optimization mechanism to dynamically adjust routing decisions within temporal windows. Experimental results demonstrate that SOLR extends stable path durations by 39.9%, reduces interruption rates by 28.5%, and decreases blocking rates by 36.4%, significantly outperforming link-state-based routing algorithms. By effectively mitigating the impact of sun outages, SOLR ensures continuous optical service connections. This interruption-tolerant framework bridges network modeling and service provisioning, offering a robust solution for mission-critical service in optical satellite networks.

Geo-distributed machine learning (GDML) can facilitate collaborative learning among geographically-dispersed data centers to meet the demands of distributed and privacy-preserving training for large-scale distributed Internet of Things applications. Unfortunately, the efficiency of distributed training tasks heavily depends on synchronized communication between multiple distributed models over bandwidth-limited wide area networks (WANs). The fine-grained Optical Transport Network (fgOTN), thanks to its adjustable bandwidth connections, represents more flexible transmission and has the ability for accurate synchronization across GDML tasks in WANs. However, flexible bandwidth assignment and complex interdependencies among tasks pose significant challenges to resource allocation for GDML in fgOTN. Specifically, flexible bandwidth assignment exacerbates resource competition among task flows, leading to decreased learning efficiency. This paper provides novel resource allocation solutions for GDML in fgOTN. We first formulate this problem as a linear programming aimed at maximizing the completion ratio of GDML tasks. Subsequently, we propose an innovative resource allocation algorithm based on genetic algorithm (GARA) for GDML in fgOTN. GARA considers both task completion and bandwidth adjustment through population generation based on prior knowledge and adaptive mutation based on completion ratio. Simulation analysis demonstrates that GARA effectively prioritizes resource allocation for high-priority tasks to alleviate resource competition, achieving the highest task completion ratio while avoiding excessive network reconfiguration.

The SFC deployment problem has proved to be NP-hard and has attracted great research interests. At present, most existing studies focus on deploying SFC through virtual network functions (VNFs) and few studies involve hybrid VNF and physical network function (PNF) environment. However, in the transition to a comprehensive network function virtualization (NFV) network, hybrid VNF and PNF environment is possible and important. Currently, VNF and PNF both play an important role in today’s cloud networks. PNF is far superior in processing speed, but VNF has advantages in flexibility and cost. Using VNF or PNF alone may not meet user needs, affecting user experience and system performance. Hence, this paper studies the SFC deployment problem in hybrid VNF and PNF environment. A cost and latency customized dynamic SFC deployment (CL-SFCD) scheme is proposed. Instead of giving priority to the PNF or the VNF roughly, the CL-SFCD scheme maps network functions to VNFs or PNFs according to the SFC requester’s personalized demands to satisfy the demands of different users. It aims to minimize the weighted value of total cost and latency. The CL-SFCD scheme is formulated as a mixed integer linear programming (MILP) model for exact results. Heuristic algorithms are developed to provide an approximate optimal solution in large-scale cloud networks. Numerical results show that the CL-SFCD scheme achieves performance that balances cost and latency based on user requirements than the VNF-only,VNF-GA, and PNF-only deployment schemes.

In the sixth-generation fixed network (F6G), network security becomes an important topic. Encryption is an effective method to prevent network attacks and realize network security. Quantum key distribution (QKD) is a promising technology to effectively address the challenge by providing secret keys due to the laws of quantum physics. New services such as high immersion experience and holographic have the characteristics of time-varying bandwidth and requirements. The introduction of optical service unit (OSU) technology makes it possible to provide the exact bandwidth used by the service. In optical transport networks, a lightpath needs to be established before service transmission, and will be removed after service transmission. Signaling is used for lightpath establishment, removal, and bandwidth adjustment. Data information transmitted in data layer and signaling information transmitted in control layer are highly vulnerable to cyberattacks, such as eavesdropping. The supply of bandwidth and key resources need to be optimized to achieve secure and stable service transmission in optical networks. Hence, how to realize bandwidth and key on demand (BKoD) provisioning for dynamic services is a key problem. To improve the flexibility of bandwidth and key resource allocation and utilization, a QKD-secured OSU-based optical transport network can be deployed. In this paper, a novel QKD-secured OSU-based optical transport network architecture is proposed and a service aware dynamic resource provisioning (SADRP) algorithm is proposed to realize BKoD. The proposed architecture uses the QKD technique to provide keys for both signaling information and data information for the first time. The proposed algorithm supplies resources according to the dynamic demand of bandwidth and key, so as to achieve the balance between dynamic demand and static resource utilization. Simulations results show that compared with the benchmark algorithm, the SADRP algorithm reduces blocking probability by 4.16%, reduces bandwidth resource utilization rate by 4.39%, reduces key resource utilization rate by 3.48%, and improves security rate by 4.17%.

Following successful adoption of cloud computing, many service providers (SPs) are now using high-performance Virtual Machines (VMs) located in large datacenters owned by public cloud infrastructure providers to deploy their virtual network functions (VNFs). Since using these VMs has a cost depending on utilization time, a complex problem of VNF placement and scheduling (VPS) must be addressed to achieve satisfactory network performance (e.g., latency) while minimizing the cost paid to lease VMs. In this study, a cost-efficient VPS scheme (CE-VPS) is proposed to address the VPS problem in public cloud networks considering dynamic requests of ordered sequences of VNFs. Our CE-VPS scheme goes beyond existing solutions as it models some important practical aspects such as an additional latency incurred by booting a VM and installing a VNF instance. Also, CE-VPS considers that VNFs can be multi-threaded or single-threaded, and that their throughput as a function of allocated computing resources must be modeled differently. CE-VPS is formulated as a mixed inter linear program (MILP) and also as an efficient heuristic algorithm. CE-VPS achieves lower cost and latency than conventional Best-Availability and Cost-Efficient Proactive VNF Placement schemes, and a better trade-off between resource consumption and latency performance than a conventional Low-Latency scheme.