As chip architectures become more and more complex, traditional system-level and function-level simulation brings problems such as difficulty in debugging and slow speed in execution. In this paper, we propose an abst...
详细信息
ISBN:
(纸本)9798350386783;9798350386776
As chip architectures become more and more complex, traditional system-level and function-level simulation brings problems such as difficulty in debugging and slow speed in execution. In this paper, we propose an abstract-level simulation model for a network processor, which ignores most hardware and functional details. It only focuses on the components that are decisive for network system performance. As a result, we significantly reduced the modeling complexity of network processors and accelerated simulations by 5 to 8 times. This provides an efficient approach for agile hardware design and architecture exploration of network processors.
This paper presents a groundbreaking approach to energy management in 6G Cloud Radio Access networks (C-RAN) by integrating Power-over-Fiber (PoF) and solar harvesting technologies to power-feed femtocells. We explore...
详细信息
ISBN:
(纸本)9783903176621;9798350351132
This paper presents a groundbreaking approach to energy management in 6G Cloud Radio Access networks (C-RAN) by integrating Power-over-Fiber (PoF) and solar harvesting technologies to power-feed femtocells. We explore how these innovations can significantly enhance energy efficiency, performance, and sustainability in 6G networks. Our work introduces a smart green-energy management methodology, focusing on fusing PoF and solar harvesting sources, and evaluates their impact on network performance in terms of throughput and energy usage optimization. Through simulations, we demonstrate the potential of the usage of solar energy in reducing the carbon footprint of 6G C-RAN scenarios, while maintaining high levels of network reliability, availability, and performance under different weather conditions. Furthermore, simulation results show that the combination of femtocell feed-in of PoF and solar harvesting technologies serves up to 11,6% additional traffic throughput compared to no solar harvesting, harnessing up to 6.7% additional green energy in the femtocells. This research paves the way for more sustainable and efficient 6G wireless networks.
The physical layer in opticalnetworks is a set of transparent optical circuits - light-paths (LP)s - that has been extensively shown as reliably approximated by additive white and Gaussian noise channels. Noise sourc...
详细信息
ISBN:
(纸本)9783903176546
The physical layer in opticalnetworks is a set of transparent optical circuits - light-paths (LP)s - that has been extensively shown as reliably approximated by additive white and Gaussian noise channels. Noise sources setting the LP generalized SNR (GSNR) are the ASE noise from amplifiers, the nonlinear interference (NLI) from the Kerr effect and crosstalk from ROADMs. Coherent transceivers can be effectively modeled by the back-to-back characterization by defining the GSNR thresholds. We will describe the physics behind such approximation including the statistical impairments due the PMD, PDL and filtering penalties, and other uncertainties. Then, we will show how the physical layer digital twin (PHY-DT) that integrates the transmission impairment models can be exploited to reliably evaluate the QoT, latency and energy consumption on the topology. Finally, we will comment on the possible use of the PHY-DT as vendor neutral planning tool for disaggregated infrastructure and as on-line service within the multi-layer hierarchical controller. The optical devices controller can be separated from the optical circuit deployment, so enabling multi-vendor networks. We will also comment on the synergistic use of AI&ML techniques assisting the PHY-DT.
We explore S, C + L-band transmission in low-core count multi-core fibers (MCFs). After reviewing progress in wideband transmission demonstrations in both SMF and MCFs we focus on two experiments demonstrating 57 km, ...
详细信息
ISBN:
(纸本)9783903176546
We explore S, C + L-band transmission in low-core count multi-core fibers (MCFs). After reviewing progress in wideband transmission demonstrations in both SMF and MCFs we focus on two experiments demonstrating 57 km, 1 Pb/s transmission with 20 THz transmission bandwidth and 319 Tb/s transmitted over 3000 km in a 4-core MCF with standard cladding diameter. Finally, we briefly explore the potential benefit of increased spectral efficiency in MCF transmission by exploiting reduced (IC-XT) at lower wavelengths.
The performance of communication-intensive distributed machine learning (DML) workloads and other emerging applications can suffer from a traffic-topology mismatch in traditional data-center networks. This degradation...
详细信息
ISBN:
(纸本)9783903176546
The performance of communication-intensive distributed machine learning (DML) workloads and other emerging applications can suffer from a traffic-topology mismatch in traditional data-center networks. This degradation can be alleviated by performing a logical network topology reconfiguration. However, how to dynamically reconfigure the logical topology and steer the bandwidth efficiently with a control plane capable of efficiently adapting to the current data center traffic patterns without considerable overhead is still an open question. This paper presents a reversibility-aware deep reinforcement learning algorithm (RA-DRL) for optical switch reconfiguration in data center networks and validates it in an experimental testbed. Using our testbed, we show that appropriate optical-switch reconfiguration, driven both by a baseline DRL and an RADRL method, can improve the training performance of DML workloads under network congestion. More importantly, by incorporating the concept of reversibility in the training of the DRL agent, we demonstrate a 5x training-time decrease for a distributed computer-vision application and an improvement in convergence time by up to 64%.
Following a summary of the constraints associated with simulations using pillar-wise-based methods to design meta-surfaces, we will present our latest progress in the domain of simulating large scale meta-surfaces, in...
详细信息
ISBN:
(纸本)9798331516369;9798331516352
Following a summary of the constraints associated with simulations using pillar-wise-based methods to design meta-surfaces, we will present our latest progress in the domain of simulating large scale meta-surfaces, including convolutional neural network-driven emulations. We highlight the significance of accurately modeling the impact of the interaction between the pillars in a meta-surface, which enables the elimination of meta-surface designs with poor optical performances (e.g. zeroth orders in a beam shaper) during the design verification stages.
Datacenter is a big computing infrastructure that aims to reduce operating costs and energy consumption. Software-Defined networking (SDN) is a novel approach of networking that seeks to segregate control plane from d...
详细信息
ISBN:
(纸本)9798331515935;9783903176690
Datacenter is a big computing infrastructure that aims to reduce operating costs and energy consumption. Software-Defined networking (SDN) is a novel approach of networking that seeks to segregate control plane from data plane in order to reduce the complexity of network resource management. optical Burst Switching (OBS) is an energy-efficient switching technology that offers flexibility to support datacenter traffic. OBS separates also control and data planes and can be the most compatible optical switching paradigm with SDN. This paper proposes an optical datacenter network architecture based on OBS and aligned with SDN specifications. The results obtained from simulation demonstrate the efficacy of this architecture, achieving zero burst loss, high throughput, and low burst delay.
In this paper, we present Dynamic Cross-layer network Orchestration (DyCroNO), a dynamic service provisioning and load balancing mechanism for IP over opticalnetworks. DyCroNO comprises of the following components: i...
详细信息
ISBN:
(纸本)9783903176546
In this paper, we present Dynamic Cross-layer network Orchestration (DyCroNO), a dynamic service provisioning and load balancing mechanism for IP over opticalnetworks. DyCroNO comprises of the following components: i) an end-end (E2E) service provisioning and virtual path allocation algorithm, ii) a lightweight dynamic bandwidth adjustment strategy that leverages the extended duration statistics to ensure optimal network utilization and guarantee the quality-of-service (QoS), and iii) a load distribution mechanism to optimize the network load distribution at runtime. As another contribution, we design a real-time deep learning technique to predict the network load distribution. We implemented a Long Short-Term Memory-based (LSTM) method with a sliding window technique to dynamically (at runtime) predict network load distributions at various lead times. Simulations were performed over three topologies: NSFNet, Cost266 and Eurolarge using real-world traffic traces to model the traffic patterns. Results show that our approach lowers the mean link load and total resources significantly while improving the resource utilization when compared to existing approaches. Additionally, our deep learning-based method showed promising results in load distribution prediction with low root mean squared error (RMSE) and similar to 90% accuracy.
The environmental impact of telecommunications is becoming increasingly relevant in our society as more data are transmitted every day. On the one hand, telecommunication networks help, for example, by reducing the nu...
详细信息
ISBN:
(纸本)9783903176546
The environmental impact of telecommunications is becoming increasingly relevant in our society as more data are transmitted every day. On the one hand, telecommunication networks help, for example, by reducing the number of travels and visits;on the other hand, the energy consumed by the network elements and devices is reaching unprecedented levels as a consequence of the growing bandwidth demand needed by novel applications and services. In this contribution, we analyze key metrics to estimate the energy efficiency, in building an opticalnetwork, of several types of optical transceivers when employed in the most common topologies of metro aggregation networks, i.e., horseshoes, rings, and hub-and-spoke. We considered three different types of coherent pluggable modules: (i) 400G ZR point-to-point (P2P);(ii) 400G ZR+ P2P;and (iii) 100G/400G point-to-multipoint (P2MP) which can be enabled by digital subcarrier multiplexing (DSCM). We show that the P2MP approach needs-due to its inherent flexibility and the ability to simplify networkdesign-the lowest number of devices, and consequently the lowest amount of CO2 emissions among the three considered metro aggregation network scenarios.
This paper discusses the forward modeling and inverse design of the electromagnetic scattering characteristics of two-dimensional etched holes in the nanometre band based on the physics-informed neural network (PINNs)...
详细信息
暂无评论