Datacom Networks Research Articles

SL-Hyper-FleX: a cognitive and flexible-bandwidth optical datacom network by self-supervised learning [Invited

This paper presents a cognitive flexible-bandwidth optical interconnect architecture for datacom networks. The proposed architecture leverages silicon photonic reconfigurable all-to-all switch fabrics interconnecting top-of-rack switches arranged in a Hyper-X-like topology with a cognitive control plane for optical reconfiguration by self-supervised learning. The proposed approach makes use of a clustering algorithm to learn the traffic patterns from historical traces. We developed a heuristic algorithm for optimizing the intra-pod connectivity graph for each identified traffic pattern. Further, to mitigate the scalability issue induced by frequent clustering operations, we parameterized the learned traffic patterns by a support vector machine classifier. The classifier is trained offline by self-labeled data to enable the classification of traffic matrices during online operations, thereby facilitating cognitive reconfiguration decision making. The simulation results show that compared with a static all-to-all interconnection, the proposed approach can improve the throughput by up to1.62×while reducing the end-to-end packet latency and flow completion time by up to3.84×and20×, respectively.

Software-defined networking control plane for seamless integration of multiple silicon photonic switches in Datacom networks.

Silicon photonics based switches offer an effective option for the delivery of dynamic bandwidth for future large-scale Datacom systems while maintaining scalable energy efficiency. The integration of a silicon photonics-based optical switching fabric within electronic Datacom architectures requires novel network topologies and arbitration strategies to effectively manage the active elements in the network. We present a scalable software-defined networking control plane to integrate silicon photonic based switches with conventional Ethernet or InfiniBand networks. Our software-defined control plane manages both electronic packet switches and multiple silicon photonic switches for simultaneous packet and circuit switching. We built an experimental Dragonfly network testbed with 16 electronic packet switches and 2 silicon photonic switches to evaluate our control plane. Observed latencies occupied by each step of the switching procedure demonstrate a total of 344 µs control plane latency for data-center and high performance computing platforms.

Heterogeneously integrated III–V/Si MOS capacitor Mach–Zehnder modulator

Hybrid silicon optical modulator brings efficiency benefits. Demand for more transmission capacity in data centres is increasing due to the continuous growth of Internet traffic. The introduction of external modulators into datacom networks is essential with advanced modulation formats. However, the large footprint of silicon photonics Mach–Zehnder (MZ) modulators will limit further increases in transmission capacity1,2,3,4. To overcome this, we introduce III–V compound semiconductors because the large electron-induced refractive-index change, high electron mobility and low carrier-plasma absorption are beneficial for overcoming the trade-offs among the voltage–length product (VπL), operation speed and insertion loss of Si MZ modulators. Here, we demonstrate an MZ modulator with a 250-µm-long InGaAsP/Si metal-oxide–semiconductor (MOS) capacitor phase-shifter and obtain a VπL of 0.09 Vcm in accumulation mode, an insertion loss of ∼1.0 dB, a cutoff frequency of ∼2.2 GHz in depletion mode and a 32-Gbit s–1 modulation with signal pre-emphasis. These results are promising for fabricating high-capacity large-scale photonic integrated circuits with low power consumption.

Directly Modulated DFB Laser on SiO$_{\bf 2}$ /Si Substrate for Datacenter Networks

Reducing the operating energy of a distributed feedback (DFB) laser is a critical issue if we are to use the device as a directly modulated light source employing wavelength division multiplexing technologies in short-distance datacom networks. A membrane buried heterostructure (BH) DFB laser on a SiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> layer is one candidate for reducing the operating energy because it provides a strong carrier and optical confinement in the active region. For low-cost fabrication, we have proposed and developed a fabrication procedure that employs the buried growth of an InP layer by using a directly bonded InP-based active layer on a SiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> /Si substrate, which enables us to use a large-scale Si wafer. To overcome the problem of the difference between the thermal expansion coefficients of Si, SiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> , and InP, we have used a thin active layer (~250 nm) on a SiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> /Si substrate as a template for the epitaxial growth of a III-V compound semiconductor. A lateral current injection structure is essential for fabricating a device with a 250-nm-thick template. Our fabricated DFB laser with a 73-μm cavity length exhibits a threshold current of 0.9 mA for continuous operation at room temperature and achieves lasing at up to 100 °C. We have also demonstrated 171-fJ/bit operation with a 25.8-Gb/s NRZ signal. These results indicate that the BH DFB laser on a SiO <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> /Si substrate is highly suitable for use as a transmitter for datacom applications.

Supply chain opportunities in high-speed electronics a strategic review of PCB, material and interconnection trends and needs for electronic assemblies for high-speed telecom, datacom networks and high-end computers 2009-2014

Supply chain opportunities in high-speed electronics a strategic review of PCB, material and interconnection trends and needs for electronic assemblies for high-speed telecom, datacom networks and high-end computers 2009-2014

Metropolitan area wavelength-division multiplexing testbeds for enterprise computing and disaster recovery

Fiber optic data communication networks face unique requirements for applications such as disaster recovery of multiterabyte and petabyte databases. We present experimental results from several enterprise computing testbeds that evaluate dense wavelength-division multiplexing (DWDM) performance in metropolitan area datacom networks. Many networking issues can be addressed through the components used within the DWDM platform; we discuss how current and emerging requirements can be addressed by new packages for 10-Gbit/s links, clock recovery circuits, and wavelength lockers. A comparison of three DWDM platforms evaluated in these testbeds is given, including features such as network bandwidth, scalability, performance over distance, and ease of management. We also report performance of datacom protocols such as enterprise system connection (ESCON), fiber connection (FICON), and FICON Bridge, when extended beyond industry standard distances using DWDM.

High-bandwidth graded-index polymer optical fiber enabling power penalty-free gigabit data transmission

A relation between the -3 dB bandwidths of graded-index plastic optical fibers (GI POFs) and the total power penalty of 50-m GI POF links is investigated in detail. The bandwidth of the GI POF is deliberately varied by controlling the index profile. It is theoretically and experimentally confirmed that sufficiently high bandwidth on the order of gigahertz is necessary for the GI POF, even for several hundreds of megabit per second (Mb/s) data rate, in order to achieve the power penalty free in the bit error rate performance of the link. In the case of silica-based multimode fiber links, it has been reported that the launch condition strongly affects the bit error rate performances; hence, a special launching technology for the silica-based multimode fiber is developed to achieve a 1-Gb/s transmission in the gigabit Ethernet protocol. In this paper, it is also found that the power penalty-free state is realized in the GI POF link, which is independent of the launch condition, when a GI POF with a nearly ideal index profile is used. The GI POF is a promising physical layer for realizing stable, high-speed and low-cost data-com networks.

Universal broadband mobility target for post-TINA infrastructure: toward the fourth generation

Future universal broadband mobile services present a challenge for telecommunications architectures, control, and management. The focus of the future mobile (fourth) generation vision is turned from capacity to services, from radio to network-wide issues. New types of applications will evolve that should be supported by an adequate programmable intelligent telecommunications infrastructure. A convergence between telecom and datacom networks will happen based on the communications middleware concept, which will provide universal secure connectivity between mobile users and their applications. Actual specifications of TINA-C do not address all such necessary issues. There is a need for technology evolution, enhancement, and integration to meet these new requirements at different levels. The author the concept of the universal broadband mobile telecommunications systems (UBMTS, or simply UBM) described as fourth-generation mobile systems. The UBMTS objective is both to extend mobile user access to the range of broadband services that will exist for broadband integrated services digital network (B-ISDN) users, and to extend the customization of new services related to personal communications systems (PCS) users.

A network strategy for survival

The telecom business is due for fundamental changes, concerning both technology and enterprise-but also concerning attitudes. This article discusses the merger between telecom and datacom, the business and technological trends-in order to outline future architecture for service and infrastructure. Our approach is that the causes of change are more fundamental than is normally conceived. Today the telecom community focuses on deregulation and competition, rather than observing that there will be different types of traffic dominating our networks -and businesses. We also tend to focus on broadband technology, rather than observing that the emerging services require online properties and a connectionless network. Our own attitude, to interpret new phenomena in telecommunications concepts, disables us from using new technology and understanding new business. An overall conclusion is, in sweeping terms, that our operations in the future will be dominated by datacom networks rather than telecom networks. This, in turn, requires a new scenario for interoperability and regulations-and calls for new business models. It is argued that the focus for regulations should be interoperability rather than infrastructure. Furthermore, a generic "hour-glass" architecture is sketched-assuming, for example, IP as the interoperability level separating the megabits per second market from the service market. The discussion in this article reflects one strand of strategic discourse on future networks within Telia and other European operators.