Comparative analysis of optical data center switches

Data center networks (DCNs) are at the forefront of a radical transformation, fueled by the rapid rise of artificial intelligence (AI), cloud-native adoption, and enterprise digitalization, which together place unprecedented demands on their design and performance. From healthcare AI and intelligent transportation to business analytics, modern DCNs are expected to support a diverse and evolving mix of bandwidth-demanding applications, each with distinct quality of service (QoS) requirements. Yet, most state-of-the-art DCN solutions rely on oversimplified DCN data traffic models, failing to address the required application-level differentiation in terms of end-to-end (E2E) latency demands. In response, we propose ALAS-DCN, a novel, to our knowledge, application-aware, low-latency, scalable DCN architecture, based on optical packet switching in the intra-rack and inter-rack domains that is both designed and rigorously evaluated using commercial QoS specifications and realistic traffic patterns of different applications coexisting within the same DCN. ALAS-DCN integrates three hierarchical optical sub-networks: intra-rack, inter-rack, and bridge, each governed by a custom wavelength division multiplexing (WDM) medium access control (MAC) protocol that dynamically prioritizes critical traffic packets, according to the specific application E2E latency demands. Unlike prior approaches, our evaluation leverages realistic workload traces generated by established traffic generators, capturing the application diversity within modern data center (DC) environments. Extensive experimental simulations across five representative DC categories, each hosting a different mix of modern applications, demonstrate that ALAS-DCN consistently achieves up to 95% throughput efficiency and sub-µs latency for high priority traffic across a variety of applications, consistently meeting their differentiated latency constraints. Scalable up to 20,000 servers, ALAS-DCN redefines the DCN as not just a transport layer but an intelligent enabler of next-generation application performance.

Optical packet switching (OPS) is being considered as one of the switching technologies for a future optical internet. For contention resolution in an optical packet switching (OPS) system, the wavelength dimension is generally used in combination with a fibre delay line (FDL) buffer. In this study, the authors propose to reduce the number of tunable wavelength converters (TWCs) by sharing TWCs for cost-effective design of an asynchronous OPS system with a shared or an output FDL buffer. Asynchronous and variable-length packets are considered in the OPS system design. To investigate the number of TWCs needed for the OPS system, an algorithm is proposed, which searches for an available TWC and an unused internal wavelength, as well as an outgoing channel. This algorithm is applied to an OPS system with a shared or an output FDL buffer. Also, the number of internal wavelengths (i.e. the conversion range of the TWC) needed for an asynchronous OPS system is presented for cost reduction of the OPS system.

At a switching node in an optical packet switching (OPS) network, packet contention occurs whenever two or more packets try to reach the same output port of the node at the same wavelength, consequently it degrades network performance. Previous works on optical packet switching flourishingly adopted optical buffering with optical fibre delay line (FDL), wavelength conversion, or deflection routing to resolve packet contention. In this paper, we focus on the optical fibre delay line approach, especially the selection of proper length of FDL under asynchronous variable length packet traffic. As the length of fibre coils equipped in a FDL optical buffering system is impossible to be reconfigured frequently, a proper length of FDL should be determined during manufacture. The motivation is to observe the impact factors on FDL optical system performance in a WDM network scenario, and provide the empirical data as supports to select the proper length of FDL in an OPS node system.

In this paper, a novel optical packet switch is proposed, which uses a set of shared fiber delay lines (FDLs) and a set of shared tunable wavelength converters (TWCs) to resolve optical packet contentions. In addition, two control algorithms, i.e., Fiber-First (FF) and Wavelength-First (WF), are proposed to schedule optical packets. Performance of the novel switch is evaluated by means of simulation experiments. Simulation results show that based on either of these two control algorithms, the switch can achieve super performance without employing a lot of FDLs and TWCs. Moreover, the performance of FF and WF is compared. Detail analyses are also given in this paper.

The service evolution and the rapid increase in traffic levels fuel the interest toward switching paradigms enabling the fast allocation of Wavelength Division Multiplexing WDM channels in an on demand fashion with fine granularities (microsecond scales). For this reason, in the last years, different optical switching paradigms have been proposed (Sabella et al., 2000): optical-packet switching (OPS), optical-burst switching (OBS), wavelength-routed OBS, etc. Among the various all-optical switching paradigms, OPS attracts increasing attention. Owing to the high switching rate, Semiconductor Optical Amplifier (SOA) is a key technology to realize Optical Packet Switches. We propose some Optical Packet Switch (OPS) architectures and illustrate their realization in SOA technology. The effectiveness of the technology in reducing the power consumption is also analyzed. The chapter is organized in three sections. The main blocks (Switching Fabric, Wavelength Conversion stage, Synchronization stage) of an OPS are illustrated in Section 2 where we also show some examples of realizing wavelength converters and synchronizers in SOA technology. Section 3 introduces SOA-based single-stage and multi-stage switching fabrics. Finally the SOA-based OPS power consumption is investigated in Section 4.

Optical Packet Switching (OPS) has emerged as a promising candidate for the future all-optical core network. A crucial issue in such networks is how the packet length distribution (PLD) influences the performance under different traffic scenarios. In this paper we investigate how a deterministic, empirical and exponential PLD influences the Packet Loss Rate (PLR) in an optical packet switch in the case of a Poisson arrival process and a bursty Hyperexponential arrival process. We consider optical packet switches with and without Fibre Delay Line (FDL) contention buffers. Our results show that in the case of OPS without FDL buffering and with a Hyperexponential arrival process, the empirical PLD yields the lowest PLR. However, in the case of OPS with FDL buffering, the deterministic PLD yields the lowest PLR.

Modeling and performance analysis of a data center network (DCN) combining optical packet switching (OPS) and agile optical circuit switching (OCS) in a unified platform are presented. Hybrid optoelectronic packet routers which combine optical switching fabric and electronic buffers are used, supporting 100 Gbit/s (25 Gbit/s × 4 wavelengths) optical packets. OCS transmission is implemented by a novel “express path,” (ExP) an instant wavelength path on demand, which enables certain flows following the ExP with no packet contention, whereas ordinary OPS packets are switched with a deflection routing algorithm and buffering for contention resolution. This architecture can support OCS for large and reliability-sensitive data transferring and OPS for latency-sensitive applications. The performance of the network is investigated via numerical simulations of top-of-rack switch (ToR)-to-ToR traffic in an N-dimensional (up to 4096 node) torus topology. We have made the following observations: 1) torus topology results in high throughput and can be scaled out by increasing the dimension; 2) contention resolution between OPS and agile OCS is needed and our proposed strategies enhance the overall performance; and 3) in a lightly to moderately loaded case of OPS traffic, more than 100 ExPs can be supported simultaneously with acceptable affect on OPS.

We demonstrate end-to-end 25Gb/s true optical packet switching featuring burst-mode reception with $\text{Hipo}\lambda\text{aos}$ Optical Packet Switch architecture. Error-free performance at 10−9 was obtained for all validated port-combinations.

Packet contention is a major issue in optical packet switching network. To resolve contention, optical buffering approach is used using Fiber Delay Lines (FDLs). Yet there is a heavy packet loss rate due to unavailability of output-port and free FDLs. This paper proposes a selective minimum delay value algorithm (SMDV) to resolve contention for shared FDL buffering architecture in order to ensure the availability of output-port for the contended packet after coming out from FDL. Simulation result shows that proposed scheme can achieve a packet loss rate of ~ even at heavy traffic load 0.9 for 32×32 switch using 32 variable length FDL.

Multi-wavelength switching in IP optical nodes adopting different buffering strategies

Multi-tenancy is a key feature of modern data centers. It allows for the existence of multiple independent virtual infrastructures, called virtual slices, on top of the same physical infrastructure, each of them specially tailored to the tenants' needs. In such a scenario, an optimal mapping of the virtual slices plays a capital role toward an efficient utilization of the data center network resources, potentially saving costs for the data center owner. However, due to the increasing trend of bringing optics to data center networks, specific virtual slice mapping mechanisms accounting for the particularities of the optical medium (e.g., wavelength continuity constraint) have to be investigated. For this, we present an integer linear programming (ILP) model for optimally mapping a set of virtual slices from different tenants in a hybrid optical circuit switching (OCS)/optical packet switching (OPS) data center network with the aim to minimize the necessary optical transponders to be equipped in the network. Additionally, we also present a lightweight heuristic for the cases in which the ILP model scalability is compromised. The benefits of the proposals are highlighted by benchmarking them against a pure OCS solution through extensive simulations.

We develop an integrated optical packet and circuit switching node in which signaling via optical packet switching is implemented. We demonstrate 80 (8λ × 10) Gbit/s optical packet switching with <;1E-4 packet-loss-rate, 8-lightpaths establishment, and dynamic wavelength-resources control.

This book presents an in-depth deliberation on optical networks in four parts, capturing the past, present, and ensuing developments in the field. Part I has two chapters presenting an overview of optical networks and the enabling technologies. Part II has three chapters dealing with the single-wavelength optical networks: optical LANs/MANs, optical access networks using passive optical network architecture, SONET/SDH, optical transport network and resilient packet ring. Part III consists of four chapters on WDM-based optical networks, including WDM-based local/metropolitan networks (LANs/MANs) using single and multihop architectures over passive-star couplers, WDM/TWDM access networks as an extension of PONs with WDM transmission, WDM metro ring networks covering circuit-switched (using point-to-point WDM and wavelength-routed transmission) plus packet-switched architectures and WDM long-haul backbone networks presenting the offline and online design methodologies using wavelength-routed transmission. Part IV deals with some selected topics in six chapters. The first deals with transmission impairments and power-consumption issues in optical networks, while the next three chapters deal with the survivable optical networks, network control and management techniques, including GMPLS, ASON, and SDN/SDON, and datacenter networks using electrical, optical, and hybrid switching techniques. The final two chapters present elastic optical networks using flexible grid for better utilization of the optical-fiber spectrum and optical packet and burst-switched networks. The three appendices present the basics of the linear programming techniques, noise processes encountered in the optical communication systems, and the fundamentals of queuing theory and its applications in telecommunication networks. (238 words)

Guest Editorial

Utilizing T-band (1000 nm to 1260 nm) for optical communications is promising for short reach, and large capacity networks, such as data centers or access networks. It is feasible to use this with low-cost coarse wavelength division multiplexing (WDM). However, a tunable wavelength light source is necessary for such applications. In this paper, we propose a new configuration for an external cavity laser, which uses a silica-based arrayed waveguide grating (AWG) for the wavelength selecting element. The external cavity laser consists of a gain chip with high reflection (HR) and anti-reflection (AR) coated facets, coupling lenses, an AWG with AR/HR coatings, and an output fiber. The AWG has 17 connection ports, which correspond to 17 wavelengths with a channel spacing of 1.67 nm. The width of the connection port waveguides was optimized to achieve high coupling efficiency. The AWG chip size is 15 mm x 30 mm. The active layer in the gain chip has InAs quantum dots. The spontaneous emission 3-dB bandwidth was 48 nm (1108 nm to 1156 nm) when a current of 150 mA was injected into the gain chip. The lasing wavelength of the external cavity laser was successfully tuned from 1129.9 nm to 1154.4 nm by selecting the connection ports of the AWG. The typical threshold current was about 130 mA.