Optical Packet Switching Research Articles

A Brief Review on the Methods that Improve Optical Burst Switching Network Performance

Abstract In this tutorial, we give an introduction to optical burst switching (OBS) networks and compare different methods that have been given to improve performance of these networks. OBS is an alternative to optical circuit switching and optical packet switching by separating data from control packets and sending them in a different wavelength. OBS networks suffer from an issue based on their bufferless nature, burst contentions. In these networks, when two data bursts are competing on the same wavelength, one of them is dropped and it leads to a significant reduction on the performance of the network. A lot of researches have been done to solve this problem. Some of them are addressed in this paper.

Performance comparison between recirculating loop buffer-based optical packet switch architectures

Due to the high cabling complications and limited scalability, the Ethernet switching is not a well-suited technology for the next-generation data center systems (NGDCs). In designing of interconnects for NGDCs, the first step is to select an appropriate switching technology. The best solution to the problems faced by NGDCs is the use of optical switching technology in the core of data centers. In this paper, three recirculating loop buffer-based optical packet switches are discussed. The first two switch discussed in this paper were proposed in the recent past, and the third switch is proposed by the authors of this paper. In this paper, a performance comparison between the switches is performed by considering physical and network layer parameters. The obtained results presented in the paper show that the performance of our proposed switch is better in comparison with other switches.

Multiple-Code Technique for Multi-Rate Transmissions in Optical Packet Switching Networks Based on OCDMA Labels

Supporting multi-rate transmission is an essential factor in current optical packet switching (OPS) networks. In this paper, the author studied a multi-rate scheme capable of forwarding packets with different signal rates based on label switching. The multiple-code (MC) technique was employed to label a packet by conveying its payload bits to multiple optical code-division multiple-access (OCDMA) labels. Spectral-amplitude-coding (SAC), which represents the chips in an OCDMA code as a set of wavelengths, was introduced to remove the multiple-access interference (MAI) from the overlapping among labels. The author tested the system effectiveness by conducting numerical analysis to formulate bit-error probability (BEP) and spectral efficiency (SE). The simulation results showed that the proposed network had a stable BEP performance when switching the packet flows of multiple data-rates.

Data Center’s Energy Savings for Data Transport via TCP on Hybrid Optoelectronic Switches

We report on possible 75% lower energy consumption for packet transport in data center networks replacing electronic with hybrid optical packet switching (optical switches with a shared electronic buffer) combined with the enhanced transmission control protocol.

An optical packet metro architecture for Fixed Mobile Convergence in the cloud era

The present paper addresses future network architectures considered for Fixed Mobile Convergence, where users can access resources transparently either through fixed access technologies or through the mobile RAN. Optical technologies rely on Wavelength Division Multiplexing (WDM) to meet continuously growing bandwidth demands. However, although in the core network optical circuit switching may efficiently support huge traffic aggregates, closer to the access, in future Metropolitan Area Networks (MAN), it is necessary to provide a finer granularity, and a dynamical bandwidth control, to very heterogeneous traffic demands. This is supported by Optical Packet Switching (OPS) relying on WDM slotted Add/Drop Multiplexing (WSADM). A flow-based transfer mode, compatible with SDN control is first specified, and applied to bi-directional MAN rings. This mode is shown to provide several methods for supporting both local and global repair in case of failure. Mathematical models are derived to assess optical packet insertion performance and are used to illustrate how WSADM rings can support MAN performance requirements. Lastly the impact of locating the IP edge, and Data Centres closer to the users than in legacy architectures, within the MAN is addressed.

Bit Error Rate Analysis of Hybrid Buffer-Based Switch for Optical Data Centers

Abstract Data centers are very critical elements which store and distribute data over the Internet. current data centers heavily rely on electronics, and thus require vast amount of power and speed is limited. These limitations can be tackled using optical technology. In general optical devices consume lesser amount of power, and support very high data rates. In this paper, a hybrid buffer-based optical packet switch which can be used in data centers as aggregated or core switches is discussed. The physical layer analysis is presented and performance is measured in terms of bit error rate (BER) obtained at various power levels. The analysis is carried out in presence of crosstalk of various components and for acceptable BER≤10–9, minimum power required for all optical, all electronic and hybrid buffer is evaluated. Finally power penalty is also evaluated which is required to overcome the effect of amplified spontaneous noise (ASE) and components crosstalk.

Improving Performance of Optical Networks by a Probable Approach

Abstract Optical Burst Switching (OBS) networks are new emerging techniques that alternate Optical Circuit Switching (OCS) and Optical Packet Switching (OPS) techniques. The high bandwidth of fiber links in these networks, besides WDM (Wavelength Division Multiplexing) helps the Internet to handle the heavy traffic. OBS networks have an important issue that needs to be tackled. These networks are bufferless in their nature, so when two bursts are trying to reserve one wavelength one of them is dropped. This drop can be assumed as an indicator of congestion; as a result, the sending rate will be decreased, despite the fact that the network may not be congested. So, wired protocols like TCP (Transmission Control Protocol) cannot work to their high potential in OBS networks. In this paper, we propose a new scheme based on a probable approach that can distinguish burst drops caused by contentions from burst drops caused by contention, so we can improve the performance of the network. Extensive simulative studies show that the proposed algorithm outperforms TCP Vegas in terms of throughput and packet delivery count.

A 1024-Port Optical Uni- and Multicast Packet Switch Fabric

Novel Data Center (DC) architectures based on resource disaggregation hold the potential to increase the resource utilization at a reduced energy and cost envelope, imposing, however, significant challenges in the underlying DC network infrastructure that has to provide high-radix and high-bandwidth connectivity, while maintaining sub-μs latency. At the same time, the explosive growth of AI applications has transformed the traffic profiles on the majority of DC workloads calling for support of advanced network functionalities, such as multicasting. In this direction, we scale-up the Hipoλaos optical packet switch architecture to a thousand-port layout, at the same time upgrading its functionality to allow latency-free intra-tray multicasting. The feasibility of a 10.24-Tb/s capacity switch fabric with a 10-Gb/s line rate in a 1024-port configuration, with four-packet buffering capacity is experimentally demonstrated, revealing error-free performance with <3.4-dB power penalty. Optical multicasting to up-to five nodes is also experimentally validated at 10 Gb/s, with a power penalty variation of less than 1 dB between the different output ports, while multicast operation across the whole required spectral range of a 1024-port switch is investigated, utilizing pairs of wavelengths with different channel spacing. Network-level analysis of the 1024-port Hipoλaos switch revealed sub-μs p90-latency and ∼85% throughput, for the unicast case, when utilizing a two-packet-buffer design. Finally, the simulation model was extended to support the Hipoλaos multicast friendly layout, revealing up to 4250% throughput improvement, as compared to the conventional layout, while maintaining sub-μs latency.

Electronically Tunable Distributed Feedback (DFB) Laser on Silicon

AbstractAn electronically tunable distributed feedback (DFB) laser heterogeneously integrated on a silicon photonics platform is experimentally demonstrated. Tuning is achieved through carrier injection into the tuning layer of a tunable twin‐guide III–V membrane on silicon. A 2 nm continuous tuning range is achieved with a single tuning current. Continuous‐wave single‐mode laser operation with a side‐mode suppression ratio larger than 44 dB across the tuning range is obtained. Measured wavelength switching times are on the order of 3 ns and non‐return‐to‐zero on‐off keying direct modulation at 12.5 Gbit s−1 is readily achieved. This work can be a major advance toward the realization of optical burst or packet switching systems for future data center networks.

Petascale: A Scalable Buffer-Less All-Optical Network for Cloud Computing Data Center

Optical packet switching is considered as a long-term solution for future data center networks due to their technological strengths, including high throughput, fine switching granularity, and excellent flexibility. However, most prior optical designs are either difficult to scale, costly to solve the packet collision, or inflexible to exploit the multiple wavelengths to increase the end-to-end connectivity. To this end, we present PETASCALE, a scalable, lossless, and high-performance optical data center network. By embedding full-mesh into a bipartite graph, PETASCALE is able to scale beyond 60 000 servers while achieving up to 2.5 and 2 times saving in a number of switches and cables, respectively, compared with fat tree topology. PETASCALE leverages the negative acknowledgment (NACK) and retransmission scheme to solve the packet collision, thus eliminating the complex and costly buffers. Then by leveraging a state retention mechanism, the contention can be notified over multiple hops in near real time. To further improve the bandwidth utilization, a wavelength routing algorithm is proposed to restrict the packet collision domain within the source pod. Moreover, PETASCALE designs a novel switch structure to support both the multi-hop NACK and wavelength routing. Our simulation results show that PETASCALE is able to reduce the end-to-end latency by more than 50% compared with an electrical fat tree network. For throughput, PETASCALE can deliver about 89% bisection bandwidth of a non-blocking network under uniform traffic pattern. While under the local traffic pattern, it can even achieve higher throughput than fat tree and other optical designs.

Experimental demonstration of an ultra-low latency control plane for optical packet switching in data center networks

Optical interconnection networks have the potential to reduce latency and power consumption while increasing the bisection bandwidth of data center networks compared to electrical network architectures. Optical circuit-switched networking has been proposed but it is reconfigurable in milliseconds. Although switches operating on nanosecond timescales have been demonstrated, centrally scheduling such switching architectures is considered to be of high complexity, incurring significant delay penalties on the total switching latency. In this paper we present a high-speed control plane design based on a central switch scheduler for nanosecond optical switching which significantly reduces the end-to-end latency in the network compared to using the best electronic switches. We discuss the implementation of our control plane on field-programmable gate array (FPGA) boards and quantify its delay components. We focus on the output-port allocation circuit design which limits the scheduling delay and the end-to-end latency. Using our FPGA-implemented control plane, for a 32 × 32 switch, we experimentally demonstrate rack-scale optical packet switching with a minimum end-to-end head-to-tail latency of 71.0 ns, outperforming current state-of-the-art electronic switches. The effect of asynchronous control plane operation on the switch performance is evaluated experimentally. Finally, a new parallel allocation circuit design is presented decreasing the scheduling delay by 42.7% and the minimum end-to-end latency to 54.6 ns. More importantly, it enables scaling to a switch double the size (64 × 64) with a minimum end-to-end latency less than 71.0 ns. In a developed cycle-accurate network emulator we demonstrate nanosecond switching up to 60% of port capacity and average end-to-end latency less than 10 μs at full capacity while maintaining zero packet loss across all traffic loads.

Arrayed Waveguide Grating and Re-Circulating Buffer Based Optical Packet Switch

Abstract Optical packet switching has gained popularity in past decade. However, due to technical challenges all- optical switching is not implemented till date, still hybrid switching where both optical and electrical switching are used simultaneously by-passing the limitations of both the technologies. In fiber delay lines (FDLs) buffering is limited due to accumulated losses and noises, and most of the times short term storage is enough in optical switching, but in case of long term storage, re-circulating type buffer can be used, but again due to accumulation of degrading terms re-circulation count is limited therefore in case of longer duration storage, electronic buffer is better choice. In this paper, we have proposed a hybrid buffer based optical switch where both optical and electronic buffers are used for storage of contending packets, and optical buffer is re-circulating in nature. The performance of the switch is measured in terms of bit error rate (BER) and packet loss probability, and BER performance is compared with recently proposed design, and it has been found that the proposed switch is comparatively much superior to earlier one.

Optics in Computing: From Photonic Network-on-Chip to Chip-to-Chip Interconnects and Disintegrated Architectures

Following a decade of radical advances in the areas of integrated photonics and computing architectures, we discuss the use of optics in the current computing landscape attempting to redefine and refine their role based on the progress in both research fields. We present the current set of critical challenges faced by the computing industry and provide a thorough review of photonic Network-on-Chip (pNoC) architectures and experimental demonstrations, concluding to the main obstacles that still impede the materialization of these concepts. We propose the employment of optics in chip-to-chip (C2C) computing architectures rather than on-chip layouts toward reaping their benefits while avoiding technology limitations on the way to manycore set-ups. We identify multisocket boards as the most prominent application area and present recent advances in optically enabled multisocket boards, revealing successful 40 Gb/s transceiver and routing capabilities via integrated photonics. These results indicate the potential to bring energy consumption down by more than 60% compared to current QuickPath Interconnect (QPI) protocol, while turning multisocket architectures into a single-hop low-latency setup for even more than four interconnected sockets, which form currently the electronic baseline. We go one step further and demonstrate how optically-enabled eight-socket boards can be combined via a 256 × 256 Hipoλaos Optical Packet Switch into a powerful 256-node disaggregated system with less than 335 ns latency, forming a highly promising solution for the latency-critical rack-scale memory disaggregation era. Finally, we discuss the perspective for disintegrated computing via optical technologies as a mean to increase the number of synergized high-performance cores overcoming die area constraints, introducing also the concept of cache disintegration via the use of future off-die ultrafast optical cache memory chiplets.

Label Stacking Scenarios in Hybrid Wavelength and Code-Switched GMPLS Networks

Multi-protocol label switching (MPLS) is a promising solution to implement high-speed internet protocol (IP) networks by reducing the layer number. To meet the increasing demand for data traffic, optical packet switching (OPS) is integrated under IP to provide high bandwidth to end users. Generalized MPLS (GMPLS) is perfectly compatible with the routing algorithm in IP/MPLS as it supports packet-switching functions. In this paper, we investigate the label stacking scenarios in GMPLS networks. In GMPLS, label stacking is done to reduce the node complexity by appending multiple labels to a single packet. Wavelength-division multiplexing (WDM) and optical code-division multiplexing (OCDM) signals have been widely used as identifying labels. As the labels can be permutated among the wavelengths or code dimensions, the structure of a label stack can be varied. However, studies on the relationship between label stacking scenarios and network performance are limited. To investigate this issue, we propose three label stacking models: sequential code distribution; sequential wavelength distribution, and random label distribution. The simulation results show that the sequential wavelength assignment, wherein the labels are uniformly distributed among the wavelengths, exhibits the best system performance in terms of the label-error rate (LER).

Hybrid wavelength conversion with dynamic pump-wavelength selection for synchronous optical packet switching

Abstract In this paper, we propose and study the performance of two conversion schemes for optical packet switch (OPS) with the objective of reducing packet contention, and hence, improving the overall packet loss ratio (PLR). In both schemes, conversion is performed in two layers. The first layer contains a dedicated wavelength converter (WC) for each input wavelength, while the second conversion layer consists of a dynamic pump-wavelength selection (DPS) of parametric wavelength converters (PWCs) shared in a Feed-Back fashion. The first scheme uses non-circular WCs (NCWCs) with DPS of chained PWCs (called LDPSC, for short). In the LDPSC scheme, an incoming packet is accepted only if it could be assigned to its required output using only the NCWCs in the first layer or more than one PWCs in a chained fashion in the second layer; otherwise it will be dropped. The second scheme uses circular wavelength conversion (CWCs) with DPS of PWCs without chaining (named CDPS, for short). In the CDPS scheme, an incoming packet is accepted only if it can be assigned to its required output using only CWCs in the first layer or only one of the PWCs in the second layer; otherwise it will be dropped. For each of the two schemes, a matching algorithm along with a Dynamic pump conversion (DPC) algorithm is presented. The performances of the proposed schemes are compared to two conventional schemes; namely, the DPSC scheme that uses a shared pool of DPS of chained PWCs, and the LDPS scheme that uses a dedicated NCWC for each input wavelength and a shared pool of DPS of PWCs without chaining. Simulation results show that, the new CDPS scheme achieves the same performance (in terms of PLR) of the LDPS scheme, while reducing the required number of PWCs by more than 44 % and 57 % for conversion distances (that is the maximum number of wavelengths adjacent to any input wavelength on any one of its sides), d = 1 and 2; respectively. Moreover, using conversion distance d = 1 , the new LDPSC scheme reduces the PLR compared to the LDPS scheme by up to 85 % and 22 % for traffic loads, 0.5 and 1; respectively. In addition, for d = 1 , the new LDPSC scheme reduces the PLR compared to the DPSC scheme by up to 99.8 % and 30 % for traffic loads, 0.5 and 1; respectively.

Multicasting in a High-Port Sub-$\mu$ sec Latency Hipo$\lambda$ aos Optical Packet Switch

We demonstrate a Hipoλaos optical packet switch (OPS) configuration that can support optical data multicasting in a sub-μsec 256-port layout. The proposed multicast OPS relies on the low-latency and high-port Hipoλaos OPS architecture, extending its capabilities from unito multicast operation through simple modifications at its egress wavelength-routing stage. It exploits the WDM cyclic properties of the arrayed waveguide grating router stage in conjunction with multi-wavelength conversion (WC) in semiconductor optical amplifier-Mach-Zehnder interferometer (SOA-MZI) toward scaling from a unito a multicast OPS design, with the level of multi-wavelength WC defining the multicast degree. Optical multicasting up to five nodes is experimentally demonstrated at 10 Gb/s, revealing error-free operation at 10 -9 with less than 2-dB power penalty compared with the unicast operation. The broadband operational characteristics of the multicast operation have been also investigated at 10 Gb/s, utilizing pairs of wavelengths with different channel spacing and validating the multicast credentials of the switch across all the necessary wavelength channels for a 256-port configuration. Finally, a mixed traffic scenario involving both unicast and multicast optical traffic is demonstrated at 10 Gb/s, revealing error-free performance for all cases, with a <;0.2 dB excess power penalty of multicasted packets over the unicast data.

AWG-based WDM ring networks: High-performance and low-cost system designs

Optical packet switching (OPS) ring networks are regarded as being promising solutions for future optical metropolitan area networks. In this paper, we propose novel arrayed-waveguide grating (AWG)-based wavelength division multiplexing (WDM) ring network systems that are aimed at achieving low-cost, scalable, and high-performance systems. By exploiting the cyclic routing properties and spatial wavelength reuse of a passive 2*2 AWG device, we redesign the ring network into a two-subring network, connected through the AWG device. The network is governed by a novel distributed medium access control (MAC) scheme, and we further derive a highly efficient and fair bandwidth allocation in accordance with a quota. A previously proposed experimental optical packet-switched WDM metro-ring network, called HOPSMAN, has shown to have exceptional performance among the OPS ring network studies. We show simulation and analytic results to demonstrate that our proposed system have achieved enhanced performance of the TT (tunable transmitter)-TR (tunable receiver)-based HOPSMAN by 1.5 times but have used only low-cost TT-FR (fixed-tuned receiver) or FT (fixed-tuned transmitter)-FR optical devices. Thus, the proposed networks achieve exceptional network throughput and also greatly reduce system costs.

Analysis of Hybrid Buffer Based Optical Data Center Switch

AbstractDue to the explosive growth in internet traffic, servers are facing bottleneck in speed and bandwidth requirements. To meet these ever increasing demands fiber optic technology can be used. This paper discusses, a hybrid buffer based optical packet switch design with its pros and cons. This hybrid buffer offers both electronic and optical buffering. This paper investigates the packet loss rate (PLR) performance of hybrid buffer under variety of conditions. The usage of buffer under different loading conditions has discussed. The sharing of electronic and optical buffers with respect to arrival of traffic is simulated, and delay ratio of electronic to optical buffering is computed. Average energy consumption at various loading condition is evaluated. The performance of switch is also investigated in a hypothetical five nodes network, and PLR is obtained under four different combinations of buffering and deflection of contending packets.

Recirculating Buffer and Arrayed Waveguide Grating-Based Switch for Optical Data Centers

Abstract Recently, an explosive growth has been observed in data traffic, which restricts the speed of exiting data communication network. To solve such problem, fiber optical-based optical communication is preferred choice. In this paper, a switch design is proposed to tackle contention problem in Optical Packet Switching (OPS) and data center networks. The proposed switch is compared with recently proposed switch design and advantages of proposed design are presented. To show superiority of the proposed design over earlier one, mathematical modeling is presented and expressions for noise figures and Bit Error Rate (BER) are derived. Using the mathematical framework, switch analysis is done in terms of BER and packet loss probability, and it has been found that power required for correct operation of the switch is lesser in comparison to earlier one. The effect of recirculation count on packet loss probability is detailed. The network layer analysis in terms of packet loss probability is also done using Monte Carlo simulation, and relation between physical and network layer parameters is also discussed.

High-Port and Low-Latency Optical Switches for Disaggregated Data Centers: The Hipoλaos Switch Architecture [Invited

The emergence of disaggregation in data center (DC) architectures as a way to increase resource utilization introduces significant challenges to the DC switching infrastructure, which has to ensure high bandwidth and low-latency communication along with high-radix connectivity. This paper examines the network requirements in disaggregated systems and reviews the credentials of state-of-the-art high-radix optical switch architectures. We also demonstrate a novel optical packet switch design, called Hipoλaos, that satisfies these requirements by combining N-port broadcast-and-select and N × N arrayed waveguide grating router-based forwarding schemes in N2-port connectivity configurations. The Hipoλaos switch provides sub-μs latency and high throughput performance by utilizing distributed control and optical feed-forward buffering in a modified Spanke architecture. Feasibility of a 256 port Hipoλaos layout with a four-packet buffering stage is experimentally demonstrated at 10 Gb/s, revealing error-free performance with a mean power penalty value of 2.19 dB. Simulation analysis is carried out for a 256 node system and eight traffic profiles; this analysis reveals a low latency value of only 605 ns, with throughput reaching 85% even when employing only two-packet buffers. Finally, the architecture’s scalability up to 1024 × 1024 designs is discussed, along with a power consumption analysis and a roadmap toward an integrated version of the switch.