Future Data Center Networks Research Articles

End-to-End 1024-Port Optical Packet Switching With 25 Gb/s Burst-Mode Reception for Data Centers

Despite the fact that Optical Packet Switching (OPS) emerges as a promising solution for future Data Center (DC) networks, towards increasing capacity and radix, while retaining sub-μs latency performance, the requirement for ultra-fast burst-mode reception has been a serious restraining factor. We attempt to overcome this limitation and demonstrate, for the first time to our knowledge, an end-to-end optical packet switch link through the 1024-port 25.6 Tb/s Hipoλaos OPS, featuring burst-mode reception with <; 50 ns locking time. The switch performance for unicast traffic is evaluated via Bit-Error-Rate measurements and error-free performance at 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">9</sup> is reported for all validated port-combinations, with a mean power penalty of 2.88 dB. Moreover, multicast flows from two different ports of the switch were successfully received validating the architecture's credentials for efficient multicast packet delivery. Taking one step further towards a realistic evaluation of an OPS-enabled DC, a simulation analysis was conducted, proving that low-latency performance, including the burst-mode reception time-overhead, can be successfully realized in a Hipoλaos-switched DC with up-to 100% throughput for a variety of traffic profiles.

Development of an Epitaxial Growth Technique Using III-V on a Si Platform for Heterogeneous Integration of Membrane Photonic Devices on Si

The rapid increase in total transmission capacity within and between data centers requires the construction of low-cost, high-capacity optical transmitters. Since a tremendous number of transmitters are required, photonic integrated circuits (PICs) using Si photonics technology enabling the integration of various functional devices on a single chip is a promising solution. A limitation of a Si-based PIC is the lack of an efficient light source due to the indirect bandgap of Si; therefore, hybrid integration technology of III-V semiconductor lasers on Si is desirable. The major challenges are that heterogeneous integration of III-V materials on Si induces the formation of dislocation at high process temperature; thus, the epitaxial regrowth process is difficult to apply. This paper reviews the evaluations conducted on our epitaxial growth technique using a directly bonded III-V membrane layer on a Si substrate. This technique enables epitaxial growth without the fundamental difficulties associated with lattice mismatch or anti-phase boundaries. In addition, crystal degradation correlating with the difference in thermal expansion is eliminated by keeping the total III-V layer thickness thinner than ~350 nm. As a result, various III-V photonic-device-fabrication technologies, such as buried regrowth, butt-joint regrowth, and selective area growth, can be applicable on the Si-photonics platform. We demonstrated the growth of indium-gallium-aluminum arsenide (InGaAlAs) multi-quantum wells (MQWs) and fabrication of lasers that exhibit &gt;25 Gbit/s direct modulation with low energy cost. In addition, selective-area growth that enables the full O-band bandgap control of the MQW layer over the 150-nm range was demonstrated. We also fabricated indium-gallium-arsenide phosphide (InGaAsP) based phase modulators integrated with a distributed feedback laser. Therefore, the directly bonded III-V-on-Si substrate platform paves the way to manufacturing hybrid PICs for future data-center networks.

Multiwavelength membrane laser array using selective area growth on directly bonded InP on SiO2/Si

The cost and power consumption of optical transmitters are now hampering further increases in total transmission capacity within and between data centers. Photonic integrated circuits (PICs) based on silicon (Si) photonics with wavelength-division multiplexing (WDM) technologies are promising solutions. However, due to the inefficient light emission characteristics of Si, incorporating III-V compound semiconductor lasers into PICs via a heterogeneous integration scheme is desirable. In addition, optimizing the bandgap of the III-V material used for each laser in a WDM transmitter becomes important because of recent strict requirements for optical transmitters in terms of data speed and operating temperature. Given these circumstances, applying a direct-bonding scheme is very difficult because it requires multiple bonding steps to bond different-bandgap III-V materials that are individually grown on different wafers. Here, to achieve wideband WDM operation with a single wafer, we employ a selective area growth technique that allows us to control the bandgap of multi-quantum wells (MQWs) on a thin InP layer directly bonded to silicon (InP-on-insulator). The InP-on-insulator platform allows for epitaxial growth without the fundamental difficulties associated with lattice mismatch or antiphase boundaries. High crystal quality is achieved by keeping the total III-V layer thickness less than the critical thickness (430 nm) and compensating for the thermally induced strain in the MQWs. By carrying out one selective MQW growth, we successfully fabricated an eight-channel directly modulated membrane laser array with lasing wavelengths ranging from 1272.3 to 1310.5 nm. The fabricated lasers were directly modulated at 56-Gbit/s with pulse amplitude modulation with four-amplitude-level signal. This heterogeneous integration approach paves the way to establishing III-V/Si WDM-PICs for future data-center networks.

C-band 56 Gbit/s on/off keying system over a 100 km dispersion-uncompensated link using only receiver-side digital signal processing.

In this Letter, we present a record C-band 56 Gbit/s intensity-modulation/direct-detection optical on/off keying (OOK) system over a 100 km dispersion-uncompensated link using only the receiver-side digital signal processing (DSP). The proposed DSP mainly includes an adaptive moment estimation (Adam)-based polynomial nonlinear equalizer (PNLE), autoregression (AR)-based post filter, and maximum likelihood sequence estimation (MLSE). Due to square-law detection, chromatic dispersion induces 11 nulls on the 28 GHz spectrum of the 56 Gbit/s OOK signal after the 100 km standard-single-mode-fiber transmission. Adam-based PNLE eliminates major linear and nonlinear distortions, but it cannot compensate the nulls. After the Adam-based PNLE, the AR-based post filter and MLSE further deal with the inter-symbol interference caused by the nulls to improve the system performance. The proposed C-band 56 Gbit/s OOK system shows great potential for future metro networks and data center networks.

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.

Photonic-Frame-Based TCP proxy architecture in optically interconnected data center networks

Optical switching is a good candidate for achieving the key requirements of high scalability, low latency and low power consumption in future large-scale data center networks (DCNs). However, circuit-based optical switching has a disadvantage of a long reconfiguration time. Packet-based optical switching also faces challenges such as the complexity of optical header processing and the difficulty of buffering in optical switches to resolve contention. To address these problems, we present a packet-switched optical network architecture design with a photonic frame wrapper that can generate variable-size photonic frames for optically interconnected intra-DCN. The proposed architecture forms hybrid DCN architecture consisting of a flattened optical switch layer for the data path and a conventional electrical switch layer for the control path. The photonic frame wrapper can increase bandwidth utilization by minimizing guard-time insertion in the optical switch domain. We also introduce a photonic frame wrapper with proxy solution to fully exploit the gain of the proposed architecture even in inter-DCN communication with long delays, which is compliant with standard TCP instances at end nodes. In intra-DCN, the average queuing delay is analyzed to find the efficient combinations between the guard time and time-slot length for the variable-sized photonic frame. We also analyzed the bandwidth overhead of guard time, to show it is comparable to that of Ethernet based DCN. In inter-DCN, the proxy solution demonstrates a noticeable TCP performance enhancement compared with the photonic frame wrapper in terms of bandwidth efficiency and end-to-end delay by theoretical and OPNET simulation analyses.

$\chi ^{(2)}$ Modulator With 40-GHz Modulation Utilizing BaTiO3 Photonic Crystal Waveguides

Future telecommunication and data center networks as well as quantum optical communication systems will require optical modulators with wide bandwidths, large extinction, low operating voltage, and small size. We report the first quantitative demonstration of slow light enhancement of the electro-optic (EO) coefficient in a χ (2) ferroelectric waveguide at microwave modulation frequencies. This is demonstrated in a compact (1 mm) photonic crystal (PC) device with a voltage-length product (Vπ·L) of 0.66 V-cm at 10 GHz and measured EO modulation out to 40 GHz. A local enhancement factor of 12 and effective EO coefficient of 900 pm/V is measured in the PC region at a wavelength of 1530 nm. By further optimizing the PC structure, devices with EO 3-dB bandwidths greater than 40 GHz and voltage-length product of 0.16 V-cm are predicted with 100-μm interaction length.

Resource Allocation in Electrical/Optical Hybrid Switching Data Center Networks

Hybrid switching combines the strengths of electronic packet switching and high-speed optical circuit switching and is thus widely believed to be a cost-effective solution for current and future data center networks (DCNs). In designing a hybrid switching DCN, it is essential to know how many resources should be devoted to optical circuit switching and electronic packet switching so the performance requirements in both planes can be satisfied, and, at the same time, objectives such as minimizing the overall system building cost or the power consumption can be realized. In this paper, we introduce BLOC, the Blocking LOss Curve, as a tool to characterize hybrid switching DCNs. With BLOC, the three interacting components in hybrid switching systems-resource allocation, traffic partitioning, and system performance-can be naturally integrated and studied together. We show how BLOC can be used to obtain feasible resource allocation/traffic partition combinations. We also show how system parameters, such as flow arrival characteristics, may affect resource allocation and system costs. Our study reveals that the total amount of resource required to satisfy 95th percentile traffic can be as low as 60% of that required for 99.9th percentile traffic.

A Scalable, Partially Configurable Optical Switch for Data Center Networks

Optical circuit switching may be instrumental in meeting the cost, energy, and aggregate bandwidth requirements of future data center networks. However, conventional MEMS beam-steering cross-connects cannot provide submillisecond switching with the port count necessary for data centers. Here, we investigate a novel noncrossbar selector switch architecture and pupil-division switching layout to improve optical switching performance by relaxing the requirement of arbitrary switch configurability. This architecture and switch design enable MEMS beam-steering micromirrors to scale to microsecond response speeds while supporting high port count and low loss switching, and can realize a number of useful interconnection topologies. We present the design, fabrication, and experimental characterization of a proof-of-principle prototype using a single comb-driven MEMS mirror to achieve 150 μs switching of 61 ports between four preprogrammed interconnection mappings. We demonstrate the scalability of this switch architecture with a detailed optical design of a 2048-port selector switch with 20 μs switching time.

Elastic optical ring with flexible spectrum ROADMs: An optical switching architecture for future data center networks

Data center networks (DCN) are facing significant increase in both scale and bandwidth requirement, and various intra-DCN optical circuit switching architectures have been proposed to handle the inter-rack traffic. As the size of data centers are becoming larger, ring based architectures have received considerable interests, because a ring can provide connectivity to a large number of racks that are distributed far apart in a data center. However, due to the specific data center traffic pattern, fixed spectrum allocation scheme in existing ring based DCNs will lead to low spectrum utilization and hence low system throughput. In this paper, we propose to introduce flexible spectrum allocation into ring based intra-DCN optical switching, so that spectrum utilization can be improved. We design an elastic ring intra-DCN optical circuit switching architecture with flexible spectrum reconfigurable optical add-drop multiplexers (FS-ROADMs). Adopting flexible spectrum allocation in elastic ring, a spectrum allocation problem which is essentially a very complex yet interesting optimization problem arises. For a given set of inter-rack traffic requests, we develop an integer linear program formulation as well as a heuristic algorithm to solve the flexible spectrum allocation problem. We study how the inter-rack traffic pattern may affect the system performances in both fixed and elastic ring cases. For elastic ring, we also analyze and evaluate the computational complexity in solving the spectrum allocation problem. Our results demonstrate that, to handle uneven inter-rack traffic pattern in DCNs, elastic ring can remarkably improve system throughput, though its computational complexity of spectrum allocation problem is slightly higher than that of the fixed ring.

Optical interconnection networks for cloud data centers [Guest Editorial

The rise of Cloud Computing and Big Data has created the need for morepowerful warehouse-scale cloud data centers, which usually comprise hundreds of thousands ofservers with highaccess bandwidth (>10Gb/s) or more, and cost-effective intra- and inter- data center networks(DCNs) to provide switching and interconnect among all servers. However, the current DCNs sufferfrom limited throughput, large latencies and highpower consumption, and will not be able to sustainthe projected future DCN traffic demandswithout consuming excessive amounts of power. Optical interconnection networks capable of supporting high throughput, high bandwidth, low latency, low power consumptionand full-bandwidth communication among all servers have been proposedas a promising solution by performing switchingat the optical domain, thus greatly improving the performance of the cloud data centers.

Exploiting Efficient and Scalable Shuffle Transfers in Future Data Center Networks

Distributed computing systems like MapReduce in data centers transfer massive amount of data across successive processing stages. Such shuffle transfers contribute most of the network traffic and make the network bandwidth become a bottleneck. In many commonly used workloads, data flows in such a transfer are highly correlated and aggregated at the receiver side. To lower down the network traffic and efficiently use the available network bandwidth, we propose to push the aggregation computation into the network and parallelize the shuffle and reduce phases. In this paper, we first examine the gain and feasibility of the in-network aggregation with BCube, a novel server-centric networking structure for future data centers. To exploit such a gain, we model the in-network aggregation problem that is NP-hard in BCube. We propose two approximate methods for building the efficient IRS-based incast aggregation tree and SRS-based shuffle aggregation subgraph, solely based on the labels of their members and the data center topology. We further design scalable forwarding schemes based on Bloom filters to implement in-network aggregation over massive concurrent shuffle transfers. Based on a prototype and large-scale simulations, we demonstrate that our approaches can significantly decrease the amount of network traffic and save the data center resources. Our approaches for BCube can be adapted to other servercentric network structures for future data centers after minimal modifications.

Smart grid technologies for future radio and data center networks

This article explores smart grid technologies that can be applied to a telecommunication network to achieve energy-efficient networking, autonomous operation, and adaptation to realtime electricity pricing schemes. With the fast penetration of renewable energy sources within base stations and data centers, the telecommunication operator can establish an active role in the energy market by adjusting power consumption in real time. In the telecommunications sector, energy management technologies have recently emerged with BS management schemes and virtual machine migration/allocation strategies. In the energy sector, smart grid technologies and new standards enable real-time management and pricing. The only brick missing is the orchestration of the technologies in the two sectors to enable smart telecommunications network operation in terms of energy consumption. In this article, concepts such as demand response, supply load control, and the model of the prosumer in the smart grid are correlated to the operation of modern radio and data center networks. The main outcome of the research is to provide new ideas for net zero service delivery and explore the role of the telecommunication provider in the energy market where dynamic electricity pricing is expected to hold a critical role in decision making processes.

Next-Generation Optically-Interconnected High-Performance Data Centers

The communication requirements imposed by the unabated growth in both the size and computational density of modern data centers will soon outpace the fundamental capabilities of conventional electronic interconnection networks. Optical interconnects represent a potentially disruptive technology that can simultaneously satisfy the throughput, latency, and energy demands of these next-generation systems. In this paper, we propose an end-to-end photonic networking platform for future optically-interconnected data center networks. A reconfigurable hybrid photonic network building block and a protocol-agnostic optical network interface comprise the key elements of our proposed system. A modular optical switch fabric prototyping platform is designed and implemented to support a wide variety of potential photonic technologies. We successfully demonstrate nanosecond-scale optical packet-switching at 1-Tb/s per port (40-Gb/s ×25λs) through our prototype, configured as a 4×4 switch, and confirm error-free transmission across all wavelengths. In order to flexibly allocate appropriate bandwidth for heterogeneous applications, we further implement and evaluate a novel wavelength-reconfigurable optical packet- and circuit-switch by utilizing our optical switch platform. Finally, we detail our implementation of an optical network interface card (O-NIC) designed to bridge the gap between the well-established protocol layers utilized in current networks and the physical layer of the proposed photonic networks and demonstrate its functionality by the streaming of high-definition video through the end-to-end optical network.

A Survey on Optical Interconnects for Data Centers

Data centers are experiencing an exponential increase in the amount of network traffic that they have to sustain due to cloud computing and several emerging web applications. To face this network load, large data centers are required with thousands of servers interconnected with high bandwidth switches. Current data center networks, based on electronic packet switches, consume excessive power to handle the increased communication bandwidth of emerging applications. Optical interconnects have gained attention recently as a promising solution offering high throughput, low latency and reduced energy consumption compared to current networks based on commodity switches. This paper presents a thorough survey on optical interconnects for next generation data center networks. Furthermore, the paper provides a qualitative categorization and comparison of the proposed schemes based on their main features such as connectivity and scalability. Finally, the paper discusses the cost and the power consumption of these schemes that are of primary importance in the future data center networks.

A hybrid optical packet and wavelength selective switching platform for high-performance data center networks

We propose and experimentally demonstrate for the first time a hybrid optical packet and wavelength selective switching platform for high-performance data center networks. This architecture based on cascaded silicon microrings and semiconductor optical amplifiers (SOAs) supports wavelength reconfigurable packet and circuit switching, and is highly scalable, energy efficient and potentially integratable. By combining the wavelength-selective behavior of the microring and the broadband behavior of the SOA switch, we are able to achieve fast switching transitions, high extinction ratios, and low driving voltages, which are all requirements for future optical high-performance data center networks. Routing correctness and error-free operation (<10(-12)) are verified for both 10-Gb/s and 40-Gb/s packets and streaming data with format transparency.