Data Center Interconnection Networks Research Articles

A Bidirectional Wavelength Division Multiplexed (WDM) Free Space Optical Communication (FSO) System for Deployment in Data Center Networks (DCNs).

Data centers are crucial to the growth of cloud computing. Next-generation data center networks (DCNs) will rely heavily on optical technology. Here, we have investigated a bidirectional wavelength-division-multiplexed (WDM) free space optical communication (FSO) system for deployment in optical wireless DCNs. The system was evaluated for symmetric 10 Gbps 16-quadrature amplitude modulation (16-QAM) intensity-modulated orthogonal frequency-division multiplexing (OFDM) downstream signals and 10 Gbps on-off keying (OOK) upstream signals, respectively. The transmission of optical signals over an FSO link is demonstrated using a gamma-gamma channel model. According to the bit error rate (BER) results obtained for each WDM signal, the bidirectional WDM-FSO transmission could achieve 320 Gbps over 1000 m free space transmission length. The results show that the proposed FSO topology offers an excellent alternative to fiber-based optical interconnects in DCNs, allowing for high data rate bidirectional transmission.

Competition analysis in data center interconnection networks using the Lotka-Volterra model

Competition analysis in data center interconnection networks using the Lotka-Volterra model

Variable Expanding Structure for Data Center Interconnection Networks

Data centers are fundamental facilities that support high-performance computing and large-scale data processing. To guarantee that a data center can provide excellent properties of expanding and routing, the interconnection network of a data center should be designed elaborately. Herein, we propose a novel structure for the interconnection network of data centers that can be expanded with a variable coefficient, also known as a variable expanding structure (VES). A VES is designed in a hierarchical manner and built iteratively. A VES can include hundreds of thousands and millions of servers with only a few layers. Meanwhile, a VES has an extremely short diameter, which implies better performance on routing between every pair of servers. Furthermore, we design an address space for the servers and switches in a VES. In addition, we propose a construction algorithm and routing algorithm associated with the address space. The results and analysis of simulations verify that the expanding rate of a VES depends on three factors:n,m, andkwhere thenis the number of ports on a switch, themis the expanding speed and thekis the number of layers. However, the factormyields the optimal effect. Hence, a VES can be designed with factormto achieve the expected expanding rate and server scale based on the initial planning objectives.

POID: a passive all-optical inter-rack interconnect for data-centers

With the increase in demand for bandwidth-thirsty cloud-based services, bandwidth demands for rack-to-rack interconnect in data centers are increasing exponentially. In this paper, an optical interconnect architecture dubbed “POID” is proposed. The novel architecture is proposed using passive optical technology, wavelength division multiplexing, and innovative architectural concepts. The proposed architecture adequately supports the rack-to-rack traffic with latencies as low as few hundred nanoseconds while providing Tera bits per second throughput. The performance of the proposed architecture is evaluated through simulation studies and compared to other recent related data center interconnection networks from recent literature. The proposed POID architecture provides up to 10% improvement in throughput and up to about 82% reduction in latency compared to other comparative designs of data center interconnects using current optical fiber and switching technologies. The proposed PIOD incurs 22% power penalty compared to the existing best performing comparable designs.

Silicon-Organic Hybrid (SOH) Mach-Zehnder Modulators for 100 Gbit/s on-off Keying

Electro-optic modulators for high-speed on-off keying (OOK) are key components of short- and medium-reach interconnects in data-center networks. Small footprint, cost-efficient large-scale production, small drive voltages and ultra-low power consumption are of paramount importance for such devices. Here we demonstrate that the concept of silicon-organic hybrid (SOH) integration perfectly meets these challenges. The approach combines the unique processing advantages of large-scale silicon photonics with unrivalled electro-optic (EO) coefficients obtained by molecular engineering of organic materials. Our proof-of-concept experiments demonstrate generation and transmission of OOK signals at line rates of up to 100 Gbit/s using a 1.1 mm-long SOH Mach-Zehnder modulator (MZM) featuring a π-voltage of only 0.9 V. The experiment represents the first demonstration of 100 Gbit/s OOK on the silicon photonic platform, featuring the lowest drive voltage and energy consumption ever demonstrated for a semiconductor-based device at this data rate. We support our results by a theoretical analysis showing that the nonlinear transfer characteristic of the MZM can help to overcome bandwidth limitations of the modulator and the electric driver circuitry. We expect that high-speed, power-efficient SOH modulators may have transformative impact on short-reach networks, enabling compact transceivers with unprecedented efficiency, thus building the base of future interfaces with Tbit/s data rates.

Supporting Dynamic Bandwidth Adjustment Based on Virtual Transport Link in Software-Defined IP Over Optical Networks

Packet and optical networks are managed independently in both current carrier networks and datacenter interconnection networks, which results in a lack of service agility and is a significant source of capital expenditure and operating expense overhead for the network operator. In our previous works, we have proposed and experimentally demonstrated a software-defined networking- based architecture (service-aware architecture) for converged control of packet and optical networks. In this paper, we first propose a concept called virtual transport link (VTL) to support dynamic link aggregation in software- defined IP over optical networks. Then, the VTL-based dynamic bandwidth adjustment (VTL-DBA) method is proposed for improving the resource utilization and for simplifying service provisioning. The feasibility of our proposals is experimentally verified on our software-defined IP over optical networks testbed. The performance of the VTL-DBA method is also evaluated based on our simulation platform in terms of resource usage, resource utilization, and blocking probability. For carrying quality guaranteed services, VTL-DBA saves 25% of transport resources on average compared with static bandwidth allocation methods, and it shows an increase of 15% in the average resource utilization. For carrying both quality guaranteed services and quality tolerant services, VTL-DBA reduces the service blocking probability to only 10% on average with higher resource utilization than static bandwidth allocation methods.

Scalability analysis methodology for passive optical interconnects in data center networks using PAM

A framework is developed for modeling the fundamental impairments in optical datacenter interconnects, i.e., the power loss and the receiver noises. This framework makes it possible, to analyze the trade-offs between data rates, modulation order, and number of ports that can be supported in optical interconnect architectures, while guaranteeing that the required signal-to-noise ratios are satisfied. To the best of our knowledge, this important assessment methodology is not yet available. As a case study, the trade-offs are investigated for three coupler-based top-of-rack interconnect architectures, which suffer from serious insertion loss. The results show that using single-port transceivers with 10 GHz bandwidth, avalanche photodiode detectors, and quadratical pulse amplitude modulation, more than 500 ports can be supported.

Performance analysis of optical interconnects’ architectures for data center networks

Electrical interconnects in Data Center Networks (DCNs) suffer from various problems which include high energy consumption, high latency, fixed throughput of links and limited reconfigurability. In...

Data center interconnection networks are not hyperbolic

Topologies for data center interconnection networks have been proposed in the literature through various graph classes and operations. A common trait to most existing designs is that they enhance the symmetric properties of the underlying graphs. Indeed, symmetry is a desirable property for interconnection networks because it minimizes congestion problems and it allows each entity to run the same routing protocol. However, despite sharing similarities these topologies all come with their own routing protocol. Recently, generic routing schemes have been introduced which can be implemented for any interconnection network. The performances of such universal routing schemes are intimately related to the hyperbolicity of the topology. Roughly, graph hyperbolicity is a metric parameter which measures how close is the shortest-path metric of a graph from a tree metric (the smaller the gap the better). Motivated by the good performances in practice of these new routing schemes, we propose the first general study of the hyperbolicity of data center interconnection networks. Our findings are disappointingly negative: we prove that the hyperbolicity of most data center interconnection topologies scales linearly with their diameter, that is the worst-case possible for hyperbolicity. To obtain these results, we introduce original connection between hyperbolicity and the properties of the endomorphism monoid of a graph. In particular, our results extend to all vertex and edge-transitive graphs. Additional results are obtained for de Bruijn and Kautz graphs, grid-like graphs and networks from the so-called Cayley model.

Performance evaluation of hybrid optical switch architecture for data center networks

In response to the need for high bandwidth and power efficient data center interconnection networks, different interconnects have been proposed based on the optical technology used: micro-electromechanical system (MEMS), optical cross connects (OXCs), arrayed waveguide grating routers (AWGRs) and semiconductor optical amplifier (SOAs). MEMS switches are based on mature technology, have low insertion loss and cross-talk, and are data rate independent. They are also the most scalable and the cheapest class of optical switches. However, the reconfiguration time of these switches is of the order of tens of milliseconds while fast optical switches have switching time in the range of a few nanoseconds. Fast optical switches can be based on AWGRs in conjunction with tunable wavelength converters or tunable lasers or they are based on SOAs in broadcast-and-select architecture. In this paper, we propose an optical interconnect architecture for the large scale data centers. The proposed interconnect: Hybrid Optical Switch Architecture (HOSA) is a hybrid design that features slow and fast optical switches. The hybrid design leverages strengths of both types of optical switches. To reduce complexity, we employ a single stage core topology that can be easily scaled up (in capacity) and scaled out (in the number of racks) without requiring major re-cabling and network reconfiguration. We investigate the scalability of the HOSA and show that by using a single stage core topology, it can be scaled to a hundreds of thousands of servers. We also investigate a trade-off between cost and power consumption of our design by comparing it with other well-known interconnects by using analytical modelling. We demonstrate power efficiency as compared to other conventional interconnects on account of upfront CAPEX but the additional CAPEX incurred in deploying our solution instead of traditional architecture is mitigated to some extent by reduced OPEX, due to its greater energy efficiency. We evaluate the performance of the system using network-level simulation by considering diverse workload communication patterns and system design parameters. Our results show low latency and high throughput with different workload communication patterns.

Blocking probability in optical interconnects in data center networks

Cloud computing and Web-based applications are creating a need for powerful data centers. Data centers have a great need for high bandwidth, low latency, low blocking probability, and low bit-error rate to sustain the interaction between different applications. Current data center networks (DCNs) suffer from several problems such as high-energy consumption, high latency, fixed throughput of links, and limited reconfigurability. Electronic switches are low radix and have high latency due to a large hop count since each hop employs a store-and-forward mechanism. Optical interconnects, on the other hand, offer several advantages such as low-energy consumption, high bandwidth, reconfigurability, malleability to changing traffic, high-radix switch design, fast switching transition times, and wavelength multiplexing. These benefits provide the incentive to shift from electrical interconnects to optical interconnects in DCNs. Despite several advantages over their electrical counterparts, the performance of optical interconnects can be further improved by considering some performance parameters of optical interconnects. One such important parameter for the performance of any communication network is the blocking probability. This paper makes a comprehensive investigation of the performance of optical interconnects in different DCN architectures on the basis of blocking probability and concludes by suggesting ways to reduce the blocking.

Performance Evaluation of a Hybrid Optical/Electrical Interconnect

In response to the need for higher-speed and affordable-cost datacenter interconnection networks, hybrid optical/electrical architectures have been proposed. Even still, a number of implementation issues remain open and little is known about the performance of real applications over such networks. To fill this gap, we present a hybrid network architecture, called HydRA, comprising commodity off-the-shelf equipment, calculate its total price using current list prices, and show its price competitiveness against fat-tree alternatives. We also report on a prototype implementation of our HydRA network that uses Ethernet switches and a custom-built software controller. We show that our HydRA network prototype manages to accelerate the execution of real parallel workloads, when compared to equal-cost electronic-only fat-tree networks.

Analytical model of quality of service scheduling for optical aggregation in data centers

Optical switching technologies represent a promising solution for data center interconnection networks to support the increasing bandwidth requirements of current cloud-based applications, while reducing interconnection complexity and energy consumption. Furthermore, the heterogeneity of intra- and inter-data center traffic characteristics requires some form of quality of service management. This paper describes modeling and design aspects of data center optical interconnections with particular emphasis on the aggregation level, where hybrid switching and packet scheduling are jointly applied to effectively implement service differentiation. Priority scheduling of three different service profiles is applied to maximize intra- and inter-data center traffic throughput, while guaranteeing time transparency for delay-sensitive services and zero loss/fixed delay for guaranteed connections. An analytical model is defined and validated to assess loss of real time and throughput of best effort traffic, in asynchronous packet context, when considering best effort traffic saturating the channels of the optical link. The model can also be used to dimension the optical output interface of the aggregation level switch.

OSA: An Optical Switching Architecture for Data Center Networks With Unprecedented Flexibility

A detailed examination of evolving traffic characteristics, operator requirements, and network technology trends suggests a move away from nonblocking interconnects in data center networks (DCNs). As a result, recent efforts have advocated oversubscribed networks with the capability to adapt to traffic requirements on-demand. In this paper, we present the design, implementation, and evaluation of OSA, a novel Optical Switching Architecture for DCNs. Leveraging runtime reconfigurable optical devices, OSA dynamically changes its topology and link capacities, thereby achieving unprecedented flexibility to adapt to dynamic traffic patterns. Extensive analytical simulations using both real and synthetic traffic patterns demonstrate that OSA can deliver high bisection bandwidth (60%-100% of the nonblocking architecture). Implementation and evaluation of a small-scale functional prototype further demonstrate the feasibility of OSA.