Design, implementation, and characterization of a hybrid optical interconnect for a four-stage free-space optical backplane demonstrator

A four-stage unidirectional ring free-space optical interconnect system was designed, analyzed, implemented, and characterized. The optical system was used within a complementary metal-oxide semiconductor-self-electro-optic-effect-device-based optical backplane demonstrator that was designed to fit into a standard VME chassis. This optical interconnect was a hybrid microlens-macrolens system, in which the microlens relays were arranged in a maximum lens-to-waist configuration to route the optical beams from the optical power supply to the transceiver arrays, while the macrolens optical relays were arranged in a telecentric configuration to route optical signal beams from stage to stage. The following aspects of the optical system design are discussed: the optical parameters for the hybrid optical system, the image mapping of the two-dimensional array of optical beams from stage to stage, the alignment tolerance of the hybrid relay system, and the power budget of the overall optical interconnect. The implementation of the optical system, including the characterization of optical components, subsystem prealignment, and final system assembly, is presented. The two-dimensional array of beams for the stage-to-stage interconnect was adjusted with a rotational error of <0.05 degrees and a lateral offset error of <3.5 mum. The measured throughput is in good agreement with the lower-bound predictions obtained in the theoretical results, with an optical power throughput of -20.2 dB from the fiber input of the optical power supply to the modulator array and -25.5 dB from the fiber input to the detector plane.     

Design methods for space-variant optical interconnections to achieve optimum power throughput

Optoelectronic systems based on space-variant optics give great freedom to the system designer in terms of interconnect topologies. One feature of space-variant systems is that they can achieve a high interconnect density. However, this density is achieved by having large arrays of diffractive elements with very small apertures relative to the propagation distances involved. Thus diffraction losses from the finite apertures can significantly affect power throughput for these types of systems, regardless of the diffractive efficiencies of the optical elements involved. Therefore it is desirable that this loss be minimized. We present several space-variant optical interconnect design methods (for both one-to-one and fan-out interconnects) and compare them in terms of power throughput for diffraction-limited interconnect distances. Both numerical simulations and experimental results are presented.     

Distributed crossbar interconnects with vertical-cavity surface-emitting laser-angle multiplexing and fiber image guides

We report on the implementation of an optical crossbar interconnectconsisting of a centralized free-space beam-steering subsystem, adistributed array of vertical-cavity surface-emitting lasers andphotoreceivers, a fiber image guide, and a large-core polymer fiberarray. The interconnect can in principle handle more than 350cross-bar channels, but our implementation demonstrated only 240channels. Transmissions of 500-Mbit/s per channel weredemonstrated. Approximately 12.7-dB end-to-end optical channelattenuation was measured. Details of component design, packaging, system integration, and testing are presented and discussed.     

Three-dimensional optoelectronic stacked processor by use of free-space optical interconnection and three-dimensional VLSI chip stacks

We present a demonstration system under the three-dimensional (3D) optoelectronic stacked processor consortium. The processor combines the advantages of optics in global, high-density, high-speed parallel interconnections with the density and computational power of 3D chip stacks. In particular, a compact and scalable optoelectronic switching system with a high bandwidth is designed. The system consists of three silicon chip stacks, each integrated with a single vertical-cavity-surface-emitting-laser-metal-semiconductor-metal detector array and an optical interconnection module. Any input signal at one end stack can be switched through the central crossbar stack to any output channel on the opposite end stack. The crossbar bandwidth is designed to be 256 Gb/s. For the free-space optical interconnection, a novel folded hybrid micro-macro optical system with a concave reflection mirror has been designed. The optics module can provide a high resolution, a large field of view, a high link efficiency, and low optical cross talk. It is also symmetric and modular. Off-the-shelf macro-optical components are used. The concave reflection mirror can significantly improve the image quality and tolerate a large misalignment of the optical components, and it can also compensate for the lateral shift of the chip stacks. Scaling of the macrolens can be used to adjust the interconnection length between the chip stacks or make the system more compact. The components are easy to align, and only passive alignment is required. Optics and electronics are separated until the final assembly step, and the optomechanic module can be removed and replaced. By use of 3D chip stacks, commercially available optical components, and simple passive packaging techniques, it is possible to achieve a high-performance optoelectronic switching system.     

Design considerations and algorithms for partitioning optoelectronic multichip modules

There is considerable interest in the development of optical interconnects for multichip modules (MCM's) to improve their performance. For effective utilization of the optical and electronic technologies, a methodology for partitioning the system is required. The key question to be answered is which technology should be used for each interconnect in a given netlist: optical or electronic. We introduce the computer-aided design approach for partitioning optoelectronic systems into optoelectronic MCM's. We first discuss the design trade-off issues in an optoelectronic system design, including speed, power dissipation, area, and diffraction limits for free-space optics. We then define a formulation for optoelectronic MCM partitioning and describe new algorithms for optimizing this partitioning based on the minimization of the power dissipation. The models for the algorithms are discussed in detail, and an example of a multistage interconnect network is given. Different results, with the number and size of chips being variable, are presented in which improvement for the system packaging has been observed when the partitioning algorithms are applied.     

All-optical crossbar switch using wavelength division multiplexing and vertical-cavity surface-emitting lasers

A design for an all-optical crossbar network utilizing wavelength-tunable vertical-cavity surface-emitting laser (VCSEL) technology and a combination of free-space optics and compact optical waveguides is presented. Polymer waveguides route the optical signals from a spatially distributed array of processors to a central free-space optical crossbar, producing a passive, all-optical, fully connected crossbar network directly from processor to processor. The analyzed network could, relatively inexpensively, connect local clusters of tightly integrated processors. In addition, it is also believed that such a network could be extended, with wavelength reuse, to connect much larger numbers of processors in a multicluster network.     

Free-space optical crossbar network integrated in a single block of LiNbO3 crystal

A free-space optical crossbar network integrated in a single block of LiNbO(3) crystal is proposed, which consists of stages of 2×2 switches making use of the electro-optic effect of crystal and in-between routing devices for permutation based on double refraction and internal double reflection on interfaces. Two basic configurations are suggested. A control algorithm for the crossbar network is discussed, which may control a nonblocking interconnection between any input and output. The integrated crossbar network is low energy loss, nonblocking, easy to assemble, and insensitive to environment. A 3×3 crossbar network is designed and the experiment is demonstrated.     

Optical smart-pixel-based Clos crossbar switch

We present the design of an optically interconnected Clos crossbar switch that uses three smart-pixel devices. This optical Clos architecture is also well matched to a space-wavelength switch that arbitrarily permutes data streams between wavelength-division multiplexed channels on an array of fibers. We have designed a hybrid complementary metal-oxide semiconductor-self-electro-optic device (CMOS-SEED) crossbar smart-pixel array for use in a 16-channel optical Clos switch. The crossbar devices also have an 8 x 8 array of multiple-quantum-well diodes that can be configured electrically as modulators with eight bit planes of randomly addressable local memory or as receivers with adjustable gain and threshold. We show that the current hybrid-SEED technology should support a 1024-channel Clos switch operating at 500 Mbits/s per channel if pixel power consumption can be reduced.     

Design and analysis of an adaptive board-to-board dynamic holographic interconnect

We describe the design and analysis of an adaptive free-space optical interconnect between two circuit boards in a standard electronic backplane. An array of vertical-cavity surface-emitting lasers is used as the transmitter, and this communicates with a detector array on the receiver circuit board. Routing is achieved with a holographic crossbar that has a ferroelectric liquid-crystal spatial light modulator to display binary phase computer-generated holograms. A detailed analysis of a 48-channel interconnect designed to operate at 1 (Gbytes/s)/channel indicates that such a switch will operate successfully given typical components and card misalignments.     

Design and characterization of a microchannel optical interconnect for optical backplanes

The design, modeling, and experimental characterization of a microchannel-based free-space optical interconnect is described. The microchannel interconnect was used to implement a representative portion of an optical backplane that was based on field-effect transistor, self-electro-optic device smart-pixel transceivers. Telecentric relays were used to form the optical interconnect, and two modes based on two different optical window clusterings were implemented. The optical system design, including the optical geometry for different degrees of clustering of windows supported by a lenslet relay and the image mapping associated with a free-space optical system, is described. A comparison of the optical beam properties at the device planes, including the spot size and power uniformity of the spot array, as well as the effects of clipping and misalignment for the different operating modes, is presented. In addition, the effects of beam clipping and misalignment for the different operating modes is presented. We show that microchannel free-space optical interconnects based on a window-clustering scheme significantly increase the connection density. A connection density of 2222 connections/cm(2) was achieved for this prototype system with 2 x 2 window clustering.     

Fiber-image-guide-based bit-parallel optical interconnects

We propose a scalable bit-parallel optical interconnect method for use in large-bandwidth interprocessor communications. Flexible fiber image guides are used to transmit spatially parallel optical data between a vertical-cavity surface-emitting laser array and a photodetector array. We have studied a lens-based and a fiber-image-taper-based input-output coupling scheme and have modeled power-loss mechanisms and resolution-degradation mechanisms associated with the schemes. We have also performed some experiments to confirm the operational principles of the proposed schemes and to investigate the power efficiency and imaging-resolution performance of the interconnect schemes. Our study indicates that the proposed interconnects can offer a scalable method to transmit hundreds of channels of multigigabyte per second per channel optical data in parallel.     

Field-programmable logic devices with optical input-output

A field-programmable logic device (FPLD) with optical I/O is described. FPLD's with optical I/O can have their functionality specified in the field by means of downloading a control-bit stream and can be used in a wide range of applications, such as optical signal processing, optical image processing, and optical interconnects. Our device implements six state-of-the-art dynamically programmable logic arrays (PLA's) on a 2 mm x 2 mm die. The devices were fabricated through the Lucent Technologies-Advanced Research Projects Agency-Consortium for Optical and Optoelectronic Technologies in Computing (Lucent/ARPA/COOP) workshop by use of 0.5-mum complementary metal-oxide semiconductor-self-electro-optic device technology and were delivered in 1998. All devices are fully functional: The electronic data paths have been verified at 200 MHz, and optical tests are pending. The device has been programmed to implement a two-stage optical switching network with six 4 x 4 crossbar switches, which can realize more than 190 x 10(6) unique programmable input-output permutations. The same device scaled to a 2 cm x 2 cm substrate could support as many as 4000 optical I/O and 1 Tbit/s of optical I/O bandwidth and offer fully programmable digital functionality with approximately 110,000 programmable logic gates. The proposed optoelectronic FPLD is also ideally suited to realizing dense, statically reconfigurable crossbar switches. We describe an attractive application area for such devices: a rearrangeable three-stage optical switch for a wide-area-network backbone, switching 1000 traffic streams at the OC-48 data rate and supporting several terabits of traffic.     

Performance-based adaptive power optimization for digital optical interconnects

Optical links are traditionally set to transmit maximum power for worst-case loss and consequently to dissipate more power than is required. We describe a technique to minimize power consumption based on the measured bit-error rate (BER) of the link. This technique uses a novel power-negotiation algorithm that optimizes the link power setting to achieve minimum power dissipation for a target BER. A 0.5 microm complementary metal-oxide semiconductor optical transceiver chip was fabricated, and a free-space optical interconnect system was built for validation. The results showed that the algorithm was able to find the optimum power settings for the VCSELs for a target BER and to account for dynamic changes such as variation in the optical loss in the system.     

Relay lens-free birefringence-customized stackable beam-steering modules for optical interconnects

A previously suggested birefringence-customized modular optical interconnect technique is extended for lens-free relay operation. Various lens-free relay imaging models are developed. We claim that the lens-free relay system is important in simplifying an optical interconnect system whenever the imaging conditions permit. To verify the validity of various proposed concepts, we experimentally implemented some 8 x 8 optical permutation modules. High-power efficiency and low channel cross talk were experimentally observed. In general, the larger the channel spacing, the less the cross talk. A quantitative cross-talk measurement of the lens-free relay system shows that, for a fixed channel width of 0.5 mm and channel spacings of 0.5, 1, and 2 mm, a less than -20-dB cross-talk performance can be guaranteed for lens-free relay distances of 40, 280, and 430 mm, respectively.     

High-speed free-space interconnect based on optical ring topology: experimental demonstration

The design and implementation of a high-speed optical-ring-topology-based free-space optical interconnect is described. This interconnect system operates at 500 MHz and consists of 16 laser transmitters, a four-channel free-space interconnect, and a high-speed receiver. A nearest-neighbor interconnect is demonstrated. At the data rate of 500 MHz the total system throughput is 8 Gbits/s. The system can easily be operated at much higher data rates since the rate is limited only by the electronic circuitry. A discussion is given about device issues such as optical switching devices, and practical system-design issues such as integration and interface with current electronic systems are considered. This interconnect is promising to the implementation of ultrafast massively parallel single-instruction multiple-data machines.     

Design, implementation, and characterization of an optical power supply spot-array generator for a four-stage free-space optical backplane

The design and implementation of a robust, scalable, and modular optical power supply spot-array generator for a modulator-based free-space optical backplane demonstrator is presented. Four arrays of 8 x 4 spots with 6.47-mum radii (at 1/e(2) points) pitched at 125 mum in the vertical direction and 250 mum in the horizontal were required to provide the light for the optical interconnect. Tight system tolerances demanded careful optical design, robust optomechanics, and effective alignment techniques. Issues such as spot-array generation, polarization, power efficiency, and power uniformity are discussed. Characterization results are presented.     

Tolerancing of board-level-free-space optical interconnects

For optical interconnects to become a mature technology they must be amenable to electronic packaging technology. Two main obstacles to including free-space optical interconnects are alignment and heat-dissipation issues. Here we study the issues of alignment tolerancing that are due to assembly and manufacturing variations (passive-element tolerancing) over long board-level distances (>10 cm) for free-space optical interconnects. We also combine these variations with active optoelectronic device variations (active-element tolerancing). We demonstrate a computer-aided analysis procedure that permits one to determine both active- and passive-element tolerances needed to achieve some system-level specification, such as yield or cost. The procedure that we employ relies on developing a detailed design of the system to be studied in a standard optical design program, such as code v. Using information from this model, we can determine the integrated power falling on the detector, which we term optical throughput, by performing Gaussian propagation or general Fresnel propagation (if significant vignetting occurs). This optical throughput can be used to determine system-level performance criteria, such as bit-error rate. With this computer-aided analysis technique, a sensitivity analysis of all the variations under study is made on a system with realistic board-level interconnect distances to find each perturbation's relative effects (with other perturbations set to 0) on the power falling on the detector. This information is used to set initial tolerances for subsequent tolerancing analysis and design runs. A tolerancing analysis by Monte Carlo techniques is applied to determine if the yield or cost (yield is denned as the percentage of systems that have acceptable system performance) is acceptable. With a technique called parametric sampling, a subsequent tolerancing design run can be applied to optimize this yield or cost with little increase in computation. We study a design example and show that most of the tolerances can be achieved with current technology.     

Scalable optical hypercube-based interconnection network for massively parallel computing

Two important parameters of a network for massively parallel computers are scalability and modularity. Scalability has two aspects: size and time (or generation). Size scalability refers to the property that the size of the network can be increased with nominal effect on the existing configuration. Also, the increase in size is expected to result in a linear increase in performance. Time scalability implies that the communication capabilities of a network should be large enough to support the evolution of processing elements through generations. A modular network enables the construction of a large network out of many smaller ones. The lack of these two important parameters has limited the use of certain types of interconnection networks in the area of massively parallel computers. We present a new modular optical interconnection network, called an optical multimesh hypercube (OMMH), which is both size and time scalable. The OMMH combines positive features of both the hypercube (small diameter, high connectivity, symmetry, simple routing, and fault tolerance) and the torus (constant node degree and size scalability) networks. Also presented is a three-dimensional optical implementation of the OMMH network. A basic building block of the OMMH network is a hypercube module that is constructed with free-space optics to provide compact and high-density localized hypercube connections. The OMMH network is then constructed by the connection of such basic building blocks with multiwavelength optical fibers to realize torus connections. The proposed implementation methodology is intended to exploit the advantages of both space-invariant free-space and multiwavelength fiber-based optical interconnect technologies. The analysis of the proposed implementation shows that such a network is optically feasible in terms of the physical size and the optical power budget.     

Compact polarization-based all-optical interconnection systems with growth capability

All-optical communication requires all-optical interconnections, thus leading to reliable, fast, and flexible modular communication means in future systems. Free-space approaches are advantageous since they fully use the two dimensions optics offer. A folded architecture based on a polarization code is proposed for dynamic optical interconnection. The suggested systems are compact and appropriate for both intracomputer and intercomputer communication. The modularity of the proposed architecture is presented, and a growth rule for the fully connected versions of the system is introduced. The proposed approach significantly reduces both the price of the interconnection systems and their complexity. Presented are 4 x 4 and 8 x 8 fully connected switches, a rearrangeable nonblocking 4 x 4 switch, and a crossbar architecture.     

Demonstration of optically controlled data routing with the use of multiple-quantum-well bistable and electro-optical devices

We present experimental results on a 1-to-64-channel free-space photonic switching demonstration system based on GaAs/GaAlAs multiple-quantum-well active device arrays. Two control schemes are demonstrated: data transparent optical self-routing usable in a packet-switching environment and direct optical control with potential signal amplification for circuit switching. The self-routing operation relies on the optical recognition of the binary destination address coded in each packet header. Address decoding is implemented with elementary optical bistable devices and modulator pixels as all-optical latches and electro-optical and gates, respectively. All 60 defect-free channels of the system could be operated one by one, but the simultaneous operation of only three channels could be achieved mainly because of the spatial nonhomogeneities of the devices. Direct-control operation is based on directly setting the bistable device reflectivity with a variable-control beam power. This working mode turned out to be much more tolerant of spatial noises: 37 channels of the system could be operated simultaneously. Further development of the system to a crossbar of N inputs and M outputs and system miniaturization are also considered.