Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04Q—SELECTING
  • H04Q11/00—Selecting arrangements for multiplex systems
  • H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
  • H04Q11/0005—Switch and router aspects
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04Q—SELECTING
  • H04Q11/00—Selecting arrangements for multiplex systems
  • H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
  • H04Q11/0062—Network aspects
• • G—PHYSICS
  • G02—OPTICS
  • G02B—OPTICAL ELEMENTS, SYSTEMS OR APPARATUS
  • G02B6/00—Light guides; Structural details of arrangements comprising light guides and other optical elements, e.g. couplings
  • G02B6/24—Coupling light guides
  • G02B6/26—Optical coupling means
  • G02B6/35—Optical coupling means having switching means
  • G02B6/354—Switching arrangements, i.e. number of input/output ports and interconnection types
  • G02B6/3542—Non-blocking switch, e.g. with multiple potential paths between multiple inputs and outputs, the establishment of one switching path not preventing the establishment of further switching paths
• • G—PHYSICS
  • G02—OPTICS
  • G02B—OPTICAL ELEMENTS, SYSTEMS OR APPARATUS
  • G02B6/00—Light guides; Structural details of arrangements comprising light guides and other optical elements, e.g. couplings
  • G02B6/24—Coupling light guides
  • G02B6/26—Optical coupling means
  • G02B6/35—Optical coupling means having switching means
  • G02B6/3596—With planar waveguide arrangement, i.e. in a substrate, regardless if actuating mechanism is outside the substrate
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04B—TRANSMISSION
  • H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
  • H04B10/25—Arrangements specific to fibre transmission
  • H04B10/2589—Bidirectional transmission
  • H04B10/25891—Transmission components
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04B—TRANSMISSION
  • H04B10/00—Transmission systems employing electromagnetic waves other than radio-waves, e.g. infrared, visible or ultraviolet light, or employing corpuscular radiation, e.g. quantum communication
  • H04B10/40—Transceivers
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04Q—SELECTING
  • H04Q11/00—Selecting arrangements for multiplex systems
  • H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
  • H04Q11/0005—Switch and router aspects
  • H04Q2011/0037—Operation
  • H04Q2011/005—Arbitration and scheduling
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04Q—SELECTING
  • H04Q11/00—Selecting arrangements for multiplex systems
  • H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
  • H04Q11/0062—Network aspects
  • H04Q2011/009—Topology aspects

Definitions

• the present disclosure relates to techniques for communicating optical signals. More specifically, the present disclosure relates to an optical cross-point macro-switch with a buffered switching matrix.
• Multistage Clos packet-switching networks are widely used in computing and telecommunications switching and routing systems to provide shared interconnectivity among many distinct endpoints or ports in these systems.
• these packet- switching networks are typically implemented as space-division switches that can scale to thousands of ports.
• Such memory buffers may also be used before and/or after the network in order to reduce head-of-line blocking, as well as to reduce switch contention because of output port blocking.
• efficiency of the network is generally limited, and careful (and relatively complex) scheduling techniques may be needed to ensure that the network is not overloaded or pushed past its critical loading into an inefficient operating regime.
• each stage may have not only routing and forwarding functionality, but may also have memory buffers and rich connectivity to preceding and following switching/routing stages.
• the need for memory at each stage may directly compete with the number of switches per stage and the number of stages that can be implemented. Therefore, the use of memory buffers at all stages in the switch may constrain the scalability of the switch.
• optical interconnects in VLSI switches can provide high-speed communication, and may permit large Clos packet- switching networks to be aggregated, e.g., by connecting smaller electrical switches with optical fiber links. While this architecture may facilitate the implementation of larger Clos packet- switching networks, it typically does not change the nature of the switching contention discussed previously. Indeed, the resulting Clos packet-switching networks usually have all the same congestion and inefficiencies, just at a larger scale.
• photonic- switching (or optical-switching) products can eliminate the electrical switching stage in favor of a 'transparent' optical switch, in which data packets are sent via beams of light from any input port to an arbitrary output port. While the speed and latency of transmission of these photonic-switching products are low, the input port and the output port contention issues (and, thus, the inefficiencies) remain.
• One embodiment of the present disclosure provides a macro-switch that includes a first integrated circuit.
• This first integrated circuit has a surface and includes: first switch sites, where each of the first switch sites includes first control logic and a first memory buffer; and second switch sites, where each of the second switch sites includes second control logic and a second memory buffer.
• the macro-switch includes a second integrated circuit having a second surface facing the surface.
• the second integrated circuit includes: optical ports that can be coupled to optical sources; optical waveguides optically coupled to the optical ports and the first switch sites; and second optical waveguides optically coupled to the first switch sites and the second switch sites. Note that the macro-switch has a fully connected topology between the first switch sites and the second switch sites.
• the macro-switch may include a cross-point switch. Moreover, the macro-switch may be non-blocking.
• the first control logic at a given first switch site may determine a given first switching schedule for the given first switch site
• the second control logic at the given second switch site may determine a given second switching schedule for the given second switch site.
• the given first switching schedule may be determined independently from other switching schedules for the first switch sites and the second switch sites
• the given second switching schedule may be determined independently from the other switching schedules for the first switch sites and the second switch sites.
• the optical waveguides between a given optical port and a given first switch site may include one optical waveguide that, during operation, conveys information from the given optical port to the given first switch site and another optical waveguide that, during operation, conveys information from the given first switch site to the given optical port.
• the second optical waveguides between a given first switch site and the given second switch site may include one optical waveguide that, during operation, conveys information from the given first switch site to the given second switch site and another optical waveguide that, during operation, conveys information from the given second switch site to the given first switch site.
• optical coupling may involve: a diffraction grating, a mirror, and/or optical proximity communication.
• the given first switch site may include transceivers that, during operation, convert input optical signals into input electrical signals and output electrical signals into output optical signals.
• the given second switch site may include second transceivers that, during operation, convert second input optical signals into second input electrical signals and second output electrical signals into second output optical signals.
• the second integrated circuit includes: a substrate; a buried- oxide (BOX) layer disposed on the substrate; and a semiconductor layer disposed on the BOX layer, where the optical waveguides and the second optical waveguides are, at least in part, implemented in the semiconductor layer.
• the substrate, the BOX layer and the semiconductor layer may constitute a silicon-on-insulator technology.
• Another embodiment provides a system that includes: a processor; a memory that stores a program module; and the macro-switch. During operation, the program module is executed by the processor.
• Another embodiment provides a method for switching optical signals using a macro-switch.
• the macro-switch conveys optical signals in optical waveguides in a second integrated circuit in the macro-switch.
• the macro-switch optically couples the optical signals from the optical waveguide to and from switch sites in a first integrated circuit in the macro-switch, where a given switch site includes control logic and a memory buffer, and where the control logic determines a switching schedule independently of other switch sites in the macro-switch.
• the macro-switch converts an optical signal to an electrical signal, performs switching, and converts the electrical signal into the optical signal, where the electrical signal is selectively stored in the memory buffer to avoid contention in the macro-switch.
• FIG. 1 is a block diagram illustrating a top view of a macro-switch in accordance with an embodiment of the present disclosure.
• FIG. 2 is a block diagram illustrating a side view of the macro-switch of FIG. 1 in accordance with an embodiment of the present disclosure.
• FIG. 3 is a block diagram illustrating a layout of a switch site in the macro-switch of FIG. 1 in accordance with an embodiment of the present disclosure.
• FIG. 4 is a block diagram illustrating an integrated circuit in the macro-switch of FIG. 1 in accordance with an embodiment of the present disclosure.
• FIG. 5 is a block diagram illustrating a system that includes the macro-switch of FIG. 1 in accordance with an embodiment of the present disclosure.
• FIG. 6 is a flow chart illustrating a method for switching optical signals using a macro-switch in accordance with an embodiment of the present disclosure.
• Embodiments of a macro-switch, a system that includes the macro-switch, and a technique for switching optical signals using the macro-switch are described.
• the macro-switch may include facing integrated circuits, one of which implements optical waveguides that convey optical signals, and the other which implements control logic, electrical switches and memory buffers at each of multiple switch sites.
• the macro-switch may have a fully connected topology between the switch sites.
• the memory buffers at each switch site provide packet buffering and congestion relief without causing undue scheduling/routing complexity.
• the macro-switch can be scaled to an arbitrarily large switching matrix (i.e., an arbitrary number of switch sites and/or switching stages).
• the macro-switch may have a 4096x4096 switching matrix (which is 6,700x larger than existing photonic switches).
• the macro-switch may have a small footprint and improved switch performance.
• the macro-switch may provide a low-power, high-bandwidth non-blocking electro-optical switch-on-a-chip.
• FIG. 1 presents a block diagram illustrating a top view of a macro-switch 100, with integrated circuits 110 and 112.
• Integrated circuit 110 may include: multiple switch sites 114 (which are sometimes referred to as 'leaf switch sites') that each include or are associated with instances of control logic (C.L.) 116 and memory buffers (M.B.) 118; and multiple switch sites 120 (which are sometimes referred to as 'spine switch sites') that each include or are associated with instances of control logic (C.L.) 122 and memory buffers (M.B.) 124.
• integrated circuit 112 may include optical waveguides 126 and 128.
• optical waveguides may optically couple optical ports (O.P.) 130 (which can be optically coupled to optical fibers or additional optical waveguides, and more generally to sources and sinks of optical signals) and switch sites 114. Furthermore, optical waveguides 128 may optically couple switch sites 114 and 120.
• O.P. 130 optically couple optical ports 130 (O.P.) 130 (which can be optically coupled to optical fibers or additional optical waveguides, and more generally to sources and sinks of optical signals) and switch sites 114.
• optical waveguides 128 may optically couple switch sites 114 and 120.
• FIG. 2 which presents a block diagram illustrating a side view of macro-switch 100
• integrated circuits 110 and 112 have surfaces 210 and 212 that face each other.
• a given switch site in switch sites 114 in FIG. 1 (such as switch site 114-1) may be optically coupled to a given one of optical ports 130 in FIG. 1 (such as optical port 130-1) by at least one of optical waveguides 126 in FIG. 1 (such as optical waveguide 126-1 in FIG. 2).
• optical waveguides 126 and 128 may convey optical signals unidirectionally. (However, in other embodiments optical waveguides 126 and 128 in FIG. 1 convey optical signals bidirectionally.)
• Input optical signals may be optically coupled from integrated circuit 112 to integrated circuit 110 by optical couplers.
• an input optical signal conveyed using optical waveguide 126-1 may be optically coupled to switch site 114-1 by optical coupler (O.C.) 210-1.
• This optical coupler may include a diffraction grating and/or a mirror.
• the optical coupling involves optical proximity communication, in which a vertical spacing between integrated circuits 110 and 112 is less than or on the same order as one or more carrier wavelengths of optical signals that are optically coupled between integrated circuits 110 and 112.
• the input optical signals may be converted into electrical signals by transceivers.
• the input optical signal from optical coupler 210-1 may be converted into an electrical signal by transceiver (TR.) 212-1.
• TR. transceiver
• one or more packets associated with this electrical signal may be processed by an instance of control logic 116 (FIG. 1) and may be selectively stored in an instance of memory buffers 118 (FIG. 1) to avoid contention in macro-switch 100.
• the instance of control logic 116 may: set an appropriate routing or switching state of a switch at switch site 114-1 (such as switch 132-1 in switches (SW.) 132 in FIG. 1); optionally access the packet in the instance of memory buffers 118 (FIG. 1); and output an associated electrical signal.
• This electrical signal may be converted into an optical signal by transceiver 212-2.
• this optical signal may be optically coupled from integrated circuit 110 into at least one of optical waveguides 128 in FIG. 1 (such as optical waveguide 128-1 in FIG. 2) by optical coupler 210-2. Then, optical waveguide 128-1 may convey the optical signal to a given switch site in switch sites 120 in FIG. 1 (such as switch site 120-1).
• the optical signal may be optically coupled from integrated circuit 112 to switch site 120-1 by optical coupler 210-3. Furthermore, at switch sites 120-1, the optical signal may be converted into an electrical signal by transceiver 212-3. As described further below with reference to FIG. 3, one or more packets associated with this electrical signal may be processed by an instance of control logic 122 (FIG. 1) and may be selectively stored in an instance of memory buffers 124 (FIG. 1) to avoid contention in macro-switch 100.
• control logic 122 FIG. 1
• memory buffers 124 FIG. 1
• the instance of control logic 122 may: set an appropriate routing or switching state of a switch at switch site 120-1 (such as switch 134-1 in switches (SW.) 134 in FIG. 1); optionally access the packet in the instance of memory buffers 124 (FIG. 1); and output an associated electrical signal.
• This electrical signal may be converted into an optical signal by transceiver 212-4.
• this optical signal may be optically coupled from integrated circuit 110 into at least one of optical waveguides 128 in FIG. 1 (which may be different from or other than optical waveguide 128-1) by optical coupler 210-4.
• the aforementioned operations may then be repeated as the optical signal is conveyed to another one of switch sites 114 (FIG. 1), where it is eventually routed as an output optical signal to one of optical ports 130 (FIG. 1).
• the return path from switch site 120-1 to the one of the optical ports 130 in FIG. 1 may involve different optical waveguides in optical waveguides 126 and 128, and a different one of switch sites 114 in FIG. 1.
• an instance of control logic 116 at a given switch site in switch sites 114 may determine a given switching schedule for this (local) switch site, and an instance of control logic 122 at a given switch site in switch sites 120 may determine a given switching schedule for this (local) switch site.
• the switching schedules may be determined independently of each other (i.e., independently of any other switching schedule in macro-switch 100).
• macro-switch 100 provides a fully connected topology between switch sites 114 and 120.
• macro-switch 100 may include a cross-point switch, which can support one-to-many or even all-to-all (full-mesh) connectivity among switch sites 114 and 120.
• optical waveguides 126 and 128 may be laid out in concentric circles in macro-switch 100 to avoid optical waveguide crossing.
• macro-switch 100 may be non-blocking and, thus, may offer high throughput.
• the macro-switch is used in a distributed compute environment (such as an enterprise data center or a modern supercomputer).
• a distributed compute environment such as an enterprise data center or a modern supercomputer.
• communication between compute nodes can be a bottleneck, both in bandwidth and latency.
• a non-blocking network configuration is often employed, such as a folded-Clos (fat-tree) network with full-bisection bandwidth.
• the network switching can be a sizable or even a dominant component of the overall system power consumption.
• Modern distributed compute environments typically use electronic switches, such as an existing Infiniband electronic switch.
• an existing Infiniband electronic switch can provide full non-blocking bisection bandwidth with a switching capacity of 130 Tbps (which, as described further below, may approximately equal that of some embodiments of macro-switch 100).
• researchers are investigating photonic switching based on a variety of switching techniques (such as MEMS, acousto-optical, magneto-optical, etc.) to steer optical inputs to optical outputs.
• M£MS-based silicon photonic switch capable of switching 50 input ports to 50 output ports has been demonstrated.
• aM£MS-based silicon photonic switch typically has a sub-microsecond switch time that is 2-3 orders of magnitude slower than the effective switch speed of the electro-optical switch sites used in macro-switch 100.
• Macro-switch 100 may use: silicon-photonics links (such as optical waveguides) that provide direct point-to-point connectivity between any two switch sites; switching nodes at each switch site to build an arbitrarily large switching matrix (i.e., an arbitrary number of switches per stage and an arbitrary number of stages); and memory buffers at each switch site to provide packet buffering and congestion relief without causing undue scheduling/routing complexity.
• silicon-photonics links such as optical waveguides
• switching nodes at each switch site to build an arbitrarily large switching matrix (i.e., an arbitrary number of switches per stage and an arbitrary number of stages); and memory buffers at each switch site to provide packet buffering and congestion relief without causing undue scheduling/routing complexity.
• the result may be a low-power, high-bandwidth, non-blocking electro-optical switch-on-a-chip.
• macro-switch 100 may include: switching and memory at each switch site; a photonic point-to-point interconnect; and a logical distribution of input and output ports across the switch sites.
• control logic such as processors or logic circuits at each switch site may: examine each incoming packet; look at the packet destination; determine where to send the packet based on routing information; determine when to send the packet; selectively store the packet in local memory in the interim; and, when ready to send, select the appropriate outgoing photonic link based on the routing table information.
• the point-to-point interconnect is implemented using silicon photonics, and employs wavelength-division multiplexing and embedded optical waveguides.
• the memory buffers at the switch sites may use one or more of a variety of types of memory, such as: DRAM, SRAM, EEPROM, flash, another type of non- volatile memory and/or another type of volatile memory.
• macro-switch 100 may use a mix of electronics and silicon photonics to achieve high bandwidth while keeping the power cost low.
• macro-switch 100 may use an on-chip network of switch sites that are optically connected by optical waveguides.
• macro-switch 100 in conjunction with the outbound optical fiber connections, may have a power consumption of 835 W (which is llx smaller) for equivalent switching capacity.
• macro-switch 100 may be implemented using a single substrate, while the existing Infiniband electronic switch may require a 28U chassis weighing almost a half a metric ton (496 kg).
• An exemplary embodiment of the macro-switch may be implemented using a square silicon substrate that is 20 cm on a side. Consequently, the perimeter of this macro-switch may total 800 mm. With an estimated effective optical fiber pitch of 300 ⁇ , this may allow for approximately 2500 total optical fibers to be connected along the periphery (leaving
• each of the input/output optical fibers may convey or carry eight carrier wavelengths at 16 Gbps per carrier wavelength, for a total of 4096 input/output carrier-wavelength pairs. This may provide a total of 65.5 Tbps input bandwidth plus 65.5 Tbps output bandwidth per macro-switch.
• macro-switch 100 may have a folded-Clos (fat-tree) internal connectivity.
• FIG. 1 is intended to depict a block diagram of the functionality of macro-switch 100, not the actual physical layout.
• the internal structure of macro-switch 100 may include 67 switch sites (including 43 leaf switch sites and 24 spine switch sites) that are connected with silicon-photonics optical waveguides.
• Optical proximity communication may optically couple the switch sites to the optical layers in integrated circuit 112 that routes the optical links, both for the internal optical waveguides and for the external optical fiber.
• power for the on-chip silicon-photonics interconnection network may be provided by off-macro-switch lasers that are connected through optical fibers to the sides of the macro-switch.
• the internal macro-switch topology may use a folded-Clos (fat-tree) among switch sites to provide non-blocking full-bisection bandwidth.
• each connection between any two of the switch sites is bidirectional, with a specified number of optical waveguides and carrier wavelengths in each direction.
• the macro-switch may include: 16 optical ports to 16 optical fibers that each convey 128 carrier wavelengths; 43 optical waveguides in each direction between the optical ports and the leaf switch sites, each of which conveys four carrier wavelengths (for a total of 192 carrier wavelengths); and 192 optical waveguides in each direction between the leaf switch sites and the spine switch sites, each of which conveys one carrier wavelength (once again, for a total of 192 carrier wavelengths).
• each leaf switch site and each spine switch site can connect up to 192 bidirectional carrier- wavelength pairs.
• FIG. 3 presents a block diagram illustrating a layout of a switch site 300 in macro- switch 100 (FIGs. 1 and 2).
• This switch site may include 192 optical serializers/deserializers (SERDES), with one per bidirectional carrier-wavelength pair.
• switch site 300 may include transmitter (Tx) routing or control logic that may examine the destination of each outgoing message or packet, and may select the outgoing carrier wavelength based on a routing table. Once the carrier wavelength is selected, the message may be sent to its queue, and the serializer may start modulating the carrier wavelength based on the message bits or the data.
• Tx transmitter
• switch site 300 may include receiver (Rx) routing or control logic that may: examine the destination of each incoming message or packet, copy the message into an input/output (IIO) scratchpad, and copy the information in the message into the queue for the transmitter routing logic.
• Rx receiver
• IIO input/output
• routing (RTG) table at switch site 300 may provide a mapping between the destination-site identifier and one or more carrier wavelengths that are associated with the destination, and which are used to convey messages or packets to the destination.
• the input/output scratchpad may be an SRAM-based scratchpad that provides direct- addressable storage for messages.
• the input/output scratchpad may be 3 MB.
• the input/output scratchpad may be organized into 12 banks, where each bank can serve data independently to the receiver and transmitter multiplexers. (Note that the banks may be accessed simultaneously in a fully pipelined manner.)
• switch site 300 may include input and output crossbars between the serializers/deserializers and the input/output scratchpad.
• each switch site in the macro-switch is estimated to occupy approximately 20 mm 2 and consumes under 8 W.
• each leaf switch site may have 96 bidirectional carrier-wavelength pairs that connect to the external optical fiber, and 96 bidirectional carrier-wavelength pairs that connect to the spine switch sites. Because the periphery of the macro-switch can have a total of 4096 bidirectional carrier-wavelength pairs, a total of 43 leaf switch sites may be needed to connect to the external optical fiber connections. (Note that the 43 leaf switch sites, each with 96 external bidirectional carrier-wavelength pairs, may connect to up to 4128 bidirectional carrier-wavelength pairs.)
• each leaf switch site may be distributed across all of the spine switch sites.
• each spine site may have one optical waveguide that contains or conveys four carrier wavelengths (per direction) connected to each leaf site. Because a spine switch site may have a maximum connectivity of 192 wavelengths in one direction, the macro-switch may need a total of 24 spine switches to connect all leaf switches in a non-blocking full-bisection bandwidth configuration.
• each switch site may need a 3 MB input/output scratchpad with 12 banks (256 KB/bank).
• the total number of optical waveguides needed to internally connect the switch sites may be approximately half the number for a macro- switch.
• the total power consumption per macro-switch including the power required for the internal optical waveguides, but excluding the power required for outgoing optical fibers, may be approximately 759 W. If 10% additional power is included for cooling, the power consumption per macro- switch is approximately 835 W.
• One consideration in the design of the macro-switch is the location of the laser sources to power the photonic links. These laser sources can be placed either outside the macro- switch or as on-chip lasers within the macro-switch. When on-chip lasers are used, additional power saving can be obtained. In particular, because only a subset of the on-macro- switch point- to-point links are typically used at any given moment due to the fully connected point-to-point architecture, it is possible to provision a set of M tunable lasers per N-channel wavelength- division-multiplexing link (where M is less than N) to deliver power to any arbitrary subset of M links by tuning to the correct set of carrier wavelengths.
• the links may be almost fully provisioned so that M is almost equal to N (i.e., the number of tunable laser outputs may be sufficient to power almost all links at any given point in time) or, in general, optimally provisioned so that the number of tunable laser outputs M is chosen so that laser (and, hence, power) utilization is optimized.
• a tuning delay e.g., less than 100 ns
• this approach may allow power efficiency to be traded off for available bandwidth.
• the macro-switch may include optical waveguides implemented in an integrated circuit, such as a photonic integrated circuit.
• FIG. 4 presents a block diagram illustrating an integrated circuit 400.
• This integrated circuit includes: a substrate 410, a buried-oxide (BOX) layer 412 disposed on substrate 410, and a semiconductor layer 414 disposed on buried-oxide layer 412.
• Optical components such as the optical waveguides may, at least in part, be included in buried-oxide layer 412 and semiconductor layer 414.
• substrate 410 and semiconductor layer 414 include silicon and buried-oxide layer 412 includes silicon dioxide.
• substrate 410, buried-oxide layer 412 and semiconductor layer 414 may constitute a silicon-on- insulator technology.
• semiconductor layer 414 may have a thickness 416 that is less than 1 ⁇ (such as 0.2-0.3 ⁇ ).
• buried-oxide layer 412 may have a thickness 418 between 0.3 and 3 ⁇ (such as 0.8 ⁇ ).
• a width of the optical waveguides may be 400- 3000 nm.
• the optical waveguides convey an optical signal (i.e., light) having wavelengths between 1.1-1.7 ⁇ , such as an optical signal having a fundamental wavelength of 1.3 or 1.55 ⁇ .
• Macro-switch 100 may be included in a system and/or an electronic device. This is shown in FIG. 5, which presents a block diagram illustrating a system 500 that includes macro-switch 100.
• system 500 includes processing subsystem 510 (with one or more processors) and memory subsystem 512 (with memory).
• system 500 may include one or more program modules or sets of instructions stored in a memory subsystem 512 (such as DRAM or another type of volatile or non- volatile computer-readable memory), which, during operation, may be executed by processing subsystem 510.
• a memory subsystem 512 such as DRAM or another type of volatile or non- volatile computer-readable memory
• the one or more computer programs may constitute a computer-program mechanism.
• instructions in the various modules in memory subsystem 512 may be implemented in: a high-level procedural language, an object-oriented programming language, and/or in an assembly or machine language.
• the programming language may be compiled or interpreted, e.g., configurable or configured, to be executed by the processing subsystem.
• Components in system 500 may be coupled by signal lines, links or buses. These connections may include electrical, optical, or electro-optical communication of signals and/or data. Furthermore, in the preceding embodiments, some components are shown directly connected to one another, while others are shown connected via intermediate components. In each instance, the method of interconnection, or 'coupling,' establishes some desired
• circuit nodes or terminals.
• Such coupling may often be accomplished using a number of circuit configurations, as will be understood by those of skill in the art; for example, AC coupling and/or DC coupling may be used.
• functionality in these circuits, components and devices may be implemented in one or more: application-specific integrated circuits (ASICs), field- programmable gate arrays (FPGAs), and/or one or more digital signal processors (DSPs).
• ASICs application-specific integrated circuits
• FPGAs field- programmable gate arrays
• DSPs digital signal processors
• system 500 may be at one location or may be distributed over multiple, geographically dispersed locations.
• System 500 may include: a VLSI circuit, a switch, a hub, a bridge, a router, a communication system (such as a wavelength-division-multiplexing communication system), a storage area network, a data center, a network (such as a local area network), and/or a computer system (such as a multiple-core processor computer system).
• a VLSI circuit such as a wavelength-division-multiplexing communication system
• a storage area network such as a wavelength-division-multiplexing communication system
• a data center such as a data center
• a network such as a local area network
• a computer system such as a multiple-core processor computer system
• the computer system may include, but is not limited to: a server (such as a multi-socket, multi-rack server), a laptop computer, a communication device or system, a personal computer, a work station, a mainframe computer, a blade, an enterprise computer, a data center, a tablet computer, a supercomputer, a network-attached-storage (NAS) system, a storage-area-network (SAN) system, a media player (such as an MP3 player), an appliance, a subnotebook/netbook, a tablet computer, a smartphone, a cellular telephone, a network appliance, a set-top box, a personal digital assistant (PDA), a toy, a controller, a digital signal processor, a game console, a device controller, a computational engine within an appliance, a consumer-electronic device, a portable computing device or a portable electronic device, a personal organizer, and/or another electronic device.
• a server such as a multi-socket,
• the embodiments of macro-switch 100 and/or system 500 may include fewer components or additional components. Although these embodiments are illustrated as having a number of discrete items, these optical components, integrated circuits and the system are intended to be functional descriptions of the various features that may be present rather than structural schematics of the embodiments described herein. Consequently, in these embodiments two or more components may be combined into a single component, and/or a position of one or more components may be changed. In addition, functionality in the preceding embodiments of macro-switch 100 and/or system 500 may be implemented more in hardware and less in software, or less in hardware and more in software, as is known in the art.
• the preceding embodiments have been illustrated with particular elements and compounds, a wide variety of materials and compositions (including stoichiometric and non- stoichiometric compositions) may be used, as is known to one of skill in the art.
• a silicon optical waveguide was illustrated in the preceding embodiments, the switching technique may be used with other materials, as is known to one of skill in the art.
• the semiconductor layer may include polysilicon or amorphous silicon.
• the materials and compounds in the macro-switch may be fabricated using a wide variety of processing techniques, including: evaporation, sputtering, molecular-beam epitaxy, chemical vapor deposition, wet or dry etching (such as photolithography or direct- write lithography), polishing, etc.
• components in macro-switch 100 may be defined using an additive process (i.e., material deposition) and/or a subtractive process (i.e., material removal), and these components may be fabricated using a wide variety of materials, including: a semiconductor, metal, glass, sapphire, silicon dioxide, organic materials, inorganic materials, a resin and/or polymers.
• a wide variety of optical components may be used in or in conjunction with macro-switch 100.
• FIG. 6 presents a flow chart illustrating a method 600 for switching optical signals using a macro-switch, such as macro- switch 100 (FIGs. 1 and 2).
• the macro-switch conveys the optical signals (operation 610) in optical waveguides in a second integrated circuit in the macro-switch.
• the macro-switch optically couples the optical signals (operation 612) from the optical waveguide to and from switch sites in a first integrated circuit in the macro-switch, where a given switch site includes control logic and a memory buffer, and where the control logic determines a switching schedule independently of other switch sites in the macro- switch.
• the macro- switch converts an optical signal to an electrical signal (operation 614), performs switching (operation 618), and converts the electrical signal into the optical signal (operation 620), where the electrical signal is selectively stored (operation 616) in the memory buffer to avoid contention in the macro-switch.

Landscapes

• Physics & Mathematics (AREA)
• Engineering & Computer Science (AREA)
• Computer Networks & Wireless Communication (AREA)
• Signal Processing (AREA)
• General Physics & Mathematics (AREA)
• Optics & Photonics (AREA)
• Electromagnetism (AREA)
• Optical Communication System (AREA)
• Optical Integrated Circuits (AREA)
• Use Of Switch Circuits For Exchanges And Methods Of Control Of Multiplex Exchanges (AREA)
• Optical Modulation, Optical Deflection, Nonlinear Optics, Optical Demodulation, Optical Logic Elements (AREA)

Priority Applications (3)

Applications Claiming Priority (4)

Publications (1)

Family

ID=56345197

Family Applications (1)

Country Status (5)

Families Citing this family (17)

Citations (4)

Family Cites Families (10)

• 2016
  • 2016-05-13 US US15/154,619 patent/US9693124B2/en active Active
  • 2016-06-07 CN CN201680030884.2A patent/CN107667537B/zh active Active
  • 2016-06-07 JP JP2017558422A patent/JP6683734B2/ja active Active
  • 2016-06-07 EP EP16734775.6A patent/EP3308484B1/en active Active
  • 2016-06-07 WO PCT/US2016/036249 patent/WO2016200831A1/en not_active Ceased

Patent Citations (4)

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events