Optimized switching fabric with multipoint optics for data centers

Abstract

To provide an express path capable of reducing latency between an edge of a data center and leaf nodes.SOLUTION: An optical communication system includes a hub optical transceiver, a power splitter, and a plurality of spoke transceivers. The hub optical transceiver is configured for receiving a spectrum of wavelengths. The power splitter is coupled to the hub optical transceiver, and operates as a passive device that is configured to replicate the spectrum of wavelengths and output a plurality of replicated spectrum of wavelengths, and each replicated spectrum of wavelengths has a corresponding power that is a fraction of a total power received from the hub optical transceiver. The plurality of spoke transceivers is coupled to the power splitter and each of the plurality of spoke transceivers is configured to receive a corresponding one of the plurality of replicated spectrum of wavelengths.SELECTED DRAWING: Figure 4A

Description

[Technical Field]

[0001] This disclosure relates to Internet Protocol (IP) switching fabrics and / or topologies, and more particularly to Optical Data Center Fabrics (ODCFs) that offer the benefits of a full IP cross-fabric and provide high-speed paths between the edge and leaf nodes of a data center using cutting-edge flexible optics technology. [Background technology]

[0002] In recent years, there has been a continuous push for online services that enable streaming-style online or cloud gaming between cloud gaming servers and clients connected via a network. The streaming style has become increasingly popular due to the availability of on-demand game titles, the ability to run more complex games, the ability to network among players for multiplayer games, the sharing of assets among players, the sharing of instant experiences among players and / or spectators, the ability to allow friends to watch video games being played by friends, and the ability to allow friends to join in on a friend's gameplay in progress.

[0003] Unfortunately, demand is also pushing against the limits of network connectivity capabilities. In particular, Internet Protocol (IP) fabrics based on cross-switching topologies or fabrics, which provide multiple multi-stage circuit switching networks, have dominated the cloud industry in recent years due to their ability to scale through cookie-cutter elements and their highly flexible routing designs. However, as five-stage or even higher CLOS switching topologies are required to provide the required host connectivity, the number of devices dedicated to "spine" functions, with no user / host connectivity, grows rapidly. These spine-only devices add cost in both switching hardware and transceivers, operational complexity that must be managed, and active points of failure.

[0004] It would also be advantageous to reduce the capital and operational costs and complexity of large-scale switching topologies or fabrics, as well as reduce latency between hosts and end-user services. It is against this background that embodiments of the present disclosure have been made. Summary of the Invention

[0005] Embodiments of the present disclosure relate to an Optical Data Center IP Switching Fabric (ODCF) and / or topology that provides the benefits of a full Internet Protocol (IP) multi-stage switching fabric, but with reduced latency without the complexity and cost, and that uses cutting-edge flexible optics technology to provide high-speed paths between the edge and leaf nodes of a data center.

[0006] An embodiment of the present disclosure discloses an optical communication system. The optical communication system includes a hub optical transceiver configured to receive wavelengths of a spectrum. The optical communication system includes a power splitter coupled to the hub optical transceiver. The power splitter operates as a passive device configured to replicate wavelengths of the spectrum and output a plurality of replicated wavelengths of the spectrum. Each wavelength of the replicated wavelengths of the spectrum has a corresponding power that is a fraction of the total power received from the hub optical transceiver. The optical communication system includes a plurality of spoke transceivers coupled to the power splitter. Each of the plurality of spoke transceivers is configured to receive a corresponding one of the wavelengths of the replicated spectrum. Each spoke transceiver is tunable to select a band of wavelengths that sets a bandwidth for each spoke transceiver. For example, a spoke transceiver can be tuned to one or more wavelengths, where the plurality of tuned wavelengths may be contiguous or non-contiguous.

[0007] An embodiment of the present disclosure provides an optical communication system. The optical communication system includes a hub optical transceiver configured to receive wavelengths of a spectrum. The optical communication system includes a power splitting layer coupled to the hub optical transceiver. The power splitting layer is configured to output a plurality of replicated wavelengths of the spectrum. Each replicated wavelength of the plurality of replicated wavelengths has a corresponding power that is a fraction of the total power received from the optical transceiver. The power splitting layer includes a plurality of power splitters arranged in one or more cascaded layers, each power splitter operating as a passive device. The optical communication system includes a plurality of spoke transceivers coupled to the power splitting layer. Each of the plurality of spoke transceivers is configured to receive a corresponding one of the plurality of replicated wavelengths. Each spoke transceiver is tunable to select a band of wavelengths that sets a bandwidth for each spoke transceiver. For example, the spoke transceivers can be tuned to one or more wavelengths, and the plurality of tuned wavelengths may be contiguous or non-contiguous.

[0008] An embodiment of the present disclosure provides an optical communication system. The optical communication system includes a hub optical transceiver configured to receive wavelengths of a spectrum. The optical communication system includes a first power splitter coupled to the hub optical transceiver. The first power splitter operates as a passive device configured to replicate wavelengths of the spectrum and output a first plurality of replicated wavelengths of the spectrum. Each wavelength of the first plurality of replicated wavelengths of the spectrum has a corresponding power that is a fraction of the total power received from the hub optical transceiver. The optical communication system includes spoke transceivers connected to the first power splitter. The spoke transceivers are configured to receive wavelengths of the replicated spectrum from the first plurality of wavelengths of the replicated spectrum. The optical communication system includes a second power splitter coupled to the spoke transceivers. The second power splitter is configured to receive wavelengths of the replicated spectrum from the first plurality of wavelengths of the replicated spectrum and output wavelengths of a second plurality of replicated wavelengths of the spectrum. The optical communication system includes a server coherent transceiver coupled to the second power splitter. The coherent transceiver is configured to receive wavelengths of the replicated spectrum from the second plurality of wavelengths of the replicated spectrum, and the coherent transceiver is tunable to select a corresponding band of wavelengths from the wavelengths of the replicated spectrum from the second plurality of wavelengths, the band of wavelengths setting a bandwidth for the coherent transceiver. For example, the coherent transceiver can be tuned to one or more wavelengths, where the multiple tuned wavelengths can be contiguous or non-contiguous.

[0009] Other aspects of the present disclosure will become apparent from the following detailed description, taken in conjunction with the accompanying drawings, illustrated by way of example of the principles of the disclosure.

[0010] The present disclosure is best understood by reference to the following detailed description taken in conjunction with the accompanying drawings. [Brief explanation of the drawings]

[0011] [Figure 1] FIG. 1 illustrates a diagram of a game cloud system for providing games over a network among one or more computing nodes located in one or more data centers, according to one embodiment of the present disclosure. [Figure 2A] FIG. 1 illustrates a diagram of multiple rack assemblies including multiple computing nodes in a representative data center of a gaming cloud system, according to one embodiment of the present disclosure. [Figure 2B] FIG. 1 is a diagram of a rack assembly including a top-of-rack switch configured to receive a full spectrum of wavelengths, according to one embodiment of the present disclosure. [Figure 3A] 1 illustrates an optical switching fabric with multi-point optics and composed of one super spine according to one embodiment of the present disclosure. [Figure 3B] 1 illustrates an optical switching fabric with multi-point optics and composed of one or more superspines, according to one embodiment of the present disclosure. [Figure 4A] 1 illustrates an optical switching fabric configured with point-to-multipoint optical systems, according to one embodiment of the present disclosure. [Figure 4B] 4B illustrates connections between components of the optical data center switching fabric of FIG. 4A according to one embodiment of the disclosure. [Figure 5A] 1 illustrates an exemplary layout of components of a data center optical switching fabric including a single splitter layer, according to one embodiment of the present disclosure. [Figure 5B] 1 illustrates a data center optical switching fabric including multiple cascaded splitter layers according to one embodiment of the present disclosure. [Figure 6A] FIG. 1 illustrates a data center switching fabric including the use of 1×4 (1*4) splitters across three cascaded splitter layers connected to a superspine switching layer according to one embodiment of the present disclosure. [Figure 6B]6B illustrates a physical data center layout implementing the switching fabric of FIG. 6A according to one embodiment of the present disclosure. [Figure 7] FIG. 1 illustrates a data center switching fabric configured to perform switching at the host / server rack layer, according to one embodiment of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0012] Although the following detailed description includes many specific details for purposes of explanation, those skilled in the art will recognize that many variations and modifications to the following details are within the scope of the present disclosure. Accordingly, the aspects of the present disclosure described below are set forth without any loss of generality to, and without imposing limitations on, the claims that follow this description.

[0013] Generally speaking, embodiments of the present disclosure provide an optical data center fabric (ODCF) and / or topology that offers the benefits of a full Internet Protocol (IP) multi-stage switching network, and further provides high-speed paths between edge and leaf nodes of the data center provided using flexible optics technology. In particular, the middle layer of switches can be replaced with passive optical power splitters in the data fabric. Benefits include reduced capital and operational costs of large optical data center fabrics and increased ease of implementing optical data center fabrics. Additional benefits include an ultra-low latency path between hosts and end-user services because latency at the middle layer is improved by using passive optical splitters rather than switches. This is because conventional switches perform optical-electrical-optical (OEO) conversion along with some routing and / or switching operations (e.g., packet queuing) that introduce latency, whereas optical power splitters are purely passive optical devices that add virtually no latency. That is, embodiments of the present disclosure significantly reduce latency through optical data center fabrics while reducing capital expenditures by removing intermediate switch layer(s) and replacing them with intermediate power splitting layer(s) (e.g., replacing expensive switch hardware with fewer cost-effective passive optical devices in the fiber optic infrastructure), and simultaneously reducing operational expenses (e.g., reduced power usage and less maintenance required). Additionally, embodiments of the present disclosure simplify network topology, resulting in easier expansion (e.g., scaling, or implementing multiple buildings in more geographic locations, etc.) and troubleshooting activities.

[0014] Based on the above general understanding of various embodiments, detailed examples of the embodiments will now be described with reference to the various figures.

[0015] Throughout this specification, references to an "application" or "game" or "video game" or "game application" or "game title" are meant to represent any type of interactive application that is directed through the execution of input commands. By way of example only, interactive applications include applications for gaming, word processing, video processing, video game processing, etc. Furthermore, the terms introduced above are interchangeable.

[0016] Embodiments of the present disclosure support a large number of computing systems and are implemented within data centers. Each computing system can be configured to perform one or more functions depending on the application running on the corresponding computing system. The computing systems within a particular data center can be configured primarily to provide a service generally (e.g., cloud gaming, cloud storage, etc.) or can each be individually configured to provide a service specific to a client. For example, a data center may be configured generally to provide cloud gaming services to multiple clients (e.g., remote computing devices, each associated with a corresponding user). In another example, a data center may be configured generally to provide a financial services platform to many customers, such as facilitating high-frequency trading of financial securities (e.g., stocks). These data centers, with the low latency, reduced infrastructure costs, ease of scaling, and multi-building implementation characteristics provided by embodiments of the present disclosure, are desirable for providing these various services. For example, in the case of cloud gaming services, these data centers are well-suited to provide the best user experience for gamers playing complex games, with little or no processing and / or display interruptions and actual on-demand play of any game title. Additionally, with respect to financial services, these data centers are better suited to provide increased trading frequency and speed (i.e., faster response times upon submission of trading orders, etc.) than more traditionally configured data centers. For simplicity and ease of explanation, embodiments of the present disclosure are described within the context of data centers providing cloud gaming, although it is understood that data centers may be configured to provide other services and / or functions, such as, for example, financial services.

[0017] FIG. 1 is a diagram of a system 100 for providing gaming over a network 150 between one or more compute nodes located in one or more data centers, each configured with an IP switching fabric employing passive optical power splitter devices to provide high-speed paths between edge and leaf nodes of the data center with reduced latency, ease of implementation, and reduced cost, according to one embodiment of the present disclosure. According to one embodiment of the present disclosure, the system is configured to provide gaming over a network between one or more cloud gaming servers, and more specifically, to provide high-speed access from the compute nodes to network storage, such as in a rack assembly. Cloud gaming involves running video games on the servers to generate game-rendered video frames, which are then sent to clients for display. In other embodiments, the system 100, including one or more data centers, may be configured to provide other services, such as financial services.

[0018] It is also understood that cloud gaming, in various embodiments (e.g., within a cloud gaming environment or a standalone system), can be executed using physical machines (e.g., central processing units—CPUs—and graphics processing units—GPUs), or virtual machines, or a combination of both. For example, virtual machines (e.g., instances) can be created using a hypervisor on host hardware (e.g., located in a data center) that utilizes one or more components of a hardware layer, such as multiple CPUs, memory modules, GPUs, network interfaces, communication components, etc. These physical resources can be arranged in racks, such as a rack of CPUs, a rack of GPUs, a rack of memory, etc., where the physical resources of the racks are accessible using a top-of-rack switch that facilitates a fabric for assembling and accessing the components used in the instances (e.g., when building the virtualized components of the instances). Generally, a hypervisor can present multiple guest operating systems in multiple instances that are configured with virtual resources. That is, each operating system can be configured with a corresponding set of virtualized resources supported by one or more hardware resources (e.g., located in a corresponding data center). For example, each operating system may be supported by a virtual CPU, multiple virtual GPUs, virtual memory, virtualized communication components, etc. Additionally, it is a configuration of an instance that can be transferred from one data center to another to reduce latency. Instant-uses defined for a user or game can be used when saving a user's game session. Instant-uses can include any number of configurations described herein to optimize fast rendering of video frames for a game session. In one embodiment, instant-uses defined for a game or user can be transferred between data centers as configurable settings.The ability to transfer instant-use settings allows for efficient migration of game play from data center to data center when users connect to play games from different geographic locations.

[0019] System 100 includes a gaming cloud system 190 implemented across one or more data centers (e.g., data centers 1 through N). As shown, an instance of gaming cloud system 190 may be located in data center N that provides management functions, where the management functions of gaming cloud system 190 may be distributed across multiple instances of gaming cloud system 190 at each data center. In some implementations, gaming cloud system management functions may be located outside of any of the data centers.

[0020] The gaming cloud system 190 includes an assigner 191 configured to assign each of the client devices (e.g., 1-N) to corresponding resources in a corresponding data center. In particular, when the client device 110 logs into the gaming cloud system 190, the client device 110 may connect to an instance of the gaming cloud system 190 at data center N, which may be geographically closest to the client device 110. The assigner 191 may perform diagnostic tests to determine available transmit and receive bandwidth to the client device 110. Based on the tests, the assigner 191 may assign resources to the client device 110 very specifically. For example, the assigner 191 may assign a particular data center to the client device 110. Furthermore, the assigner 191 may assign a particular compute thread, a particular streaming array, or a particular compute node in a particular rack assembly to the client device 110. The assignment may be performed based on knowledge of the assets (e.g., games) available on the compute nodes. Previously, client devices were generally assigned to data centers and not further assigned to rack assemblies. In this manner, assigner 191 can assign client devices requesting execution of a particular computationally intensive game application to compute nodes that may not be running the computationally intensive application. Additionally, load management of the assignment of compute-intensive game applications requested by clients can be performed by assigner 191. For example, the same compute-intensive game application requested for a short period of time may be distributed across different compute nodes in one rack assembly or different rack assemblies to reduce the load on a particular compute node, compute thread, and / or rack assembly.

[0021] In some embodiments, allocation may be performed based on machine learning. In particular, resource demand may be predicted for a particular data center and its corresponding resources. For example, if a data center is predicted to soon handle many clients running computationally intensive gaming applications, assigner 191 may assign client devices 110 with that knowledge, allocating resources that may not currently be utilizing their full capacity. In another case, assigner 191 may switch client devices 110 from gaming cloud system 190 in data center N to resources available in data center 3 in anticipation of increased load at data center N. Furthermore, future clients may be allocated resources in a distributed manner, such that resource load and demand may be distributed across the entire gaming cloud system, across multiple data centers, across multiple rack assemblies, across multiple computational threads, and / or across multiple computational nodes. For example, client device 110 may be allocated resources from gaming cloud systems in both data center N (e.g., via path 1) and data center 3 (e.g., via path 2).

[0022] Once a client device 110 is assigned to a particular computational node of a corresponding computational thread of a corresponding streaming array, the client device 110 connects to the corresponding data center via a network, i.e., the client device 110 may communicate with a data center different from the data center that performs the assignment, such as data center 3.

[0023] System 100 provides games via game cloud system 190, and according to one embodiment of the present disclosure, the games are executed remotely from the client devices (e.g., thin clients) of corresponding users playing the games. System 100 can provide game control to one or more users playing one or more games via cloud gaming network or game cloud system 190 over network 150, in either single-player or multiplayer mode. In some embodiments, cloud gaming network or game cloud system 190 can include multiple virtual machines (VMs) executing on a hypervisor of a host machine, where one or more virtual machines are configured to execute game processor modules that utilize hardware resources available to the host hypervisor. In some implementations, the game processor modules can include a process emulator for executing software applications configured for one processor or operating system to be executed on a process emulator that may be configured by a different processor and / or operating system, where the process emulator may or may not be virtualized. Network 150 can include one or more communication technologies. In some embodiments, network 150 can include fifth-generation (5G) network technology with advanced wireless communication systems.

[0024] In some embodiments, communication may be facilitated using wireless technology. Such technology may include, for example, 5G wireless communication technology. 5G is the fifth generation of cellular network technology. 5G networks are digital cellular networks in which a provider's coverage area is divided into small geographic areas known as cells. Analog signals representing sound and video are digitized by the telephone, converted by an analog-to-digital converter, and transmitted as a bit stream. All 5G wireless devices within a cell communicate over electromagnetic waves with a local antenna array and low-power automatic transceiver (transmitter and receiver) within the cell via frequency channels assigned by the transceiver from a frequency pool reused by other cells. The local antennas are connected to the telephone network and the Internet by high-bandwidth optical fiber or wireless backhaul connections. As with other cellular networks, mobile devices moving from one cell to another are automatically transferred to the new cell. It should be understood that a 5G network is merely one example type of communication network, and embodiments of the present disclosure may utilize previous generations of wireless or wired communications, as well as later generations of wired or wireless technologies following 5G.

[0025] As shown, system 100, including game cloud system 190, can provide access to multiple game applications. In particular, each of the client devices may request access to a different game application from the cloud gaming network. For example, game cloud system 190 can provide one or more game servers, which can be configured as one or more virtual machines running on one or more hosts to execute the corresponding game applications. For example, a game server can manage virtual machines supporting game processors that instantiate instances of users' game applications. Thus, multiple game processors of one or more game servers associated with multiple virtual machines are configured to execute multiple instances of one or more game applications associated with the gameplay of multiple users. In this manner, the back-end server support provides streaming of gameplay media (e.g., video, audio, etc.) of the multiple game applications to the corresponding multiple users. That is, the game servers of game cloud system 190 are configured to stream data (e.g., rendered images and / or frames of the corresponding gameplay) back to the corresponding client devices over network 150. In this manner, computationally complex game applications can continue to run on the back-end server in response to controller inputs received and forwarded by the client devices. Each server is capable of rendering images and / or frames, then encoding (eg, compressing) them and streaming them to a corresponding client device for display.

[0026] In one embodiment, the cloud gaming network of the game cloud system 190 is a distributed game server system and / or architecture. Specifically, a distributed game engine, which executes game logic, is configured as a corresponding instance of a corresponding game application. Generally, a distributed game engine takes each function of the game engine and distributes those functions to be executed by multiple processing entities. Individual functions may be further distributed across one or more processing entities. The processing entities may be configured in various configurations, such as physical hardware and / or virtual components or virtual machines and / or virtual containers, where a container is distinct from a virtual machine because it virtualizes an instance of a game application running on a virtualized operating system. The processing entities may utilize and / or rely on servers and their underlying hardware on one or more servers (computing nodes) of the cloud gaming network or game cloud system 190, which may be located on one or more racks. The coordination, allocation, and management of the execution of these functions across the various processing entities is performed by a distributed synchronization layer. In this manner, the distributed synchronization layer controls the execution of these functions, enabling the generation of media (e.g., video frames, audio, etc.) for the gaming application in response to controller inputs by the player. The distributed synchronization layer allows critical game engine components / functions to be efficiently performed across distributed processing entities (e.g., via load balancing) so that they can be distributed and restructured for more efficient processing.

[0027] 2A is a diagram of multiple rack assemblies 220 including multiple computing nodes in an exemplary data center 200A of a gaming cloud system, according to one embodiment of the disclosure. The multiple data centers may be distributed around the world, for example, in North America, Europe, and Japan. In one embodiment, the rack assembly (e.g., configured for game streaming) is centered around computing nodes that run gaming applications, video games, and / or stream audio / video of game sessions to one or more clients.

[0028] Data center 200A includes multiple rack assemblies 220 (e.g., rack assemblies 220A through 220N). Each rack assembly includes a corresponding top-of-rack (TOR) switch and multiple compute threads. For example, representative rack assembly 220N includes top-of-rack switch 240N and multiple compute threads 230 (e.g., threads 230A through 230N). Other rack assemblies may be similarly configured, with or without modifications. In particular, each compute thread includes one or more compute nodes that provide hardware resources (e.g., processors, CPUs, GPUs, etc.). For example, compute thread 230N in multiple compute threads 230 of rack assembly 220N is shown as including four compute nodes, although it is understood that a rack assembly may include one or more compute nodes. Each rack assembly is coupled to a top-of-rack switch configured to provide communication with a management server 210 configured for management of the corresponding data center. The top-of-rack switches may be coupled to an optical data center fabric configured to transmit data throughout the data center. For example, rack assembly 220N is coupled to top-of-rack switch 240N, which also provides communication to an external communication network (such as the Internet).

[0029] As shown, management server 210 in data center 200A communicates with assigner 191 (shown in FIG. 1 ) to allocate resources to client devices 110. In particular, management server 210 may coordinate with an instance of gaming cloud system 190′, and coordinate with the initial instance of gaming cloud system 190 (e.g., of FIG. 1 ), to allocate resources to client devices 110. In embodiments, allocation is performed based on asset awareness, such as knowing what resources and bandwidth are needed and present in the data center. Thus, for illustrative purposes, embodiments of the present disclosure are configured to assign client devices 110 to particular compute nodes 232 of corresponding compute threads 231 of corresponding rack assembly 220B.

[0030] Additionally, management server 210 of data center 200A includes a switch controller 215 that communicates with each of a plurality of top-of-rack (TOR) switches 240. As previously described, each rack assembly may be configured with a corresponding cluster or top-of-rack switch. For example, rack assembly 220A includes top-of-rack switch 240A, rack assembly 220B includes top-of-rack switch 240B, rack assembly 220C includes top-of-rack switch 240C, and so on, with rack assembly 220N including top-of-rack switch 240N. In particular, switch controller 215 may configure each top-of-rack switch to tune to a selectable band of wavelengths, which sets the bandwidth for the spoke transceivers of the corresponding rack assembly. In this manner, each rack assembly may be reconfigured at various times depending on how the compute nodes in the corresponding rack assembly are utilized. For example, during peak gaming periods, the rack assembly may be tuned to receive wavelengths of a sufficient band to support gaming. During off-peak gaming periods (e.g., evening periods), the same rack assembly may be tuned to receive wavelengths in a different band. For example, during off-peak periods, the rack assembly may be tuned to receive less bandwidth (i.e., compared to the bandwidth received during peak periods) that is sufficient to maintain a minimum amount of connectivity for gaming. In this way, additional bandwidth (i.e., bandwidth not being used by a rack assembly) may be dynamically allocated to other rack assemblies currently experiencing greater demand, such as rack assemblies having compute nodes performing data backups or maintenance, running deep learning algorithms, etc.

[0031] 2B is a diagram of a rack assembly including a top-of-rack switch 250 configured to receive full-spectrum wavelengths as part of an optical data center fabric, where the top-of-rack switch 250 is further tuned to receive and communicate at discrete portions of the full-spectrum wavelengths (e.g., wavelengths in a selected band that sets the bandwidth of the rack assembly) in accordance with one embodiment of the present disclosure. In this manner, the optical data center fabric provides high-speed paths between the edge (e.g., super-spine switches) and leaf nodes (e.g., top-of-rack switches) of the data center using one or more passive optical power splitter devices.

[0032] In particular, rack assembly 220X may be configured to provide game streaming functionality using compute nodes executing one or more game applications. In other implementations, the compute nodes may be used to execute other types of applications. Rack assembly 220X may include network storage (not shown) configured to store game content (e.g., game logic, game applications, game data, user data, etc.) for fast access by one or more compute nodes and / or servers (280A-280N). One or more compute nodes and / or servers may be configured as a streaming array. In other implementations, the network storage is separate from rack assembly 220X, such as in distributed storage configured to store multiple game applications (e.g., a full complement of game title packages for a game cloud system). As illustrated, the diagram of FIG. 2B shows a high-level rack design for rack assembly 220X, which may represent one or more of the multiple rack assemblies 220 of FIG. 2A. For example, rack assembly 220X could represent rack assembly 220N.

[0033] Additionally, rack assembly 220X includes a top-of-rack switch 250, which may also be referred to as a leaf node or leaf node switch. Top-of-rack switch 250 is configured to receive the full-spectrum wavelengths provided within the optical data center fabric used for data networking throughout the data center and is further tuned to receive and communicate individual portions of the full-spectrum wavelengths. For example, top-of-rack switch 250 receives the full-spectrum wavelengths from one or more passive optical power splitting layers. Notably, top-of-rack switch 250 may be communicatively coupled to other compute nodes and / or servers in the same rack assembly or different rack assemblies via the optical data center fabric. For example, top-of-rack switch 250 may be communicatively connected to a communications network (e.g., the Internet) to provide network communications outside of the rack assembly and / or data center.

[0034] As shown, the top-of-rack switch 250 includes a network interface card or controller (NIC) 260 configured to provide communications (e.g., via full-spectrum wavelengths) between the rack assembly 220X and the optical data center fabric. In one embodiment, the NIC 260 includes a coherent receiver 275X configured to tune the rack assembly 220X to select a band of wavelengths from the full-spectrum wavelengths received at the top-of-rack switch (e.g., select the band of wavelengths that set the bandwidth of the top-of-rack switch 250). In one embodiment, the coherent receiver 275X is configured within a hot-pluggable device 270. For example, the hot-pluggable device may be a small form-factor pluggable (SFP) hot-pluggable network interface module 270 that includes transceivers that provide access to the optical data center fabric, and the top-of-rack switch is configured as a spoke transceiver. In another embodiment, hot-pluggable device 270 is a four small form-factor pluggable (QSFP) hot-pluggable network interface module configured as spoke transceivers with a top-of-rack switch that includes transceivers providing access to an optical data center fabric that provides high-speed data access and greater bandwidth.

[0035] In another embodiment, each of the compute nodes and / or servers 280A-280N may optionally include a corresponding coherent receiver. For example, server 280A may include coherent receiver 275A, server 280B may include coherent receiver 275B, ..., server 280N may include coherent receiver 275N. In this manner, coordination for a band of wavelengths may be performed at each compute node and / or server instead of at the top-of-rack switch 250 of the corresponding rack assembly 220X. A more detailed description of providing coordination at the compute nodes and / or servers is provided in FIG. 7.

[0036] As mentioned above, the top-of-rack switch 250 may be controlled by a management server via a control path (not shown). For example, the top-of-rack switch 250 may be dynamically adjusted to receive wavelengths of a particular band, which sets the bandwidth for the top-of-rack switch.

[0037] FIG. 3A illustrates an optical data center switching fabric 300A having point-to-multipoint optics configured between a super-spine hub optical transceiver (e.g., switching device) 301A and multiple leaf nodes 320 (e.g., spoke transceivers) in accordance with one embodiment of the present disclosure. The leaf nodes may be top-of-rack switches in multiple rack assemblies in a data center. As shown, the super-spine hub optical transceiver 301A is communicatively coupled to each of the multiple leaf nodes 320 (e.g., leaf switches or top-of-rack switches) such that the optical data center switching fabric 300A effectively provides high-speed paths between the edge of the data center (i.e., the hub optical transceiver 301A) and the leaf nodes served by the super-spine hub optical transceiver 301A using flexible optical technology (e.g., passive optical power splitters).

[0038] In particular, the optical data center switching fabric 300A transmits data to leaf nodes using full-spectrum wavelengths instead of dividing the full-spectrum wavelengths by frequency. This can be achieved by using passive optical devices, such as optical power splitters, in one or more intermediate layers (not shown) instead of using switching devices, as further described below in connection with, for example, FIGS. 4A-4B. For example, the super-spine hub optical transceiver 301A may be configured with multiple ports (e.g., ports longer than two lanes) for transmitting and receiving data. Depending on the number of leaf nodes, one or more intermediate layers can be inserted between the super-spine hub optical transceiver 301A and multiple leaf nodes 320. That is, as the number of leaf switches increases, one lane coming from the super-spine hub optical transceiver 301A can be coupled to one or more layers of passive optical power splitters to transmit the full-spectrum wavelengths output by the super-spine hub optical transceiver 301A to two or more leaf nodes.

[0039] In some embodiments, amplifiers may be implemented between the super spine hub optical transceiver 301A and the plurality of leaf nodes 320 to improve the signal-to-noise ratio of the transmitted signal, although amplification is not expected to be necessary given the length of data paths typically encountered inside a data center. That is, the existing power from the optics inside the optical data center switching fabric 300A should be high enough in a data center implementation so that amplifiers are not required, although an amplification stage can be implemented if needed or in some other use cases, such as providing amplification at the end of a hub.

[0040] As shown, the switch controller 215A may be configured within the super-spine hub optical transceiver 301A. The switch controller 215A may be communicatively coupled to each of the plurality of leaf nodes 320 via control path 350. In this manner, the switch controller 215A can coordinate each leaf node to receive and / or process a band of wavelengths from the full spectrum of wavelengths transmitted from the super-spine hub optical transceiver 301A within the optical data center switching fabric 300A. In some cases, the switch controller 215A may be implemented within the management server 210 as shown in FIG. 2A, which may be configured within the super-spine hub optical transceiver 301A or may be located remotely from the hub optical transceiver 301A, as described above. For example, a coherent receiver at a corresponding leaf node may be tuned to receive and / or process wavelengths in a band that sets the bandwidth for that leaf node.

[0041] The optical data center switching fabric 300A includes a single super-spine hub optical transceiver 301A acting as a hub in a hub and multiple spoke configuration, where the spokes are leaf nodes. The super-spine hub optical transceiver 301A can provide data over a full spectrum of wavelengths. Purely by way of example, the full spectrum can transmit 400 gigabits per second (400G) at selectable wavelengths having specific data rates (e.g., 16, 32, or 64 gigabits per second). The full spectrum wavelengths can be further partitioned into subcarriers (e.g., 25G or any other number of subcarrier wavelengths) by various multiplexing techniques. In one embodiment, the full spectrum wavelengths are partitioned into 25G subcarrier wavelengths. In this manner, each of the spoke transceivers can be tuned to receive one or more 25G subcarrier wavelengths. For example, a spoke transceiver may be tuned to receive a 25G subcarrier wavelength, or a 50G subcarrier wavelength, or a 75G subcarrier wavelength, or a 100G subcarrier wavelength, or any increment or multiple of a 25G subcarrier wavelength.

[0042] 3B illustrates an optical data center switching fabric 300B with multipoint optics and configured with one or more superspines, according to one embodiment of the present disclosure. Each superspine is communicatively coupled to multiple leaf nodes (e.g., spoke transceivers or top-of-rack switches) 320. For example, each of the superspine hub optical transceivers (e.g., switching devices) 301A and 301B is communicatively coupled to multiple leaf nodes 320. The leaf nodes may be top-of-rack switches in multiple rack assemblies in a data center. In this manner, the optical data center switching fabric 300B effectively provides high-speed paths between the edge of the data center (i.e., the hub optical transceivers 301A and 301B) and the leaf nodes, which are provided using flexible optical technology (e.g., passive optical power splitters).

[0043] The coupling between each of the super-spine hub optical transceivers 301A and 301B and the plurality of leaf nodes 320 in the optical data center switching fabric 300B is similar to the coupling between the super-spine hub optical transceiver 301A and the plurality of leaf nodes 320 described in Figure 3A. In this manner, the optical data center switching fabric 300B of Figure 3B can provide approximately twice the operating capacity (e.g., throughput, bandwidth, etc.) of the optical data center switching fabric 300A of Figure 3A when data is transmitted using both the super-spine optical transceivers 301A and 301B. Purely by way of example, if the optical data center switching fabric 300A of FIG. 3A provides a full spectrum of wavelengths transmitted at 400 gigabits per second (i.e., 400G) using the super-spine hub optical transceiver 301A, then the optical data center switching fabric 300B of FIG. 3B can be configured to provide a full spectrum of wavelengths transmitted at 800 gigabits per second (i.e., 800G) using both the super-spine hub optical transceivers 301A and 301B. Also, the optical data center switching fabric 300B of FIG. 3B can provide the same operating capacity as the optical data center switching fabric 300A of FIG. 3A, but can also provide backup or failover services. That is, if the delivery paths of data through one or more ports of the super-spine hub optical transceiver 301A become faulty, those data paths can be restored using one or more ports of the super-spine hub optical transceiver 301B.

[0044] As shown, the switch controller 215B may be configured within the super-spine hub optical transceiver 301B. The switch controller 215B may be communicatively coupled to each of the plurality of leaf nodes 320 via control path 355. In this manner, the switch controller 215B may tune each of the leaf nodes to receive and / or process a band of wavelengths from the full spectrum of wavelengths transmitted from the super-spine hub optical transceiver 301B within the optical data center switching fabric 300B. In some cases, the switch controller 215B may be implemented within the management server 210 as shown in FIG. 2A, which may be configured within the super-spine hub optical transceiver 301B or may be located remotely from the hub optical transceiver 301B, as previously described. For example, a coherent receiver at a corresponding leaf node may be tuned to receive and / or process a band of wavelengths that establishes a bandwidth for that leaf node. Coordination of leaf nodes is performed in coordination between switch controllers 215A and 215B of super spine hub local transceivers 301A and 301B.

[0045] 4A illustrates an optical data center switching fabric 400 configured with point-to-multipoint optics, according to one embodiment of the present disclosure. In particular, the optical data center switching fabric uses point-to-multipoint optics to directly communicatively couple a hub optical transceiver 410 to multiple spoke optical transceivers 430 (e.g., leaf switches) through one or more intermediate layers containing one or more passive optical devices.

[0046] As shown, the optical data center switching fabric 400 includes hub optical transceivers 410 (e.g., superspine switches) configured to receive a spectrum of wavelengths, also referred to as full spectrum wavelengths, processed through the data center. The full spectrum wavelengths 420 deliver aggregate bandwidth for data distribution within the optical data center switching fabric 400. Purely for purposes of illustration, the full spectrum may be transmitted at selectable data rates in aggregates of 400 gigabits per second (e.g., 400G) or 800 gigabits per second (800G). As previously mentioned, the optical devices in the optical data center switching fabric 400 may be further partitioned into subcarriers, such as wavelengths and / or channels in multiples of 25G using various multiplexing techniques. For example, as shown in FIG. 4A, the full spectrum wavelengths 420 may be divided into 32 distinct 25G wavelengths or wavelength bands.

[0047] The hub optical transceiver 410 may be communicatively coupled to an optical power splitter 460. That is, the point-to-multipoint optics of the optical data center switching fabric 400 can be separated using optical power splitters, which reduces latency throughout the data center by eliminating one or more switching layers, including conventional switches. The reduced latency is achieved when the optical power splitter is a purely passive optical device that adds substantially no latency, whereas conventional switches perform optical-electrical-optical (OEO) conversion along with some routing and / or switching operations (e.g., packet queuing) that introduce latency. Furthermore, optical power splitters can be less complex and less costly than switches used in intermediate layers found in conventional switching fabrics. For example, in some embodiments, any standard optical power splitter can be used within the optical data center switching fabric 400. The power splitter 460 operates as a passive device configured to replicate the spectral wavelength 420 and output multiple replicated spectral wavelengths 420′. For example, the plurality of replicated spectral wavelengths 420' includes replicated spectral wavelength 420A, replicated spectral wavelength 420B, . . . , and replicated spectral wavelength 420N.

[0048] In an embodiment, each replicated spectrum wavelength comprises a full spectrum wavelength 420 (e.g., an 800G wavelength) transmitted by the hub optical transceiver 410, but at a lower power. That is, each replicated spectrum wavelength of the plurality of replicated spectrum wavelengths 420′ has a corresponding power that is a fraction of the total power received from the hub optical transceiver 410 by the optical power splitter 460. In one embodiment, the power splitter divides the total power, or the power received at the full spectrum or replicated full spectrum wavelengths, equally or unevenly. For example, the power splitter can divide the total power of the full spectrum wavelengths equally or unevenly among the plurality of replicated spectrum wavelengths.

[0049] The optical data center switching fabric 400 includes multiple spoke transceivers 430 coupled to a power splitter. For example, each spoke transceiver is coupled to the power splitter 460 via a separate connection (e.g., fiber optic cable). In one embodiment, each spoke transceiver is a leaf switch or leaf node, such as a top-of-rack switch, in each rack assembly of the corresponding data center. More specifically, each of the multiple spoke transceivers is configured to receive a corresponding one of multiple replicated wavelengths. That is, each spoke transceiver receives a full-spectrum wavelength, and the optical data center switching fabric 400 transmits data to the spoke transceivers using the full-spectrum wavelength instead of dividing the full-spectrum wavelength by frequency. In this way, the point-to-multipoint optics used between the hub optical transceiver 410 and each spoke optical transceiver splits or breaks the power but still transports the full-spectrum wavelength. For example, spoke optical transceiver 430A is configured to receive wavelength 420A of the replicated spectrum, spoke optical transceiver 430B is configured to receive wavelength 420B of the replicated spectrum, ..., spoke optical transceiver 430N is configured to receive wavelength 420N of the replicated spectrum.

[0050] The plurality of spoke optical transceivers 430 are adjustable to set multiple bandwidths based on the aggregate bandwidth provided by the full spectrum wavelengths 420. More specifically, each spoke transceiver is adjustable to select a band of wavelengths (i.e., select a discrete portion of the corresponding full spectrum wavelengths to be replicated) that sets the bandwidth for the corresponding spoke transceiver. For example, each spoke transceiver may include a leaf switch or a top-of-rack switch deployed within a corresponding rack assembly that serves multiple hosts and / or servers. This can be achieved by using tunable coherent receivers in the spoke optical transceivers. In particular, each spoke transceiver includes an optical coherent receiver configured to split the replicated spectrum wavelengths (e.g., full spectrum wavelengths) into selectable wavelength bands, and each spoke transceiver can be dynamically adjusted to receive the selectable wavelength bands that define the bandwidth of the corresponding spoke transceiver. For example, each spoke transceiver includes a coherent receiver configured to select a band of wavelengths from a corresponding one of the plurality of replicated spectrum wavelengths received by the spoke transceiver.

[0051] In particular, each spoke transceiver can be tuned to at least one subcarrier wavelength (e.g., a 25G wavelength), but can be tuned to multiple subcarriers to increase bandwidth through individual connections, such as through corresponding spoke transceivers (e.g., top-of-rack switches). For example, spoke optical transceiver 430A includes a coherent receiver 440A tuned to receive two 25G subcarrier wavelengths for a total of 50G of bandwidth, spoke optical transceiver 430B includes a coherent receiver 440B tuned to receive one 25G subcarrier wavelength for a total of 25G of bandwidth,..., spoke optical transceiver 430N includes a coherent receiver 440N tuned to receive four 25G subcarrier wavelengths for a total of 100G of bandwidth. Additionally, another spoke optical transceiver may include a coherent receiver tuned to receive subcarrier wavelengths in any multiple of 25G (e.g., 25G, 50G, 75G, 100G, 125G, etc.), or may be tuned to receive subcarrier wavelengths in any multiple of greater than or less than 25G. That is, some components may be tuned to select increments that are multiples of 25G, while other components may be tuned to select larger or smaller increments based on current technology or communication generation within the data center.

[0052] The optical data center switching fabric 400 achieves improved performance by eliminating intermediate layers of switches, thereby reducing latency between the hub optical transceiver 410 and each of the multiple spoke optical transceivers 430. One or more switches in one or more intermediate layers of switches are replaced with point-to-multipoint optics, which are passive devices that require no input power for operation. In one embodiment, the point-to-multipoint optics include passive optical power splitters. Depending on the number of spoke optical transceivers required within the data center, one or more intermediate layers of passive optical power splitters can be implemented, with each layer including one or more power splitters. In one embodiment, the multiple layers can include cascaded layers of power splitters.

[0053] In one embodiment, each of the hub optical transceivers 410, the optical power splitter 460, and each of the plurality of spoke optical transceivers 430 may be hot-pluggable. For example, as described above, each device may be contained within a QSFP form factor. In one implementation, one or more hub optical transceivers and one or more optical power splitters (i.e., configured in one or more layers or cascaded layers) may be installed within a rack assembly. As described above, each of the spoke optical transceivers may be installed in a corresponding rack assembly as a corresponding top-of-rack switch.

[0054] In one embodiment, a control system is coupled to the coherent receivers of the multiple spoke transceivers, the control system configured to coordinate the coherent receiver of each spoke transceiver to select a corresponding wavelength band. For example, the control system may include a switch controller 215C configured within the hub optical transceiver 410. The switch controller 215C may be communicatively coupled to each of the multiple spoke optical transceivers 430 via the control path 450. In this manner, the switch controller 215C may coordinate each of the spoke optical transceivers to receive and / or process a band of wavelengths from the full spectrum of wavelengths transmitted from the hub optical transceiver 410 within the optical data center switching fabric 400. In some cases, the switch controller 215C may be implemented within the management server 210 as shown in FIG. 2A, which may be configured within the hub optical transceiver 410 or may be located remotely from the hub optical transceiver 410, as described above. For example, a coherent receiver at a corresponding spoke optical transceiver may be tuned to receive and / or process a band of wavelengths that establishes the bandwidth for that spoke optical transceiver.

[0055] In one embodiment, the optical data center switching fabric 400 can also dynamically allocate bandwidth to the spoke optical transceivers. As traffic requirements change over time, the optical data center switching fabric 400 can be configured to increase and / or decrease capacity as needed. For example, an optical data center switching fabric 400 used for a large business system may experience heavy usage during business hours, and therefore may allocate more bandwidth to selected rack assemblies and their corresponding hosts / servers during business or peak hours. During off-peak or nighttime hours, the bandwidth allocated to each rack assembly may be dynamically adjusted to enable dynamic allocation of bandwidth between rack assemblies performing potentially more or less demanding operations. For example, bandwidth can be adjusted to be allocated and / or relocated to different servers / network segments in different rack assemblies for purposes of data backup, management operations, running machine learning algorithms, or running other off-peak applications. Furthermore, software-defined networking (SDN) can be configured to automate bandwidth allocation so that bandwidth can be dynamically and automatically allocated as requirements dictate.

[0056] In some embodiments, an amplifier may be implemented between the hub optical transceiver 410 and the plurality of spoke optical transceivers 430 to improve the signal-to-noise ratio of the transmitted signal, although it is anticipated that amplification is not necessary given the length of data paths typically encountered within a data center. That is, an amplifier may be coupled between a spoke transceiver and the power splitter 460, the amplifier configured to amplify the power of the wavelengths of the replicated spectrum received by the spoke transceiver. An amplifier may also be coupled between the hub optical transceiver 410 and the power splitter 460, the amplifier configured to amplify the power of the wavelengths of the full spectrum received by the power splitter 460.

[0057] Figure 4B illustrates the optical data center switching fabric 400 introduced in Figure 4A configured with point-to-multipoint optics, according to one embodiment of the present disclosure. The switching fabrics 400 illustrated in Figures 4A and 4B are identically configured. That is, the optical data center switching fabric 400 uses point-to-multipoint optics to directly communicatively couple a hub optical transceiver 410 to multiple spoke optical transceivers 430 (e.g., leaf switches) through one or more intermediate layers containing one or more passive optical devices.

[0058] More specifically, the optical data center switching fabric 400 shown in Figure 4B illustrates the connections between the components. As shown, the connections between the hub optical transceiver 410 and the optical power splitter 460, as well as the connections between the power splitter 460 and each of the multiple spoke optical transceivers 430, can be implemented by optical fiber cabling, which can be single or dual optical fiber cabling. In single optical fiber cabling, signals flow in one direction (i.e., upstream or downstream). In dual optical fiber cabling, signals can flow in both directions (i.e., upstream and downstream).

[0059] In one embodiment, the connection between hub optical transceiver 410 and optical power splitter 460 is a single fiber optic wiring, thus requiring separate wiring for upstream and downstream data traffic. In another embodiment, the connection between hub optical transceiver 410 and optical power splitter 460 is a dual fiber optic wiring to handle both upstream and downstream data traffic.

[0060] In one embodiment, the connection between the power splitter 460 and the corresponding spoke optical transceiver is a single fiber optic wiring, thus requiring separate wiring for upstream and downstream data traffic. In another embodiment, the connection between the power splitter 460 and the corresponding spoke optical transceiver is a dual fiber optic wiring to handle both upstream and downstream data traffic.

[0061] 5A-5B illustrate examples of optical splitting in a data center environment. Depending on the data center's objectives, capacity, and variations in bandwidth provision per rack assembly, multiple physical optical fiber infrastructure topologies can be used in various embodiments. For illustrative purposes only, several examples are shown in FIGS. 5A-5B , and it is understood that the topologies are not limited to these examples. Regardless of the topology implemented, the topology for the optical data center switching fabric of the disclosed embodiments allows for maximum flexibility with a minimum number of optical fibers, which means significantly less fiber is used and less active switching infrastructure (i.e., fewer passive optical devices are used compared to the number of active switches in the intermediate tier(s)) compared to traditional switching fabric designs (e.g., using intermediate switching tier(s)) for the same number of rack assemblies within a data center.

[0062] 5A illustrates a data center optical switching fabric 500A including a single power splitter layer, according to one embodiment of the present disclosure. The optical data center switching fabric 500A uses point-to-multipoint optics to directly communicatively connect a superspine to multiple spoke optical transceivers 550 located in multiple rack assemblies 530. As illustrated, the optical switching fabric 500A includes a hub optical transceiver, a power splitter layer, and multiple spoke optical transceivers 550.

[0063] In particular, the optical switching fabric 500A includes super-spine switching devices (e.g., switches, hub optical transceivers, etc.) configured to receive a spectrum of wavelengths (e.g., full spectrum wavelengths) 420 that deliver a total bandwidth for data distribution, where the full spectrum of wavelengths 420 has a total power. The super-spine switching devices may be configured within a super-spine rack assembly 501A that also includes passive optical devices.

[0064] Additionally, the optical switching fabric 500A includes a power splitting layer coupled to the superspine in the rack assembly 501A (e.g., a hub optical transceiver). The power splitting layer is configured to output multiple replicated wavelengths of the spectrum, each having a corresponding power that is a fraction of the total power received from the superspine 501A. In one embodiment, the power splitting layer includes multiple power splitters arranged in one or more cascaded layers, each power splitter configured to operate as a passive device and replicate a wavelength of the spectrum received at a reduced power. The power splitting layer may also be included within the rack assembly 501A.

[0065] As shown in FIG. 5A, the power splitting layer includes one layer, and more specifically, one power splitter 460A configured as a 1:16 (1*16) power splitter. That is, the full spectrum of wavelengths is split into 16 replicated spectrum wavelengths. As described above, each of the replicated spectrum wavelengths has a corresponding power that is a fraction of the total power received from the superspine switching device. In one embodiment, the fractions associated with the replicated spectrum wavelengths are uniformly distributed. In another embodiment, the fractions associated with the replicated spectrum wavelengths are non-uniformly distributed.

[0066] As shown, power splitter 460 has multiple outputs (e.g., channels) 540A-540P, e.g., 16 channels, each having a corresponding duplicate wavelength of the spectrum. The outputs are coupled to multiple top-of-rack switches 550 in multiple rack assemblies 530. For example, the output of channel 540A from power splitter 460A is coupled to top-of-rack switch 550A of rack assembly 530A, channel 540B is coupled to top-of-rack switch 550B of rack assembly 530B, channel 540C is coupled to top-of-rack switch 550C of rack assembly 530C, channel 540D is coupled to top-of-rack switch 550D of rack assembly 530D, channel 540E is coupled to top-of-rack switch 550E of rack assembly 530E, channel 540F is coupled to top-of-rack switch 550F of rack assembly 530F, channel 540G is coupled to top-of-rack switch 550G of rack assembly 530G, and channel 540H is coupled to top-of-rack switch 550F of rack assembly 530G. channel 540M is coupled to top-of-rack switch 550M of rack assembly 530M; channel 540O is coupled to top-of-rack switch 550O of rack assembly 530O; and channel 540P is coupled to top-of-rack switch 550P of rack assembly 530P. In this manner, one super-spine switch port can be connected to 16 leaf switches or top-of-rack switches.Depending on how many power splitting layers are implemented, one switch port can be connected to any number of top-of-rack switches.

[0067] In particular, the optical data center switching fabric 500A includes a plurality of spoke transceivers (e.g., top-of-rack switches) coupled to a power splitting layer. Each of the plurality of spoke transceivers is configured to receive a corresponding one of the wavelengths of the plurality of replicated spectrums, with each spoke transceiver being tunable to select a band of wavelengths that sets the bandwidth for the respective spoke transceiver. As shown in FIG. 5A, a plurality of top-of-rack switches 530 are coupled to the power splitting layer or power splitter 460A. As previously described, each top-of-rack switch receives the wavelengths of the replicated spectrum and can be further tuned to select a band of wavelengths that sets the bandwidth of the corresponding top-of-rack switch. This is achieved using a corresponding coherent optical receiver at each top-of-rack switch configured to split the wavelengths of the replicated spectrum into selectable wavelength bands that define the bandwidth of the corresponding top-of-rack switch.

[0068] As previously mentioned, a control system is coupled to the coherent receivers of the plurality of top-of-rack switches 530, the control system being configured to adjust each top-of-rack switch to select a corresponding band of wavelengths. Although the control system is not shown in FIG. 5A, the control system can include a switch controller communicatively coupled to each coherent receiver to adjust each of the top-of-rack switches to receive a corresponding band of wavelengths from the full spectrum of wavelengths 420.

[0069] 5B illustrates a data center optical switching fabric 500B including multiple cascaded splitter layers according to one embodiment of the present disclosure. The optical data center switching fabric 500B uses point-to-multipoint optics to directly communicatively connect a superspine to multiple spoke optical transceivers 560 located in multiple rack assemblies 530. As shown, the optical switching fabric 500B includes a hub optical transceiver, cascaded power splitting layers, and multiple spoke optical transceivers 550.

[0070] In particular, the optical switching fabric 500B includes super-spine switching devices (e.g., switches, hub optical transceivers, etc.) configured to receive a spectrum of wavelengths (e.g., full spectrum wavelengths) 420 that deliver a total bandwidth for data distribution, where the full spectrum of wavelengths 420 has a total power. The super-spine switching devices may be configured within a super-spine rack assembly 501B that also includes passive optical devices in cascaded power splitting layers.

[0071] In particular, optical switching fabric 500A includes cascaded power splitting layers configured to output a plurality of replicated wavelengths of the spectrum, each having a corresponding power that is a fraction of the total power received from the superspine of rack assembly 501B. In one embodiment, the cascaded power splitting layers include a plurality of power splitters arranged in one or more cascaded layers, each power splitter configured to operate as a passive device and replicate a wavelength of the spectrum received at a reduced power. Cascaded power splitting layers may also be included within rack assembly 501B.

[0072] As shown, the cascaded power splitting layer includes a first layer including one power splitter 460B configured as a one-to-four (1*4) power splitter. That is, the full spectrum of wavelengths is split into four replicated spectrum wavelengths. For example, power splitter 460 may provide one replicated spectrum wavelength as an output across channel 560, another replicated spectrum wavelength as another output across channel 561, another replicated spectrum wavelength as another output across channel 562, and another replicated spectrum wavelength as another output across channel 563. As described above, each of the replicated spectrum wavelengths is provided as an output from power splitter 460B with a corresponding power that is a fraction of the total power received from the super-spine switching device. In one embodiment, the portions associated with the replicated spectrum wavelengths are uniformly distributed. In another embodiment, the portions associated with the replicated spectrum wavelengths are non-uniformly distributed.

[0073] In particular, the optical data center switching fabric 500B includes a plurality of spoke transceivers (e.g., top-of-rack switches) coupled to a cascaded power splitting layer. Each of the plurality of spoke transceivers is configured to receive a corresponding one of the wavelengths of the plurality of replicated spectrums, and each spoke transceiver is tunable to select a band of wavelengths that sets the bandwidth for the respective spoke transceiver. As shown in FIG. 5B, a plurality of top-of-rack switches 530 are coupled to the cascaded power splitting layer. As previously described, each top-of-rack switch receives the wavelengths of the replicated spectrum and can be further tuned to select a band of wavelengths that sets the bandwidth of the corresponding top-of-rack switch. This is achieved using a corresponding coherent optical receiver at each top-of-rack switch configured to split the wavelengths of the replicated spectrum into selectable wavelength bands that define the bandwidth of the corresponding top-of-rack switch.

[0074] As shown, the cascaded power splitting layer is coupled to multiple top-of-rack switches 550 in multiple rack assemblies 530. In particular, each output from power splitter 460B is coupled to a corresponding top-of-rack switch via a corresponding power splitter. For example, the output via channel 560 is coupled to power splitter 460C configured as a one-to-four (1*4) power splitter, the output via channel 561 is coupled to power splitter 460D configured as a one-to-four (1*4) power splitter, the output via channel 562 is coupled to power splitter 460E configured as a one-to-four (1*4) power splitter, and the output via channel 563 is coupled to power splitter 460F configured as a one-to-four (1*4) power splitter. Each of the outputs provides a replicated wavelength of the spectrum.

[0075] Additionally, each of power splitters 460C, 460D, 460E, and 460F is coupled to a plurality of top-of-rack switches in a corresponding rack assembly. As shown, power splitters 460C, 460D, 460E, and 460F may be installed in a corresponding rack assembly, although the power splitters may be located elsewhere, such as in another rack assembly. In particular, power splitter 460C provides four outputs via channels 560A, 560B, 560C, and 560D. Each channel is further communicatively connected to a corresponding top-of-rack switch. For example, from power splitter 460C, channel 560A is coupled to top-of-rack switch 550A of rack assembly 530A, channel 560B is coupled to top-of-rack switch 550B of rack assembly 530B, channel 560C is coupled to top-of-rack switch 550C of rack assembly 530C, and channel 560D is coupled to top-of-rack switch 550D of rack assembly 530D. Also, from power splitter 460D, channel 561A is coupled to top-of-rack switch 550E of rack assembly 530E, channel 561B is coupled to top-of-rack switch 550F of rack assembly 530F, channel 561C is coupled to top-of-rack switch 550G of rack assembly 530G, and channel 561D is coupled to top-of-rack switch 550H of rack assembly 530H. Also, from power splitter 460E, channel 562A is coupled to top-of-rack switch 550I of rack assembly 530I, channel 562B is coupled to top-of-rack switch 550J of rack assembly 530J, channel 562C is coupled to top-of-rack switch 550K of rack assembly 530K, and channel 562D is coupled to top-of-rack switch 550L of rack assembly 530L.Also, from power splitter 460F, channel 563A is coupled to top-of-rack switch 550M of rack assembly 530M, channel 563B is coupled to top-of-rack switch 550N of rack assembly 530N, channel 563C is coupled to top-of-rack switch 550O of rack assembly 530O, and channel 563D is coupled to top-of-rack switch 550P of rack assembly 530P.

[0076] As previously mentioned, a control system is coupled to the coherent receivers of the plurality of top-of-rack switches 530, the control system being configured to adjust each top-of-rack switch to select a corresponding band of wavelengths. Although the control system is not shown in FIG. 5B, the control system can include a switch controller communicatively coupled to each coherent receiver to adjust each of the top-of-rack switches to receive a corresponding band of wavelengths from the full spectrum of wavelengths 420.

[0077] 6A illustrates a data center switching fabric 600A that includes the use of 1×4 (1*4) splitters across three cascaded power splitter layers connected to a superspine switching layer 610, in accordance with one embodiment of the present disclosure. In particular, the optical data center switching fabric 600A uses point-to-multipoint optics to directly communicatively couple the superspine switching layer (e.g., hub optical transceivers) to multiple spoke optical transceivers (e.g., leaf switches, top-of-rack switches) via one or more intermediate layers that include one or more passive optical devices.

[0078] As shown, the optical data center switching fabric 600A includes a superspine switching layer 610 configured to receive a spectrum of wavelengths, also referred to as full spectrum wavelengths, processed through the data center. The full spectrum wavelengths deliver aggregate bandwidth for data distribution within the optical data center switching fabric 600A. Purely for purposes of illustration, the full spectrum may be transmitted at selectable data rates in aggregates of 400 gigabits per second (e.g., 400G) or 800 gigabits per second (800G). As previously mentioned, the optical devices of the optical data center switching fabric 400 may be further partitioned into subcarriers, such as wavelengths in multiples of 25G, using various multiplexing techniques. For example, the full spectrum wavelengths may be divided into 32 different 25G wavelengths or wavelength bands.

[0079] The optical switching fabric 600A includes three cascaded power splitter layers coupled to a superspine switching layer 610. For example, the cascaded power splitter layers and the superspine switching layer can be located within a single rack assembly in a data center. The three cascaded splitter layers include a first cascade stage power splitter, a second cascade stage power splitter, and a third cascade stage power splitter. The three cascaded power splitter layers are configured to output multiple replicated wavelengths of the spectrum, each having a corresponding power that is a fraction of the total power received from the superspine switching layer 610. In one embodiment, the cascaded power splitter layer includes multiple power splitters arranged in one or more cascaded layers, each power splitter operating as a passive device and configured to replicate a wavelength of the spectrum it receives at a reduced power.

[0080] In particular, the first cascade stage includes a single 1×4 (1*4) power splitter 460G that provides four outputs through the channel, each of which provides a replicated spectrum of wavelengths having a power that is a fraction of the total power for the full spectrum of wavelengths provided by the superspine switching layer 610.

[0081] The second cascade stage includes four power splitters, each a 1x4 (1*4) power splitter, coupled to power splitter 460G in the first cascade stage. For example, power splitter 460H receives wavelengths of the replicated spectrum via channels from power splitter 460G in the first cascade stage and provides four additional outputs via channels, each providing a wavelength of the replicated spectrum. Power splitter 460I receives wavelengths of the replicated spectrum via channels from power splitter 460G in the first cascade stage and provides four additional outputs via channels, each providing a wavelength of the replicated spectrum. Power splitter 460J receives wavelengths of the replicated spectrum via channels from power splitter 460G in the first cascade stage and provides four additional outputs via channels, each providing a wavelength of the replicated spectrum. Additionally, power splitter 460K receives wavelengths of the replicated spectrum via channels from power splitter 460G in the first cascade stage and provides four additional outputs via channels, each output providing a wavelength of the replicated spectrum.

[0082] The third cascade stage includes 16 power splitters 460L, 460M, 460N, 460O, ..., 460X. Each power splitter is similarly configured, and each power splitter is coupled to an output from one of the power splitters in the third cascade stage. As a representative example, power splitter 460H from the second cascade stage provides four outputs coupled to power splitters 460L, 460M, 460N, and 460O. For example, each of power splitters 460L through 460O receives a corresponding wavelength of the replicated spectrum from power splitter 460H via a corresponding channel.

[0083] The three cascaded power splitter layers are coupled to a plurality of spoke optical transceivers, e.g., top-of-rack switches of a plurality of rack assemblies. Each of the plurality of spoke transceivers is configured to receive a corresponding replicated wavelength of the spectrum, where each spoke transceiver receives a full-spectrum wavelength but at a reduced power from the total power output by the superspine switching layer 610, which provides the full-spectrum wavelengths, as described above. In this manner, each spoke transceiver is configured to split the replicated wavelengths (e.g., full-spectrum wavelengths) into a selectable band of wavelengths using a corresponding coherent optical receiver. That is, each spoke transceiver can be dynamically adjusted to receive a selectable band of wavelengths that defines the bandwidth of the corresponding spoke transceiver.

[0084] 6B illustrates a physical data center layout implementing the optical data center switching fabric 600A of FIG. 6A, including multiple cascaded power splitter layers, in accordance with one embodiment of the present disclosure. The optical data center switching fabric 600B uses point-to-multipoint optics to directly communicatively connect a super-spine switching layer to multiple spoke optical transceivers 550 located in multiple rack assemblies 530. As illustrated, the optical switching fabric 600B includes a super-spine switching layer 690 (e.g., spine switches, hub optical transceivers, etc.), a cascaded power splitting layer, and multiple spoke optical transceivers 550.

[0085] The superspine switching layer (e.g., switches, hub optical transceivers, etc.) is configured to receive wavelengths of a spectrum (e.g., full spectrum wavelengths) that deliver a total bandwidth for data distribution, the full spectrum wavelengths having a total power.

[0086] The optical switching fabric 600B includes a cascaded power splitting layer configured to output a plurality of replicated wavelengths of the spectrum, each replicated wavelength of the spectrum having a corresponding power that is a fraction of the total power received from the superspine switching layer. In one embodiment, the cascaded power splitting layer includes a plurality of power splitters arranged in one or more cascaded layers, each power splitter operating as a passive device and configured to replicate a wavelength of the spectrum received at a reduced power.

[0087] The first cascade stage 631 includes one power splitter 460P configured as a one-to-four (1*4) power splitter. That is, the full spectrum of wavelengths is split into four replicated spectrum wavelengths. For example, the power splitter 460P may provide outputs across four channels, each channel having a replicated spectrum of wavelengths. As described above, each of the replicated spectrum of wavelengths is provided as an output from the power splitter 460P with a corresponding power that is a fraction of the total power received from the superspine switching layer.

[0088] The second cascade stage 632 of the power splitter includes four power splitters, each configured as a one-to-four (1*4) power splitter. For example, the second cascade stage 632 includes power splitters 460R, 460S, 460T, and 460U, each coupled to power splitter 460P in the first cascade stage 631. Each power splitter is coupled to a top-of-rack switch in a corresponding row of the rack assembly.

[0089] For example, power splitter 460R receives wavelengths of the spectrum from power splitter 460P via corresponding channels and provides additional wavelengths of the spectrum at significantly less power via four corresponding channels to each of top-of-rack switches 550A, 550B, 550C, and 550D in the first row of rack assembly 551. Power splitter 460S also receives wavelengths of the spectrum from power splitter 460P via corresponding channels and provides additional wavelengths of the spectrum at significantly less power via four corresponding channels to each of top-of-rack switches 550E, 550F, 550G, and 550H in the second row of rack assembly 552. Power splitter 460T also receives wavelengths of the spectrum from power splitter 460P via corresponding channels and provides additional wavelengths of the spectrum at significantly less power via four corresponding channels to each of top-of-rack switches 550I, 550J, 550K, and 550L in the third row of rack assembly 553. Power splitter 460U also receives wavelengths of the spectrum from power splitter 460P via corresponding channels and provides additional wavelengths of the spectrum at significantly less power via four corresponding channels to each of top-of-rack switches 550M, 550N, 550O, and 550P in the fourth row of rack assembly 554.

[0090] As previously described, each of the plurality of spoke transceivers 530 (e.g., top-of-rack switches) is configured to receive a corresponding one of the wavelengths of the plurality of replicated spectrums, with each spoke transceiver being tunable to select a band of wavelengths that sets the bandwidth for each spoke transceiver. As shown in FIG. 6B, the plurality of top-of-rack switches 530 are coupled to a cascaded power splitting layer. As previously described, each top-of-rack switch receives the wavelengths of the replicated spectrum and can be further tuned to select a band of wavelengths that sets the bandwidth of the corresponding top-of-rack switch. This is achieved using a corresponding coherent optical receiver at each top-of-rack switch configured to split the wavelengths of the replicated spectrum into selectable bands of wavelengths that define the bandwidth of the corresponding top-of-rack switch.

[0091] As shown in Figure 6B, the optical data center switching fabric can be implemented using repeatable steps. Notably, the physical connectivity of each rack assembly remains the same, such as the power connections and fiber cables that connect to the optical data center switching fabric. In this way, the bandwidth provided can be dynamically adjusted for each rack type as it is added to the optical data center fabric. This reduces the number of optical connections required, simplifying the data center architecture and the physical provisioning and cabling process.

[0092] In one embodiment, the first cascade stage 631 and the second cascade stage 632 of one or more power splitters may be located within an overhead or under-floor cable tray 620. Fiber optic cables connecting the two cascade stages 631 and 632 to the top-of-rack switches in each row of rack assemblies 551-554 may be routed through the cable tray 620. In particular, one consistent fiber cable drop may be provided for each rack assembly in the row of rack assemblies 551-554 upon installation. As previously mentioned, control of each rack assembly provides dynamic adjustment of the bandwidth received at each corresponding rack assembly, so that each rack assembly can be dynamically configured to receive selectable bandwidth at any time.

[0093] FIG. 7 illustrates a data center switching fabric configured to perform switching at the host / server rack layer, according to one embodiment of the present disclosure. The optical data center switching fabric uses point-to-multipoint optics to directly communicatively connect a superspine switching layer to multiple spoke optical transceivers located in multiple rack assemblies. In particular, data distribution is illustrated in rack assembly 700, where instead of providing coherent receivers in the top-of-rack switches of the corresponding rack assemblies, the coherent receivers can be moved further down the distribution stack from the switching fabric providing data to the rack assemblies to the hosts / servers generating traffic within one or more rack assemblies. As bandwidth requirements increase and low latency paths become more important, higher bandwidth can be provided to endpoints (e.g., hosts and / or servers) by reducing the number of steps between network endpoints.

[0094] For example, as described above, the top-of-rack switch 550' (e.g., spoke optical transceivers) can be optically configured with coherent receivers. However, as described further below, the operations performed by the coherent receivers can be further pushed to each compute node in the rack assembly.

[0095] In particular, the top-of-rack switch 550' can be configured to receive a replicated wavelength of the spectrum from the power splitter. As described above, the hub optical transceiver is configured to receive a wavelength of the spectrum (e.g., a full spectrum of wavelengths) that delivers a total bandwidth for data distribution, where the full spectrum of wavelengths has a total power. The cascaded power splitting layer includes one or more power splitters each configured to output a corresponding replicated wavelength of the spectrum, where each replicated wavelength of the spectrum has a corresponding power that is a fraction of the total power associated with the full spectrum of wavelengths provided by the hub optical transceiver. For example, the power splitter can be configured to replicate a wavelength of the spectrum and provide multiple replicated wavelengths of the spectrum.

[0096] As shown, the top-of-rack switch 550′ may provide an output via port 0 to another power splitter 460X, configurable as a 1×8 (1*8) power splitter. The power splitter 460X provides as outputs, via corresponding channels 711A, 711B, . . . , 711H, wavelengths of the corresponding replicated spectrum to each of the eight servers and / or compute nodes of the set 701 (e.g., server 0 through server 7). The top-of-rack switch 550′ may also provide an output via port 1 to another power splitter 460Y, configurable as a 1×8 (1*8) power splitter. The power splitter 460Y provides as outputs, via corresponding channels 712A, 712B, . . . , 712H, wavelengths of the corresponding replicated spectrum to each of the eight servers of the set 702 (e.g., server 0 through server 7). The top-of-rack switch 550' may also provide output via port 2 to another power splitter 460Z, which may be configurable as a 1x8 (1*8) power splitter. The power splitter 460Z provides corresponding replicated wavelengths of the spectrum as outputs to each of the eight servers (e.g., Server 0 through Server 7) of the set 703 via corresponding channels 713A, 713B, ..., 713H. Ports 3 through 7 of the top-of-rack switch 550' may similarly be configured to provide corresponding replicated wavelengths of the spectrum as outputs to an additional set of eight servers (not shown).

[0097] Further, each server may be configured with a corresponding coherent transceiver. In this manner, the coherent transceiver of the server coupled to the power splitter may be configured to receive wavelengths of the replicated spectrum. Furthermore, the coherent receiver is tunable to select a corresponding band of wavelengths from the wavelengths of the replicated spectrum that sets the bandwidth for the coherent transceiver, and correspondingly, the corresponding server. That is, the coherent receiver of the server is configured to split the wavelengths of the replicated spectrum into a band of selectable wavelengths that define the bandwidth for the corresponding server.

[0098] Accordingly, this disclosure describes an optical data center fabric (ODCF) and / or topology that provides the benefits of a full Internet Protocol (IP) multi-stage switching network, and further provides high-speed paths between the edge and leaf nodes of the data center provided using flexible optics technology.

[0099] It should be understood that the various embodiments defined herein may be combined or assembled into specific implementations that use various features disclosed herein. Thus, the examples provided are only some of the possible examples and are not intended to limit the various implementations that may be defined by combining various elements. In some examples, an implementation may include fewer elements without departing from the spirit of the disclosed or equivalent implementations.

[0100] Embodiments of the present disclosure may be practiced with a variety of computer system configurations including handheld devices, microprocessor systems, microprocessor-based or programmable consumer electronics, minicomputers, mainframe computers, etc. Embodiments of the present disclosure may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a wire-based or wireless network.

[0101] With the foregoing embodiments in mind, it will be appreciated that embodiments of the present disclosure can employ various computer-implemented operations involving data stored in computer systems. These operations are operations requiring physical manipulation of physical quantities. Any of the operations described herein that form part of embodiments of the present disclosure are useful machine operations. Embodiments of the present disclosure also relate to devices or apparatus for performing these operations. Apparatus can be specially constructed for the required purposes, or the apparatus can be a general-purpose computer selectively activated or configured by a computer program stored in the computer. In particular, various general-purpose machines can be used with computer programs written in accordance with the teachings herein. Alternatively, it may be more convenient to construct a more specialized apparatus to perform the required operations.

[0102] The present disclosure can also be embodied as computer-readable code on a computer-readable medium. The computer-readable medium can be any data storage device that can store data, which can then be read by a computer system. Examples of computer-readable media include hard drives, network-attached storage (NAS), read-only memory, random-access memory, CD-ROMs, CD-Rs, CD-RWs, magnetic tape, and other optical and non-optical data storage devices. The computer-readable medium can also include computer-readable tangible media distributed over network-connected computer systems so that the computer-readable code is stored and executed in a distributed fashion.

[0103] Although the method operations have been described in a particular order, it should be understood that other housekeeping operations may be performed between operations, or operations may be adjusted to occur at slightly different times, or may be distributed within a system that allows processing operations to occur at various intervals relative to processing, so long as the processing of the overlay operation is performed in the desired manner.

[0104] Although the foregoing disclosure has been described in some detail for clarity of understanding, it will be apparent that certain changes and modifications may be practiced within the scope of the appended claims. Accordingly, the present embodiments are to be considered as illustrative and not restrictive, and embodiments of the present disclosure are not to be limited to the details provided herein, but may be modified within the scope of the appended claims and their equivalents.

Claims

1. 1. An optical communication system, comprising: a controller configured to adjust each of a plurality of spoke transceivers to select a corresponding band of wavelengths from a spectrum of wavelengths, the wavelengths of the spectrum including a plurality of subcarriers, the plurality of subcarriers including wavelengths of a corresponding and unique band of wavelengths from the wavelengths of the spectrum; the plurality of spoke transceivers are coupled to a first power splitter that receives the wavelengths of the spectrum, the first power splitter configured to generate a plurality of replicated wavelengths of the spectrum, each of the plurality of spoke transceivers configured to receive a corresponding one of the wavelengths of the replicated spectrum at a first power; the first power splitter disposed in a first power splitting stage receives wavelengths of the spectrum at a second power from a second power splitter disposed in a second power splitting stage, the second power being greater than the first power; each of the plurality of spoke transceivers having a coherent receiver coupled to the controller, the coherent receiver being adjustable by the controller to select a wavelength in the corresponding band from the corresponding one of the wavelengths in the plurality of replicated spectra; Optical communication system.

2. the wavelengths of the replicated spectrum received by the spoke transceivers have a power that is a fraction of the total power for the wavelengths of the spectrum received from a hub optical transceiver configured to provide the wavelengths of the spectrum to power splitters in the first power splitting stage and the second power splitting stage; 2. The optical communication system according to claim 1.

3. the first power splitter divides the second power for the wavelengths of the spectrum received from the second power splitter equally or unevenly among the wavelengths of the plurality of replicated spectra; 2. The optical communication system according to claim 1.

4. 1. An optical communication system, comprising: a hub optical transceiver configured to receive a spectrum of wavelengths at a total power, the spectrum of wavelengths including a plurality of subcarriers, each of the plurality of subcarriers including a corresponding and unique band of wavelengths in the spectrum; a power splitter optically coupled to the hub optical transceiver and configured to receive the wavelengths of the spectrum at the total power, the power splitter configured to generate one or more replicated wavelengths of the spectrum at a reduced power; a spoke transceiver coupled to the power splitter and disposed within a rack assembly serving a plurality of servers, the spoke transceiver configured to receive wavelengths of the replicated spectrum at a power that is a fraction of the total power; a control system coupled to the spoke transceivers, the control system configured to adjust the spoke transceivers to select a band of wavelengths from the wavelengths of the replicated spectrum; Optical communication system.

5. a coherent receiver within a network interface of the spoke transceiver; the coherent receiver is coupled to the control system; the coherent receiver is adjustable by the control system to select the wavelength from the wavelengths of the replicated spectrum.

5. The optical communication system according to claim 4.

6. the coherent receiver is a hot-pluggable device; 6. The optical communication system according to claim 5.

7. the wavelengths of the replicated spectrum received by the spoke transceivers have the power that is the fraction of the total power for the wavelengths of the spectrum received from the hub optical transceiver.

5. The optical communication system according to claim 4.

8. an amplifier coupled between the spoke transceiver and the power splitter and configured to generate the power for the wavelengths of the replicated spectrum.

8. The optical communication system according to claim 7.

9. one or more of the hub optical transceiver, the power splitter, and the spoke transceivers are quad small form factor pluggable (QSFP) devices; 5. The optical communication system according to claim 4.

10. a plurality of spoke transceivers coupled to the power splitter and the control system, the plurality of spoke transceivers being disposed within a plurality of rack assemblies, the plurality of rack assemblies serving a corresponding plurality of servers; each of the plurality of spoke transceivers configured to receive a corresponding replicated spectrum wavelength from the one or more replicated spectrum wavelengths generated by the power splitter; the control system is configured to adjust each of the plurality of spoke transceivers to select a corresponding band of wavelengths from the corresponding replicated spectrum of wavelengths.

5. The optical communication system according to claim 4.

11. 1. An optical communication system, comprising: a hub optical transceiver configured to receive a spectrum of wavelengths at a total power, the spectrum of wavelengths including a plurality of subcarriers, each of the plurality of subcarriers including a corresponding and unique band of wavelengths in the spectrum; a power splitting layer comprising a plurality of power splitters optically coupled to the hub optical transceiver and configured to output a plurality of replicated wavelengths of the spectrum; a plurality of spoke transceivers coupled to the power splitting layer, each of the plurality of spoke transceivers configured to receive a corresponding one of the wavelengths of the plurality of replicated spectra at a corresponding power that is a fraction of the total power; a control system coupled to the plurality of spoke transceivers, the control system configured to adjust each of the plurality of spoke transceivers to select a corresponding band of wavelengths from the corresponding one of the plurality of replicated spectrum wavelengths; Optical communication system.

12. each of the plurality of power splitters configured to replicate a wavelength of the received spectrum at a reduced power; 12. The optical communication system according to claim 11.

13. The power splitting layer is configured into one or more cascading layers, the cascading layers comprising: a first power splitter coupled to the hub optical transceiver and configured to replicate wavelengths of the spectrum and output a plurality of first replicated wavelengths of the spectrum; a first power splitter coupled to the first power splitter and configured to replicate the wavelengths of the plurality of replicated first spectra and output a plurality of wavelengths of a second replicated spectrum; 12. The optical communication system according to claim 11.

14. Each of the plurality of spoke transceivers a coherent receiver in a network interface coupled to the control system; the coherent receiver is adjustable by the control system to select wavelengths in the corresponding band from the corresponding one of the wavelengths in the plurality of replicated spectra.

12. The optical communication system according to claim 11.

15. the coherent receiver is a hot-pluggable device; 15. The optical communication system according to claim 14.

16. each of the plurality of replicated wavelengths of the spectrum having a corresponding power that is the fraction of the total power for the wavelength of the spectrum received from the hub optical transceiver; 12. The optical communication system according to claim 11.

17. 12. The optical communication system of claim 11, further comprising an amplifier coupled between a spoke transceiver and a power splitter of the power splitting layer and configured to generate power for wavelengths of the replicated spectrum received by the spoke transceiver.

18. one or more of the hub optical transceiver, the plurality of power splitters, and the plurality of spoke transceivers are quad small form factor pluggable (QSFP) devices; 12. The optical communication system according to claim 11.

Images