1. Field
The present disclosure relates to spine-and-leaf networks, and in particular relates to optical interconnection assemblies for use in spine-and-leaf network networks that facilitate scale-out.
2. Technical Background
A data center is a location that houses computers and related telecommunications equipment and components for the purpose of processing (e.g., receiving, storing, managing and transmitting) data. Data centers often need to be expanded or “scaled out,” wherein hardware is added to accommodate the increasing data-processing demands. It is thus desirable that the data-center hardware be configured in a manner that is scalable, i.e., that can support scale-out of the hardware such that the data-processing performance of the data center improves in direct proportion to the added capacity.
Traditional data-center architectures have relied on a three-tier switching architecture whereby network reliability and scale-out capability is accomplished through switch redundancy. However, the three-tier switching architecture is not optimal for certain types of data centers, such as Internet data centers, that process relatively large amounts of data.
One type of network architecture that is well-suited for use in high-capacity data centers is called a “spine-and-leaf” (S/L) architecture, which flattens the network to reduce latency and simplifies redundancy. In this regard,
The ability to scale-out the S/L network 10 also depends on the data rates employed, e.g., ten (10) Gigabit Ethernet (10-GbE) or forty (40) GbE. Presently, many spine-switch components and leaf-switch connection components are rated for a 40 GbE data rate. However, a 40-GbE mesh would limit the network's ability to be scaled out because many leaf switches support only four 40-GbE uplink connection components to interface with the spine switch, which effectively limits the network to four spine switches. As a result, because the number and bandwidth of client connections is limited by the bandwidth capacity of the spine switches, this effectively limits the overall connection capacity of the S/L network 10.
One approach to overcoming this type of scale-out limitation involves creating a 10-GbE mesh to allow for four (4) times the amount of scale-out capability, i.e., sixteen 10-GbE connection components that allow for sixteen (16) spine switches, as opposed to provide four (4) 40-GbE connection components that limits the mesh to a maximum of four (4) spine switches. This 10-GbE mesh can be created by using cabling in the form of fiber optic cable jumpers terminated with LC duplex connectors to break out each 40-GbE connection component into 4×10-GbE connection components to obtain the sixteen (16) 10-GbE connection components. However, this creates cabling complexity because the number of individually routed 10-GbE optical fibers continues to increase exponentially as additional spine switches are added. At the same time, each optical fiber is potentially routed to a unique switch, thus defeating one of the advantages of the high-bandwidth connection components of the spine and leaf switches.
No admission is made that any reference cited herein constitutes prior art. Applicant expressly reserves the right to challenge the accuracy and pertinence of any cited documents.
Embodiments include optical interconnection assemblies supporting multiplexed data signals for providing network connectivity to a plurality of servers, clients and/or other computing devices. Related components, methods, and systems are also disclosed. For example, a spine-and-leaf (S/L) network requires that each spine switch be communicatively connected to every leaf switch in the network, and vice versa. However, as additional switches are added to a S/L network, cabling complexity increases exponentially, because individual connections must be made between each new spine/leaf switch combination. In accordance with exemplary embodiments disclosed herein, multiple channels can be provided from a spine or leaf switch to a central optical interconnection assembly using a single multiplexed signal from each spine and leaf switch. In this manner, the optical interconnection assemblies disclosed herein can be employed in a S/L network to provide a desired spine-leaf fabric that can be more easily scaled while minimizing cabling complexity to and from each individual spine and leaf switch.
To minimize cabling, exemplary optical interconnection assemblies disclosed herein have a plurality of spine-side multiplexer/demultiplexer pairs for communicating multiplexed communications signals between the optical interconnection assembly and one or more spine switches. The optical interconnection assemblies also have a plurality of leaf-side multiplexer/demultiplexer pairs for communicating multiplexed communications signals between the optical interconnection assembly and one or more leaf switches. Within the optical interconnection assembly, each spine-side demultiplexer is connected to every leaf-side multiplexer via at least one path, and each leaf-side demultiplexer is connected to every spine-side multiplexer via at least one path. In this manner, the optical interconnection assembly provides at least one discrete channel from each leaf switch to every spine switch, and vice versa. Also in this manner, each spine switch is connected directly to the optical interconnection assembly, and each leaf switch is also connected directly to the optical interconnection assembly.
Each leaf-side demultiplexer receives a leaf-side downlink signal from the leaf switch and demultiplexes the leaf-side downlink signal into a plurality of component downlink signals. For downlink signals, which travel in a leaf-to-spine direction (e.g., from leaf-connected servers to spine switches), the leaf-side demultiplexer provides at least one component downlink signal to every spine-side multiplexer, such that each spine-side multiplexer receives at least one component downlink signal from every leaf-side demultiplexer. Each spine-side multiplexer then multiplexes the downlink component signals received from the plurality of leaf-side demultiplexers into a spine-side downlink signal and provides the spine-side downlink signal to one of the spine switches. For uplink signals, which travel in a spine-to-leaf direction (e.g., from spine switches to leaf-connected servers), the spine-side demultiplexers distribute spine-side uplink signals received from the spine switches to the leaf-side multiplexers in a similar manner, such that each leaf-side multiplexer receives at least one component uplink signal from every spine-side demultiplexer. Each leaf-side multiplexer multiplexes the received uplink component signals into a respective leaf-side uplink signal and provides the leaf-side downlink signal to one of the leaf switches.
In this manner, the self-contained optical interconnection assembly permits additional spine and leaf switches to be more easily integrated into a network by connecting the spine and leaf switches directly to the optical interconnection assembly, as opposed to directly connecting each new spine to each additional leaf individually. Thus, the self-contained optical interconnection assembly reduces cabling complexity during network build-out, while maintaining increased scalability of the network.
One embodiment of the disclosure relates to an optical interconnection assembly for directing communication signals between spine and leaf connection components of a spine-and-leaf network. The optical interconnection assembly comprises a plurality of leaf-side demultiplexers, each having a leaf-side demultiplexer input and a plurality of leaf-side demultiplexer outputs. The optical interconnection assembly further comprises a plurality of spine-side multiplexers, each having a plurality of spine-side multiplexer inputs and a spine-side multiplexer output. The optical interconnection assembly further comprises a plurality of downlink optical paths. Each of the plurality of downlink optical paths is optically connected between a leaf-side demultiplexer output to a spine-side multiplexer input. Each spine-side multiplexer is configured to receive a component downlink signal on a downlink optical path from every leaf-side demultiplexer and multiplex the received component downlink signals into a multiplexed spine-side downlink signal.
Another embodiment of the disclosure relates to a spine and leaf (S/L) network comprising at least one leaf switch each having a plurality of leaf connection components, a plurality of spine switches each having a plurality of spine connection components, and at least one optical interconnection assembly for directing communication signals between spine and leaf connection components of the spine-and-leaf network. Each optical interconnection assembly comprises a plurality of leaf-side demultiplexers each having a leaf-side demultiplexer input and a plurality of leaf-side demultiplexer outputs. Each optical interconnection assembly further comprises a plurality of spine-side multiplexers, each having a plurality of spine-side multiplexer inputs and a spine-side multiplexer output. Each optical interconnection assembly further comprises a plurality of downlink optical paths. Each of the plurality of downlink optical paths is optically connected between a leaf-side demultiplexer output to a spine-side multiplexer input such that each leaf-side demultiplexer is optically connected to every spine-side multiplexer by at least one downlink optical path. Each spine-side multiplexer is configured to receive a component downlink signal on a downlink optical path from every leaf-side demultiplexer and multiplex the received component downlink signals into a multiplexed spine-side downlink signal.
Another embodiment of the disclosure relates to a method of directing communication signals between spine and leaf connection components of a spine-and-leaf network. The method comprises receiving a multiplexed leaf-side downlink signal from a leaf connection component at one of a plurality of leaf-side demultiplexer inputs of a plurality of leaf-side demultiplexers of an optical interconnection assembly. The method further comprises demultiplexing each multiplexed leaf-side downlink signal into a plurality of component downlink signals. The method further comprises providing each of the plurality of component downlink signals to a different one of a plurality of downlink optical paths of the optical interconnection assembly via one of a plurality of leaf-side demultiplexer outputs of the optical interconnection assembly. The method further comprises receiving a component downlink signal at each one of a plurality of spine-side multiplexer inputs of a plurality of spine-side multiplexers of the optical interconnection assembly via a downlink optical path from one of the leaf-side demultiplexers. Each spine-side multiplexer receives a component downlink signal from every one of the plurality of leaf-side demultiplexers. The method further comprises multiplexing, at each spine-side multiplexer, the received component downlink signals into a multiplexed spine-side downlink signal. The method further comprises providing each multiplexed spine-side downlink signal to a spine connection component at the spine-side multiplexer output.
Another embodiment of the disclosure relates to an optical interconnection assembly for directing communication signals between spine and leaf connection components of a spine-and-leaf network. The optical interconnection assembly comprises a plurality of spine-side demultiplexers, each having a spine-side demultiplexer input and a plurality of spine-side demultiplexer outputs. The optical interconnection assembly further comprises a plurality of leaf-side multiplexers, each having a plurality of leaf-side multiplexer inputs and a leaf-side multiplexer output. The optical interconnection assembly further comprises a plurality of uplink optical paths. Each of the plurality of uplink optical paths is optically connected between a spine-side demultiplexer output to a leaf-side multiplexer input. Each leaf-side multiplexer is configured to receive a component uplink signal on an uplink optical path from every spine-side demultiplexer and multiplex the received component uplink signals into a multiplexed leaf-side uplink signal.
Additional features and advantages will be set forth in the detailed description which follows, and in part will be readily apparent to those skilled in the art from the description or recognized by practicing the embodiments as described in the written description and claims hereof, as well as the appended drawings.
It is to be understood that both the foregoing general description and the following detailed description are merely exemplary, and are intended to provide an overview or framework to understand the nature and character of the claims.
The accompanying drawings are included to provide a further understanding, and are incorporated in and constitute a part of this specification. The drawings illustrate one or more embodiment(s), and together with the description serve to explain principles and operation of the various embodiments.
Various embodiments will be further clarified by the following examples.
Embodiments include optical interconnection assemblies supporting multiplexed data signals for providing network connectivity to a plurality of servers, clients and/or other computing devices. Related components, methods, and systems are also disclosed. To minimize cabling, exemplary optical interconnection assemblies disclosed herein have a plurality of spine-side multiplexer/demultiplexers for communicating multiplexed communications signals between the optical interconnection assembly and one or more spine switches. The optical interconnection assemblies also have a plurality of leaf-side multiplexer/demultiplexers for communicating multiplexed communications signals between the optical interconnection assembly and one or more leaf switches. Within the optical interconnection assembly, each spine-side demultiplexer is connected to every leaf-side multiplexer via at least one path, and each leaf-side demultiplexer is connected to every spine-side multiplexer via at least one path. In this manner, the optical interconnection assembly provides at least one discrete channel from each leaf switch to every spine switch, and vice versa. Also in this manner, each spine switch is connected directly to the optical interconnection assembly, and each leaf switch is also connected directly to the optical interconnection assembly.
Each leaf-side demultiplexer receives a leaf-side downlink signal from the leaf switch and demultiplexes the leaf-side downlink signal into a plurality of component downlink signals. For downlink (i.e., leaf-to-spine) signals, the leaf-side demultiplexer provides at least one component downlink signal to every spine-side multiplexer, such that each spine-side multiplexer receives at least one component downlink signal from every leaf-side demultiplexer. Each spine-side multiplexer then multiplexes the downlink component signals received from the plurality of leaf-side demultiplexers into a spine-side downlink signal and provides the spine-side downlink signal to one of the spine switches. For uplink (i.e., spine-to-leaf) signals, the spine-side demultiplexers distribute spine-side uplink signals received from the spine switches to the leaf-side multiplexers in a similar manner, such that each leaf-side multiplexer receives at least one component uplink signal from every spine-side demultiplexer. Each leaf-side multiplexer multiplexes the received uplink component signals into a respective leaf-side uplink signal and provides the leaf-side downlink signal to one of the leaf switches.
In this manner, the self-contained optical interconnection assembly permits additional spine and leaf switches to be more easily integrated into a network by connecting the spine and leaf switches directly to the optical interconnection assembly, as opposed to directly connecting each new spine to each additional leaf individually. Thus, the self-contained optical interconnection assembly reduces cabling complexity during network build-out, while maintaining increased scalability of the network.
One embodiment of the disclosure relates to an optical interconnection assembly for directing communication signals between spine and leaf connection components of a spine-and-leaf network. In this regard,
In the example embodiment of
In the exemplary S/L network 10 of
Referring now to
The optical interconnection assembly 18 further comprises a plurality of downlink optical paths 32D. In the example of
In this embodiment, each multiplexed leaf-side downlink signal is comprised of four different wavelengths, which output to the four (4) respective leaf-side demultiplexer outputs 22L of each leaf-side demultiplexer 24L. In this embodiment, each spine-side multiplexer input 28S also receives one component downlink signal 40D of each of the four wavelengths, so that the different component downlink signals 40D can be re-multiplexed into the spine-side downlink multiplexed signal without interfering with each other.
In this regard, each numbered multiplexer input 28S/L and demultiplexer output 22S/L refers to a specific wavelength (λ1-4). In other words, each demultiplexer 20S/L is configured to receive and demultiplex a multiplexed input signal 38D/46U having four component signals 40D/U on wavelengths 1-4. Each demultiplexer 20S/L outputs component signals 40 D/U having λ 1 to output 1, component signals 40 D/U having λ 2 to output 2, component signals 40 D/U having λ 3 to output 3, and component signals 40 D/U having λ 4 to output 4. Likewise, each multiplexer 24S/L is configured to receive component signals 40 D/U having λ 1 at input 1, component signals 40 D/U having λ 2 at input 2, component signals 40 D/U having λ 3 at input 3, and component signals 40 D/U having λ 4 at input 4. The received component signals 40D/U are then multiplexed into a single multiplexed output signal 38U/46D.
Thus, by connecting each demultiplexer output 22L/S to a multiplexer input 28L/S of the same wavelength (λ 1-4), and vice versa, it is ensured that the individual component signals 40D/U will not interfere with each other when they are multiplexed back into their respective output signals 38U/46D.
Referring back to
Each spine-side multiplexer 24S is optically connected to leaf-side multiplexer 20L(1) by at least one downlink optical path 32D, and is also connected to a spine connection component 17 (see
Referring now to
With continuing reference to
In this manner, parallel uplink and downlink signals can be communicated between any leaf switch 14 and any spine switch 16, while minimizing cabling complexity. Because only one fiber optic connection is required between each leaf switch 14 or spine switch 16 and the optical interconnection assembly 18, additional spine switches 16 can be added to the S/L network 10, without manually connecting each new spine switch 16 to every leaf switch 14, or vice versa.
The optical interconnection assembly 18 of FIGS. 2 and 3A-3D can employ a variety of hardware configurations and/or form-factors. In this example, each optical interconnection assembly 18 may be housed in a fiber optic module 52, but other form-factors are also possible, including, without limitation, a furcated cable or other assembly.
In this example, spine connection components 17 of each spine switch 16 are optically connected to spine-side multiplexer/demultiplexer pair 26S of optical interconnection assemblies 18 via one or more optical-fiber cables (i.e., spine-side optical paths 44D/U) while leaf connection components 15 of each leaf switch 14 are optically connected to leaf-side multiplexer/demultiplexer pairs 26L of the optical interconnection assemblies 18 via one or more optical-fiber cables (i.e., leaf-side optical paths 36D/U). In this embodiment, each of the spine-side and leaf-side multiplexer/demultiplexer pair 26S/L is connected to one or more fiber optic adapters 54, which are each configured to receive and optically connect optical-fiber cables 44D/U and 36D/U to the interconnection assembly 18. The optical-fiber cables 44D/U and 36D/U may be relatively short in some embodiments, and may also be referred to hereinafter as “patch cords” or “jumpers” as the term is used in the industry. In the present example, patch cords 44D/U and 36D/U are each 40 GbE.
In order for S/L network 10 to be fully meshed at 40 GbE, at least one spine connection component 17 of each spine switch 16 needs to be connected to at least one leaf connection component 15 of each leaf switch 14. Put another way, each spine switch 16 is connected to every leaf switch 14. The configuration of optical-fiber array 32D/U (also called a “harness”) in each optical interconnection assembly 18 defines a mesh that serves to connect at least one spine connection component 17 to at least one leaf connection component 15 in a manner that makes S/L network 10 more easily scalable without reducing the patch-cord cabling to 10 GbE and without adding additional cabling complexity.
In this regard,
To accomplish the scale-out of S/L network 10 of
It should be understood from the above examples that the optical interconnection assembly 18 is not limited to a “4×4” configuration. In this regard,
The spine-side multiplexer/demultiplexer pair bandwidth BWS is related to the number NS of WDM channels (cable pairs 32D/U) at each spine-side multiplexer/demultiplexer pair 26S and to the data rate D carried by each of the cables by the relationship BWS=NS·D. Likewise, the leaf-side multiplexer/demultiplexer pair bandwidth BWL is related to the number NL of WDM channels (cable pairs 32D/U) at each leaf-side multiplexer/demultiplexer pair 26L and to the data rate D carried by each of the cables by the relationship BWL=NL·D.
The spine-side multiplexer/demultiplexer pairs 26S and the leaf-side multiplexer/demultiplexer pairs 26L of optical interconnection assembly 60 are related by the equation
M
S
·BW
S
=M
L
·BW
L. (1)
Substituting for BWS and BWL in equation (1) using the above relationship for these terms yields the following relationship:
M
S
·N
S
·D=M
L
·N
L
·D. (2)
Equation (2) can be simplified into the following relationship:
N
S
/N
L
=M
L
/M
S. (3)
Equation (3) represents the basic relationship between the number MS of spine-side multiplexer/demultiplexer pairs 26S, the number ML of leaf-side multiplexer/demultiplexer pairs 26L, and the respective number NS and NL of cable pairs 32D/U at each of the spine-side and leaf-side multiplexer/demultiplexer pairs 26S/L. One or more optical interconnection assemblies 60 that are configured according to equation (3) can be used to scale-out the corresponding S/L network 10.
Table 1 below sets forth three example configurations for optical interconnection assembly 60 based on equation (3).
Thus, three different example configurations for optical interconnection assembly 60 have the following relationships, respectively: 1) BWS=BWL and NS=NL; 2) BWS=3·BWL and NS=3·NL; and 3) BWS=3·BWL and NS=4·NL. It should be understood that Example 1 is a “balanced” configuration, in which NS is equal to NL, while Examples 2 and 3 are “unbalanced” configurations, in which NS may be larger or smaller than NL. In many embodiments, the multiplexing and demultiplexing configuration and cabling complexity for a balanced configuration may be less complex than for an unbalanced configuration, because the number of frequencies required to permit each spine-side multiplexer/demultiplexer pair to communicate with every leaf-side multiplexer/demultiplexer pair is the same in both the uplink and downlink directions.
Unless otherwise expressly stated, it is in no way intended that any method set forth herein be construed as requiring that its steps be performed in a specific order. Accordingly, where a method claim does not actually recite an order to be followed by its steps or it is not otherwise specifically stated in the claims or descriptions that the steps are to be limited to a specific order, it is no way intended that any particular order be inferred.
It will be apparent to those skilled in the art that various modifications and variations can be made without departing from the spirit or scope of the disclosure. Since modifications combinations, sub-combinations and variations of the disclosed embodiments incorporating the spirit and substance of the disclosure may occur to persons skilled in the art, the disclosure should be construed to include everything within the scope of the appended claims and their equivalents.
This application claims the benefit of priority under 35 U.S.C. §119 of U.S. Provisional Application Serial No. 61/977,686 filed on Apr. 10, 2014, the content of which is relied upon and incorporated herein by reference in its entirety. This application is related to U.S. patent application Ser. No. 14/177,443 filed Feb. 11, 2014, the content of which is incorporated herein by reference in its entirety.
Number | Date | Country | |
---|---|---|---|
61977686 | Apr 2014 | US |