A method, system, apparatus, and storage medium for setting up a universal substrate scale-out

Abstract

This invention provides a method, system, device, and storage medium for setting up scale-out on a universal substrate. The method includes: mounting an OAM module on the backplane of the universal substrate and setting the scale-out to a retimer card; mounting multiple GENZ4C+ connectors on the universal substrate to connect the scale-out signals and the management signals of the retimer card from the universal substrate to the retimer card; configuring the retimer card according to the OAM module; and selecting the path of a multiplexer by identifying different OAM modules to enable BMC or OAM module management of QSFP-DD modules. This invention integrates the scale-out portion of the OAM module onto a separate board, greatly improving the board's reusability. Through different compatible designs, it ensures the compatibility of the UBB backplane, allowing a single UBB board to be used with different OAM modules.

Description

Technical Field

[0001] This invention relates to the field of servers, and more specifically, to a method, system, apparatus, and storage medium for setting up a universal baseboard scale-out. Background Technology

[0002] With the rapid development of information technology in today's society, the demand for AI servers in data centers is increasing. Artificial intelligence (AI) applications are developing rapidly, driving an explosive growth in new hardware accelerators for machine learning (ML), deep learning (DL), and high-performance computing (HPC). Traditional PCIe CEM-sized cards are not optimized for AI workloads, which require ever-increasing bandwidth and interconnect flexibility for data / model parallelism. Therefore, OAI / OAM was born. OAM (OCP accelerator module) can interconnect not only on a single server but also multiple UBB (Universal Baseboard) devices via Ethernet.

[0003] like Figure 1 As shown, the UBB specification defines the scale-out method for the OAM module, where one port connects to the RETIMER, which in turn connects to the QSFP-DD, and then to a switch via fiber optic cable to interconnect with other UBBs and build large-scale AI training models. The QSFP-DD, RETIMER, and OAM are all on a single UBB board. Currently, the scale-out and OAM components are on the same board. This design cannot meet the design requirements for using multiple OAM modules, and it lacks flexibility in combining various OAMs and retimers. Changing the OAM module may require redesigning the entire UBB backplane, significantly increasing costs. Furthermore, the current design places the QSFP-DD and OAM modules on the same plane, compromising overall heat dissipation. The current solution also increases the complexity of UBB backplane cabling and design costs. Summary of the Invention

[0004] In view of this, the purpose of this invention is to provide a method, system, computer device, and computer-readable storage medium for setting up a universal substrate scale-out. This invention integrates the scale-out part of the OAM module onto a separate board, greatly improving the reusability of the board. At the same time, through different compatible designs, it ensures the compatibility of the UBB backplane, enabling a single UBB board to be paired with different OAM modules.

[0005] To achieve the above objectives, one aspect of the present invention provides a method for setting up a scale-out on a universal substrate, comprising the following steps: setting an OAM module on the backplane of the universal substrate and setting the scale-out to a retimer card; setting multiple GENZ4C+ connectors on the universal substrate to connect the scale-out signal and the management signal of the retimer card from the universal substrate to the retimer card; configuring the retimer card according to the OAM module; and selecting the path of a multiplexer by identifying different OAM modules to enable the BMC or OAM module to manage the QSFP-DD module.

[0006] In some implementations, the method further includes: setting two retimer chips on the retimer card to connect two QSFP-DDs to achieve interconnection with other systems.

[0007] In some implementations, configuring the retimer card according to the OAM module includes: configuring the retimer card using a BMC in response to the OAM module not supporting configuration of the retimer card; and configuring the retimer card by converting the I2C connection of the OAM module to the CPLD to MDC / MDIO in response to the OAM module supporting configuration of the retimer card.

[0008] In some embodiments, the method further includes performing unified signal processing on a plurality of GENZ4C+ connectors on the general-purpose substrate.

[0009] In another aspect of this invention, a system for setting up scale-out on a universal substrate is provided, comprising: a first setting module configured to set an OAM module on the backplane of the universal substrate and set the scale-out to a retimer card; a second setting module configured to set up a plurality of GENZ4C+ connectors on the universal substrate to connect the scale-out signal and the management signal of the retimer card from the universal substrate to the retimer card; a configuration module configured to configure the retimer card according to the OAM module; and a management module configured to select the path of a multiplexer by identifying different OAM modules to enable BMC or OAM module management of QSFP-DD modules.

[0010] In some implementations, the system also includes an interconnect module configured to: set two retimer chips on the retimer card to connect two QSFP-DDs to enable interconnection with other systems.

[0011] In some implementations, the configuration module is configured to: configure the retimer card using a BMC in response to the OAM module not supporting configuration of the retimer card; and convert the I2C connection of the OAM module to the CPLD to MDC / MDIO to configure the retimer card in response to the OAM module supporting configuration of the retimer card.

[0012] In some embodiments, the system also includes a processing module configured to perform unified signal processing on a plurality of GENZ4C+ connectors on the universal substrate.

[0013] In another aspect of the present invention, a computer device is provided, comprising: at least one processor; and a memory storing computer instructions executable on the processor, the instructions, when executed by the processor, implementing the steps of the method described above.

[0014] In another aspect, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method steps.

[0015] The present invention has the following beneficial technical effects:

[0016] 1. By designing the Scale Out section and OAM module on different boards, the Retimer card can be designed with less than half of the stack-up, reducing the overall cost;

[0017] 2. It can effectively support OAM from different manufacturers, and can adapt to different OAM design requirements by simply replacing the Retimer card, thus improving design flexibility;

[0018] 3. Through extensive compatibility design, it can meet the management needs of different Retimer cards. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of a common substrate in the prior art;

[0021] Figure 2 A schematic diagram of an embodiment of the method for setting up a universal substrate scale-out provided by the present invention;

[0022] Figure 3 This is an overall topology diagram of an embodiment of the present invention;

[0023] Figure 4 This is a topology diagram of the retimer card according to an embodiment of the present invention;

[0024] Figure 5 This is a topology diagram of MDC / MDIO according to an embodiment of the present invention;

[0025] Figure 6 This is a QSFP-DD management topology diagram according to an embodiment of the present invention;

[0026] Figure 7 A schematic diagram of an embodiment of the system for setting up a universal substrate scale-out provided by the present invention;

[0027] Figure 8 A schematic diagram of the hardware structure of an embodiment of a computer device with a universal substrate scale-out provided by the present invention;

[0028] Figure 9 This is a schematic diagram of an embodiment of a computer storage medium with a universal substrate scale-out provided by the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to specific examples and the accompanying drawings.

[0030] It should be noted that all uses of "first" and "second" in the embodiments of the present invention are for the purpose of distinguishing two entities or parameters with the same name but different names. It is clear that "first" and "second" are only for the convenience of expression and should not be construed as limiting the embodiments of the present invention. Subsequent embodiments will not explain this in detail.

[0031] In a first aspect, an embodiment of a method for setting up a universal substrate scale-out is proposed. Figure 2 The diagram shown is an embodiment of the method for setting up a universal substrate scale-out provided by the present invention. Figure 2 As shown, the embodiments of the present invention include the following steps:

[0032] S1. Place the OAM module on the backplane of the universal substrate and set the scale out to the retimer card;

[0033] S2. Multiple GENZ4C+ connectors are provided on the general-purpose substrate to connect the scale-out signal and the management signal of the retimer card from the general-purpose substrate to the retimer card.

[0034] S3. Configure the retimer card according to the OAM module; and

[0035] S4. Select the multiplexer path by identifying different OAM modules to enable BMC or OAM module management of QSFP-DD modules.

[0036] The embodiments of this invention flexibly accommodate the scale-out design requirements of multiple OAMs, ensuring that a single UBB backplane can be used to support multiple OAM designs, while reducing the PCB stack-up of the UBB backplane, achieving an optimal combination of design and cost; in addition, by not placing the QSFP-DD connector and the OAM module on the same plane, the overall heat dissipation effect and power consumption requirements are improved. Figure 3 This is an overall topology diagram of an embodiment of the present invention. Four GENZ4C+ connectors are placed on the UBB board to connect the scale-out signals and the retimer management signals from the UBB board to the retimer board, thereby separating the scale-out from the UBB board and improving the overall compatibility and reusability of the system.

[0037] The OAM module is mounted on the backplane of a universal substrate, and the scale-out signal is configured on the retimer card. Multiple GENZ4C+ connectors are placed on the universal substrate to connect the scale-out signal and the retimer card's management signals from the universal substrate to the retimer card. The retimer card is configured according to the OAM module. The multiplexer path is selected by identifying different OAM modules to enable BMC or OAM module management of the QSFP-DD module.

[0038] The OAM module is defined according to the OCP specification, and the pin definitions for OAM are consistent across manufacturers. Therefore, this embodiment of the invention places the OAM module only on the UBB backplane, ensuring compatibility with current mainstream OAM vendors such as Intel and Cambricon. Since the scale-out portion has inconsistent definitions across different manufacturers, this embodiment of the invention places it on a dedicated retimer card to ensure the reusability of the UBB backplane.

[0039] In some implementations, the method further includes: setting two retimer chips on the retimer card to connect two QSFP-DDs to achieve interconnection with other systems. Figure 4 This is a topology diagram of the retimer card according to an embodiment of the present invention. Figure 4As shown, two retimer chips are placed on the retimer card and connected to two QSFP-DDs to enable interconnection with other systems.

[0040] In some embodiments, the method further includes performing unified signal processing on multiple GENZ4C+ connectors on the universal substrate. Given that different OAM and Retimer manufacturers have different requirements for polarity reversal and lane reversal, this embodiment of the invention performs unified signal processing on the four Genz4C+ connectors of the UBB, ensuring that the same UBB backplane can be used regardless of whether the OAM module or Retimer card is replaced in the future.

[0041] In some implementations, configuring the retimer card according to the OAM module includes: configuring the retimer card using a BMC in response to the OAM module not supporting configuration of the retimer card; and configuring the retimer card by converting the I2C connection of the OAM module to the CPLD to MDC / MDIO in response to the OAM module supporting configuration of the retimer card.

[0042] Given the different management methods for Retimers among various OAM vendors, this embodiment of the invention incorporates the following compatibility design based on current OAM standards: When a vendor's OAM does not support Retimer configuration, the BMC is used to directly configure the Retimer; when the vendor's OAM supports Retimer configuration, the OAM's I2C is connected to the CPLD and converted to MDC / MDIO for Retimer configuration. The CPLD will switch the MDIO MUX based on the different OAM standards. It is worth noting that the scale-out portion of most mainstream OAM vendors currently uses the Ethernet protocol; therefore, the retimer in this embodiment is primarily a Phyretimer, managed using MDC / MDIO. Figure 5 This is a topology diagram of MDC / MDIO according to an embodiment of the present invention.

[0043] Given that different OAM vendors manage the QSFP-DD section in different ways, this embodiment of the invention also incorporates a compatibility design. The CPLD selects the MUX path by identifying different OAMs, enabling BMC or OAM to manage the QSFP-DD module. Figure 6 This is a QSFP-DD management topology diagram according to an embodiment of the present invention.

[0044] In this embodiment of the invention, the Retimer card is designed as a 1U standard card module, which makes it easier to ensure heat dissipation and space efficiency.

[0045] This invention integrates the Scale Out portion of the OAM module onto a separate board, significantly improving board reusability. Simultaneously, through various compatible designs, it ensures the compatibility of the UBB backplane, allowing a single UBB board to be paired with different OAM modules. This effectively solves the following problems: 1. If the Scale Out portion and OAM module were designed onto a single board, at least one layer would be added to the current UBB stack-up, increasing design costs by at least 15%. With the independent design described above, the Retimer card requires less than half the stack-up, reducing overall costs. 2. It effectively solves compatibility issues. Current OAM designs from different manufacturers are not entirely consistent. The above design effectively ensures compatibility with OAM from different manufacturers, allowing for adaptation to different OAM design requirements by simply replacing the Retimer card, thus improving design flexibility. 3. Through numerous compatible designs, it can meet the management needs of different Retimer cards.

[0046] It should be particularly noted that the steps in each embodiment of the above-described method for setting up a universal substrate scale-out can be interchanged, substituted, added, or deleted from each other. Therefore, these reasonable permutations and combinations of the method for setting up a universal substrate scale-out should also fall within the protection scope of this invention, and the protection scope of this invention should not be limited to the embodiments.

[0047] Based on the above objectives, a second aspect of the present invention provides a system for setting up a universal substrate scale-out. For example... Figure 7 As shown, system 200 includes the following modules: a first setting module configured to set the OAM module on the backplane of a universal substrate and set the scale-out signal to the retimer card; a second setting module configured to set multiple GENZ4C+ connectors on the universal substrate to connect the scale-out signal and the management signal of the retimer card from the universal substrate to the retimer card; a configuration module configured to configure the retimer card according to the OAM module; and a management module configured to select the path of the multiplexer by identifying different OAM modules to realize BMC or OAM module management of QSFP-DD modules.

[0048] In some implementations, the system also includes an interconnect module configured to: set two retimer chips on the retimer card to connect two QSFP-DDs to enable interconnection with other systems.

[0049] In some implementations, the configuration module is configured to: configure the retimer card using a BMC in response to the OAM module not supporting configuration of the retimer card; and convert the I2C connection of the OAM module to the CPLD to MDC / MDIO to configure the retimer card in response to the OAM module supporting configuration of the retimer card.

[0050] In some embodiments, the system also includes a processing module configured to perform unified signal processing on a plurality of GENZ4C+ connectors on the universal substrate.

[0051] Based on the above objectives, a third aspect of the present invention provides a computer device comprising: at least one processor; and a memory storing computer instructions executable by the processor to perform the following steps: S1, mounting an OAM module on the backplane of a universal substrate and setting a scale-out signal to a retimer card; S2, mounting a plurality of GENZ4C+ connectors on the universal substrate to connect the scale-out signal and the management signal of the retimer card from the universal substrate to the retimer card; S3, configuring the retimer card according to the OAM module; and S4, selecting a multiplexer path by identifying different OAM modules to enable BMC or OAM module management of a QSFP-DD module.

[0052] In some implementations, the steps further include: setting two retimer chips on the retimer card to connect two QSFP-DDs to enable interconnection with other systems.

[0053] In some implementations, configuring the retimer card according to the OAM module includes: configuring the retimer card using a BMC in response to the OAM module not supporting configuration of the retimer card; and configuring the retimer card by converting the I2C connection of the OAM module to the CPLD to MDC / MDIO in response to the OAM module supporting configuration of the retimer card.

[0054] In some embodiments, the steps further include: performing unified signal processing on the multiple GENZ4C+ connectors on the general-purpose substrate.

[0055] like Figure 8 The diagram shown is a hardware structure schematic of an embodiment of the computer device with the above-described universal substrate scale-out provided by the present invention.

[0056] For example Figure 8The device shown is an example, which includes a processor 301 and a memory 302.

[0057] Processor 301 and memory 302 can be connected via a bus or other means. Figure 8 Taking the example of a connection between China and Israel via a bus.

[0058] The memory 302, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the method for setting the general-purpose substrate scale-out in the embodiments of this application. The processor 301 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions, and modules stored in the memory 302, thereby implementing the method for setting the general-purpose substrate scale-out.

[0059] Memory 302 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created by using the method of setting up a general-purpose substrate scale-out, etc. Furthermore, memory 302 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, memory 302 may optionally include memory remotely located relative to processor 301, and these remote memories can be connected to a local module via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0060] One or more computer instructions 303 corresponding to the method of setting the general substrate scale out are stored in memory 302. When executed by processor 301, the method of setting the general substrate scale out in any of the above method embodiments is executed.

[0061] Any embodiment of the computer device that performs the above-described method for setting up a general-purpose substrate scale-out can achieve the same or similar effects as any of the aforementioned method embodiments.

[0062] The present invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs a method for setting a general-purpose substrate to scale out.

[0063] like Figure 9 The diagram shown is a schematic representation of an embodiment of the computer storage medium with a universal substrate scale-out provided by the present invention. Figure 9Taking the computer storage medium shown as an example, the computer-readable storage medium 401 stores a computer program 402 that, when executed by a processor, performs the above method.

[0064] Finally, it should be noted that those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program for setting up a general-purpose substrate scale-out can be stored in a computer-readable storage medium. When executed, the program can include the processes of the embodiments of the above methods. The storage medium for the program can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc. The above computer program embodiments can achieve the same or similar effects as any of the corresponding foregoing method embodiments.

[0065] The above are exemplary embodiments disclosed in this invention. However, it should be noted that various changes and modifications can be made without departing from the scope of the embodiments of this invention as defined by the claims. The functions, steps, and / or actions of the methods according to the disclosed embodiments described herein do not need to be performed in any particular order. Furthermore, although the elements disclosed in the embodiments of this invention may be described or claimed individually, they may be understood as multiple unless explicitly limited to a singular number.

[0066] It should be understood that, as used herein, the singular form “a” is intended to include the plural form as well, unless the context clearly supports an exception. It should also be understood that, as used herein, “and / or” refers to any and all possible combinations of one or more of the associated listed items.

[0067] The embodiment numbers disclosed in the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0068] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0069] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of the invention (including the claims) is limited to these examples. Within the framework of the invention, technical features of the above embodiments or different embodiments can be combined, and many other variations of different aspects of the invention exist, which are not provided in the details for the sake of brevity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the invention should be included within the protection scope of the invention.

Claims

1. A method for setting up a universal substrate scale-out, characterized in that, Includes the following steps: The OAM module is placed on the backplane of the universal substrate, and the scale out setting is set on the retimer card; Multiple GENZ4C+ connectors are provided on the general-purpose substrate to connect the scale-out signal and the management signal of the retimer card from the general-purpose substrate to the retimer card; Configure the retimer card according to the OAM module; and By identifying different OAM modules, the multiplexer path can be selected to enable BMC or OAM module management of QSFP-DD modules; The step of configuring the retimer card according to the OAM module includes: in response to the OAM module not supporting the configuration of the retimer card, configuring the retimer card using BMC; and in response to the OAM module supporting the configuration of the retimer card, converting the I2C connection of the OAM module to the CPLD to MDC / MDIO to configure the retimer card.

2. The method according to claim 1, characterized in that, The method also includes: Two retimer chips are set on the retimer card to connect to two QSFP-DDs to enable interconnection with other systems.

3. The method according to claim 1, characterized in that, The method also includes: Unified signal processing is performed on multiple GENZ4C+ connectors on the general-purpose substrate.

4. A system for setting up a universal substrate scale-out, characterized in that, include: The first setting module is configured to set the OAM module on the backplane of the universal substrate and set the scale out to the retimer card; The second setting module is configured to set multiple GENZ4C+ connectors on the general-purpose substrate to connect the scale-out signal and the management signal of the retimer card from the general-purpose substrate to the retimer card. The configuration module is used to configure the retimer card according to the OAM module; and The management module is configured to select the multiplexer path by identifying different OAM modules to enable BMC or OAM module management of QSFP-DD modules; The configuration module is configured to: in response to the OAM module not supporting configuration of the retimer card, configure the retimer card using BMC; and in response to the OAM module supporting configuration of the retimer card, connect the I2C of the OAM module to the CPLD to convert it to MDC / MDIO to configure the retimer card.

5. The system according to claim 4, characterized in that, The system also includes an interconnect module, configured for: Two retimer chips are set on the retimer card to connect to two QSFP-DDs to enable interconnection with other systems.

6. The system according to claim 4, characterized in that, The system also includes a processing module configured for: Unified signal processing is performed on multiple GENZ4C+ connectors on the general-purpose substrate.

7. A computer device, characterized in that, include: At least one processor; as well as A memory storing computer instructions executable on the processor, which, when executed by the processor, implement the steps of the method according to any one of claims 1-3.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1-3.

Images