Cloud computing measurement and control system and method based on intelligent network card DPU

Abstract

The application discloses a cloud computing measurement and control system and method based on an intelligent network card DPU, belongs to the field of combination of spaceflight measurement and control and cloud computing, and comprises an intelligent network card DPU module, data transceiving matching is carried out between a transceiving protocol stack and a digital radio frequency front-end device in the intelligent network card DPU module, and the intelligent network card DPU module further comprises a digital front-end data self-adaptive following module, a measurement and control acceleration pooling module, an intelligent network card clouding module and a measurement and control algorithm acceleration module. The application realizes high-throughput measurement and control signal transmission, realizes accurate tracking of different rates by the network card, and satisfies reliable, low-delay data transmission of the digital front-end and the cloud platform.

Description

Technical Field

[0001] This invention relates to the field of aerospace telemetry and control combined with cloud computing, and more specifically, to a cloud computing telemetry and control system and method based on a smart network interface card (DPU). Background Technology

[0002] With the increasing number of space launch missions, the number of satellites in orbit is growing exponentially; in 2018 alone, countries launched a total of 476 satellites, spacecraft, and other spacecraft. Simultaneously, with the advancement and implementation of the military-civilian integration policy, the defense and military aerospace industry is gradually expanding into civilian and commercial applications, commercial aerospace is booming, 5G technology is progressing rapidly, and my country has also made significant progress in building a unified space-ground network. All of this places increasingly higher demands on existing aerospace tracking, telemetry, and command (TT&C) systems. The development of a new generation of ground-based TT&C systems not only requires multi-target, multi-mission, high bandwidth, low latency, high speed, and networking capabilities, but also features information resource development and sharing, unified management, and unified scheduling, with an increasing emphasis on resource integration and efficient utilization. Traditional technical solutions are struggling to meet these development needs. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of the prior art and provide a cloud computing measurement and control system and method based on a smart network interface card (NIC) DPU, which can realize high-throughput measurement and control signal transmission, enable the NIC to accurately track different rates, and meet the requirements of reliable and low-latency data transmission for digital front-end and cloud platform.

[0004] The objective of this invention is achieved through the following solution:

[0005] A cloud computing measurement and control system based on a smart network interface card (NIC) DPU includes:

[0006] The intelligent network interface card (NIC) DPU module performs data transmission and reception matching with the digital radio frequency front-end device based on the transceiver protocol stack; and the intelligent NIC DPU module also includes a digital front-end data adaptive following module, a measurement and control acceleration pooling module, an intelligent NIC cloudification module, and a measurement and control algorithm acceleration module.

[0007] The digital front-end data adaptive following module includes a digital phase-locked loop unit, which is used to track the UDP transmission frame frequency of the digital front-end device and send uplink data to the radio frequency front-end while receiving digital front-end data, using the restored data rate.

[0008] The measurement and control acceleration pooling module includes an interconnection module, which connects multiple smart network cards.

[0009] The smart network card cloudification module includes a CPU and cloud platform scheduler software. A corresponding agent program runs on the CPU of the smart network card cloudification module, and the agent program communicates with the cloud scheduling center.

[0010] The measurement and control algorithm acceleration module FPGA chip and image repository utilize the FPGA chip to provide user algorithm space, allowing users to select the measurement and control algorithm library provided by the smart network card, create a standard algorithm library, and upload it to the image repository.

[0011] Furthermore, it includes a digital radio frequency front-end device for sending data to the smart network interface card (DPU) module.

[0012] Furthermore, the digital phase-locked loop unit includes a fully digital phase-locked loop unit (DPLL), and while receiving downlink data, the frequency of the received data frame is locked using the fully digital phase-locked loop unit (DPLL). Uplink data is then retrieved from the data buffer and sent according to this frequency, ensuring that the uplink and downlink rates are consistent.

[0013] Furthermore, the interconnection module utilizes the accelerator's 100G optical port for interconnection, enabling external network processing for data exchange with external systems and internal network processing for collaboration between multiple cards, thus achieving interoperability between any smart network cards.

[0014] Furthermore, for low-speed data, the interconnect module communicates directly with the server via the PCIe bus.

[0015] Furthermore, the agent program reports smart network interface card (NIC) information and executes tasks according to the scheduler's instructions; the cloud scheduling center enables the agent program to download and run software from the image repository, and also enables the software to configure parameters. Based on the status reported by the agent program, the health status of the smart NIC is determined. When a smart NIC malfunctions and is detected by the cloud scheduling center, its services are migrated to other smart NICs for operation.

[0016] Furthermore, the cloud scheduling center communicates with the CPU's agent program to download algorithms to the smart network card; if the existing algorithm library cannot meet the user's needs, the user can design an algorithm according to the rules, compile it into an image, and publish it to the image repository.

[0017] A cloud computing measurement and control method based on a smart network interface card (NIC) DPU, based on any of the systems described above, further includes the following steps:

[0018] S1, upload the algorithm software of the measurement and control system to the image repository, and configure the cloud platform system to use the image repository;

[0019] S2 uses a cloud scheduling system to pull server software and corresponding accelerator card software from the image repository for deployment.

[0020] S3 sends data to the accelerator card through the digital front-end device and is configured to send data at different rates. The smart network card DPU module receives the data and, based on the tracking of the data rate, sends the uplink data from the server to the digital front-end at the same rate, and detects the reliability of the uplink data through the digital front-end.

[0021] S4, after the smart network card DPU module receives downlink data, it performs calculations using the measurement and control acceleration algorithm loaded on the acceleration card, and then sends the calculation results to the CPU for further processing.

[0022] Furthermore, following step S4, the following step is also included:

[0023] S5: When a smart network card is manually removed or malfunctions, the scheduling terminal detects the fault, selects an idle smart network card from the pool, loads the current task image into its agent program, and sets the IP address of the faulty network card through the agent program. The data from the digital front end is directly transmitted to the new smart network card, thereby realizing automatic and rapid migration of services and uninterrupted service.

[0024] Furthermore, in step S1, the algorithm software includes capture algorithm software, decoding algorithm software, and custom response algorithm software.

[0025] The beneficial effects of this invention include:

[0026] The technical solution of this invention achieves dual-channel throughput of up to 100G for measurement and control signal transmission. Utilizing an adaptive algorithm, it enables the network card to accurately track different speeds with latency errors at the microsecond level, meeting the requirements of reliable, low-latency data transmission for digital front-ends and cloud platforms. By designing an additional FPGA computing chip on the smart network card and loading an algorithm library, the CPU's workload is offloaded, satisfying the real-time requirements of traditional measurement and control technologies. Furthermore, the dual-channel network card enables internal and external network networking, achieving large-scale pooling capabilities. By designing a CPU on the smart network card and running a corresponding agent program, and improving the cloud scheduling software, the smart network card enables the cloudification of measurement and control technologies. Attached Figure Description

[0027] 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 drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1This is a schematic diagram of a 100G adaptive speed high-precision follower according to an embodiment of the present invention;

[0029] Figure 2 This is a schematic diagram of the intelligent network interface card pooling of the measurement and control system according to an embodiment of the present invention;

[0030] Figure 3 This is a schematic diagram of the cloudification of the intelligent network card in the measurement and control system according to an embodiment of the present invention. Detailed Implementation

[0031] All features disclosed in all embodiments of this specification, or steps in all methods or processes implied in the disclosure, may be combined and / or extended or replaced in any way, except for mutually exclusive features and / or steps.

[0032] In view of the problems mentioned in the background, the technical concept of this invention adopts a cloud computing architecture and conducts research on virtual baseband pool technology based on real-time cloud computing, forming a cloud-based aerospace ground telemetry, tracking, and command (TT&C) system. This can improve the overall utilization efficiency of the ground system, reduce operating costs, and keep pace with the times. It can be used to promote innovation and upgrading of the aerospace industry, create new business models, and conform to the inevitable trend of future military aerospace equipment and commercial aerospace military-civilian integration technology development. However, after further consideration, the inventors of this invention found that the large data flow, high resource consumption, and strong real-time requirements of the TT&C algorithm will place a heavy load on the CPU and introduce uncertainty in the time accuracy of the scheduling system. Even with the adoption of technologies such as core binding, there will still be a data delay deviation of tens of millimeters, which cannot meet the requirements of the TT&C system.

[0033] Faced with the contradiction between demand and software implementation, after creative thinking, the technical solution of this invention integrates advanced technical architecture with relevant hardware and software to design an intelligent network card suitable for cloud measurement and control systems. In the conception of this invention, the cloud measurement and control intelligent network card provided is different from ordinary network cards. It implements the transceiver protocol stack on the intelligent network card and accurately matches it with the data transmission and reception implementation of the radio frequency front end. It automatically adapts to different rates, up to 100G. At the same time, it also provides measurement and control algorithms such as acquisition algorithm and decoding algorithm for use, offloading the CPU pressure, and fully supports pooling and cloud management to achieve true cloud measurement and control.

[0034] Specifically, this smart network interface card (NIC) utilizes an additional FPGA computing chip on the NIC, loading an algorithm library to offload the CPU and meet the real-time requirements of traditional measurement and control technologies. Furthermore, it achieves large-scale pooling capabilities by enabling internal and external network connectivity through dual-channel NICs. By designing a CPU on the NIC to run a corresponding agent program and improving cloud scheduling software, the smart NIC enables cloud-based measurement and control technologies. This smart NIC achieves dual-channel measurement and control signal transmission with a maximum throughput of 100Gbps. Adaptive algorithms enable precise tracking of different data rates, with latency errors at the microsecond level, meeting the requirements for reliable, low-latency data transmission between digital front-ends and cloud platforms.

[0035] Further inventive concepts include a digital front-end data adaptive following module, a measurement and control acceleration pooling module, a smart network card cloudification module, and a measurement and control algorithm acceleration module.

[0036] The digital front-end data adaptive tracking module includes a fully digital phase-locked loop (DPLL) unit. In specific implementations, a 100G high-precision digital front-end data adaptive tracking technology solution is provided. This solution utilizes a digital phase-locked loop to solve the high-speed synchronization of the digital front-end at 100G speeds, achieving accuracy at the microsecond level. It can accurately track the UDP frame frequency, ensuring that the RF digital front-end's buffer does not overflow due to mismatched UDP frame transmission and reception rates. A phase-locked loop is a phase feedback control system. In a digital phase-locked loop, because the error control signal is a discrete digital signal rather than an analog voltage, the change in the controlled output voltage is discrete rather than continuous. Furthermore, all loop components are implemented using digital circuits; therefore, the phase-locked loop used in this embodiment is a fully digital phase-locked loop, or DPLL for short. The smart network card in this embodiment uses a fully digital DPLL to track the UDP transmission frame frequency of the digital front-end device, accurately receive digital front-end data, and simultaneously transmit uplink data to the RF front-end using the restored data rate. Figure 1 As shown, the digital front-end data adaptive following module includes the following improved process: while receiving downlink data, the frequency of the received data frame is precisely locked using a fully digital phase-locked loop (DPLL) unit. Uplink data is then retrieved from the data buffer and sent according to this frequency, ensuring consistent uplink and downlink rates, with an adaptability range from 0 to 100G.

[0037] The measurement and control acceleration pooling module includes an interconnect module. The smart network interface cards (NICs) in the measurement and control system not only handle network transmission but also process measurement and control algorithm data. For complex tasks, multiple smart NICs are needed to assist in processing. The interconnect module utilizes the accelerator's 100G optical port for interconnection. The external network handles the exchange of data with external systems, while the internal network handles collaboration between multiple NICs, enabling interoperability between any two NICs with latency in the microsecond range. For low-speed data, the interconnect module can also directly communicate with the server via the PCIe bus, such as... Figure 2As shown.

[0038] The smart NIC cloudification module includes a CPU and cloud platform scheduler software. After the smart NICs are pooled, a corresponding scheduler and policies are still needed to achieve flexible management and allocation, realizing true cloudification. Therefore, the technical solution of this invention improves the design of the CPU on the smart NIC and improves the cloud platform scheduler software. A corresponding agent program runs on the CPU of the smart NIC, which is responsible for communicating with the improved cloud scheduling center. It can report smart NIC information and execute tasks according to the scheduler's instructions. The cloud scheduling center allows the agent to download and run software from the image repository, and can also configure the software parameters. Based on communication with the agent, the cloud scheduling center knows the usage status of each smart NIC, facilitating arbitrary scheduling. Based on the status reported by the agent, it judges the health status of the NIC. When a smart NIC malfunctions, the cloud scheduling center automatically detects it and then migrates the services on it to other smart NICs for operation, such as... Figure 3 As shown.

[0039] The measurement and control algorithm acceleration module includes an FPGA chip and an image repository. For example... Figure 3 As shown, the improved design of the measurement and control intelligent network card includes an FPGA chip, providing user algorithm space. Users can select commonly used measurement and control algorithm libraries provided by the intelligent network card, such as acquisition and decoding algorithms. These will be compiled into a standard algorithm library and uploaded to a mirror repository. The scheduler only needs to communicate with the CPU agent to automatically download algorithms to the intelligent network card. If the existing algorithm library cannot meet the user's needs, the user can design their own algorithm according to the rules, compile it into an image, and publish it to the mirror repository. For acquisition and decoding tasks with very high real-time requirements and large computational demands, the CPU alone is insufficient. This function can be enabled to offload the CPU, freeing it to perform other algorithms and tasks.

[0040] like Figure 1 As shown, a physical network system is built using servers and switches, and a smart network interface card (NIC) is used to replace the traditional NIC. A cloud scheduler, the smart NIC's CPU agent software, and image repository software are installed. The entire process includes the following steps:

[0041] 1) Upload the capture, decoding algorithms, and custom response algorithms of the measurement and control system to the image repository, and configure the cloud platform system to use the algorithm repository.

[0042] 2) Deploy the server software and corresponding accelerator card software by pulling them from the image repository through the cloud scheduling system.

[0043] 3) Data is sent to the accelerator card through the digital front-end device, and the digital front-end can be configured to send data at different rates. The smart network card receives the data and, based on the precise tracking of the data rate, sends the uplink data from the server to the digital front-end at the same rate, and detects the reliability of the uplink data through the digital front-end.

[0044] 4) After receiving downlink data, the smart network card performs calculations using the measurement and control acceleration algorithm loaded on the acceleration card, and then sends the calculation results to the CPU for further processing.

[0045] 5) When a smart network card is manually removed or malfunctions, the scheduling terminal automatically detects it, selects an idle smart network card from the pool, loads the current task image into its agent program, and automatically sets the IP address of the faulty network card through the agent program. The data from the digital front end is directly transmitted to the new smart network card, thereby realizing automatic and rapid migration of services and uninterrupted service.

[0046] It should be noted that, within the scope of protection defined in the claims of this invention, the following embodiments can be combined and / or extended or replaced in any logical manner from the above specific embodiments, such as the disclosed technical principles, disclosed technical features or implicitly disclosed technical features.

[0047] Example 1

[0048] This invention provides a cloud computing measurement and control system based on a smart network interface card (NIC) DPU, comprising:

[0049] The intelligent network interface card (NIC) DPU module performs data transmission and reception matching with the digital radio frequency front-end device based on the transceiver protocol stack; and the intelligent NIC DPU module also includes a digital front-end data adaptive following module, a measurement and control acceleration pooling module, an intelligent NIC cloudification module, and a measurement and control algorithm acceleration module.

[0050] The digital front-end data adaptive following module includes a digital phase-locked loop unit, which is used to track the UDP transmission frame frequency of the digital front-end device and send uplink data to the radio frequency front-end while receiving digital front-end data, using the restored data rate.

[0051] The measurement and control acceleration pooling module includes an interconnection module, which connects multiple smart network cards.

[0052] The smart network card cloudification module includes a CPU and cloud platform scheduler software. A corresponding agent program runs on the CPU of the smart network card cloudification module, and the agent program communicates with the cloud scheduling center.

[0053] The measurement and control algorithm acceleration module FPGA chip and image repository utilize the FPGA chip to provide user algorithm space, allowing users to select the measurement and control algorithm library provided by the smart network card, create a standard algorithm library, and upload it to the image repository.

[0054] Example 2

[0055] Based on Embodiment 1, a digital radio frequency front-end device is included for sending data to the smart network card DPU module.

[0056] Example 3

[0057] Based on Embodiment 1, the digital phase-locked loop unit includes a fully digital phase-locked loop unit (DPLL). Furthermore, while receiving downlink data, the DPLL unit is used to lock the frequency of the received data frame. Uplink data is then retrieved from the data buffer and sent according to this frequency, ensuring that the uplink and downlink rates are consistent.

[0058] Example 4

[0059] Based on Example 1, the interconnection module uses the accelerator's 100G optical port for interconnection, enabling external network processing and exchange of external data, and internal network processing for collaboration between multiple cards, thus achieving interoperability between any smart network cards.

[0060] Example 5

[0061] Based on Example 1, for low-speed data, the interconnect module communicates directly with the server via the PCIe bus.

[0062] Example 6

[0063] Based on Example 1, the agent program reports smart network interface card (NIC) information and executes tasks according to the scheduler's instructions; the cloud scheduling center enables the agent program to download and run software from the image repository, and also enables the software to configure parameters. Based on the status reported by the agent program, the health status of the smart NIC is determined. When a smart NIC malfunctions and is detected by the cloud scheduling center, its services are migrated to other smart NICs for operation.

[0064] Example 7

[0065] Based on Example 6, the cloud scheduling center communicates with the CPU's agent program to download algorithms to the smart network card; if the existing algorithm library cannot meet the user's needs, the user can design an algorithm according to the rules, compile it into an image, and publish it to the image repository.

[0066] Example 8

[0067] A cloud computing measurement and control method based on a smart network interface card (DPU), based on any of the systems described in Examples 1 to 7, further includes the following steps:

[0068] S1, upload the algorithm software of the measurement and control system to the image repository, and configure the cloud platform system to use the image repository;

[0069] S2 uses a cloud scheduling system to pull server software and corresponding accelerator card software from the image repository for deployment.

[0070] S3 sends data to the accelerator card through the digital front-end device and is configured to send data at different rates. The smart network card DPU module receives the data and, based on the tracking of the data rate, sends the uplink data from the server to the digital front-end at the same rate, and detects the reliability of the uplink data through the digital front-end.

[0071] S4, after the smart network card DPU module receives downlink data, it performs calculations using the measurement and control acceleration algorithm loaded on the acceleration card, and then sends the calculation results to the CPU for further processing.

[0072] Example 9

[0073] Based on Example 8, after step S4, the following step is also included:

[0074] S5: When a smart network card is manually removed or malfunctions, the scheduling terminal detects the fault, selects an idle smart network card from the pool, loads the current task image into its agent program, and sets the IP address of the faulty network card through the agent program. The data from the digital front end is directly transmitted to the new smart network card, thereby realizing automatic and rapid migration of services and uninterrupted service.

[0075] Example 10

[0076] Based on Example 8, in step S1, the algorithm software includes capture algorithm software, decoding algorithm software, and custom response algorithm software.

[0077] The units described in the embodiments of the present invention can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0078] According to one aspect of the present invention, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and executes the computer instructions, causing the computer device to perform the methods provided in the various optional implementations described above.

[0079] In another aspect, embodiments of the present invention also provide a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the methods described in the above embodiments.

[0080] All parts not covered in this invention are the same as or can be implemented using existing technologies.

[0081] The above technical solution is only one embodiment of the present invention. For those skilled in the art, based on the application methods and principles disclosed in the present invention, it is easy to make various types of improvements or modifications, and not limited to the methods described in the above specific embodiments of the present invention. Therefore, the methods described above are only preferred and are not restrictive.

[0082] In addition to the examples above, other embodiments may be obtained by those skilled in the art based on the above disclosure or by making modifications using knowledge or technology in related fields. The features of each embodiment may be interchanged or replaced. Modifications and changes made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

Claims

1. A cloud computing measurement and control system based on a smart network interface card (NIC) DPU, characterized in that, include: The intelligent network interface card (NIC) DPU module performs data transmission and reception matching with the digital radio frequency front-end device based on the transceiver protocol stack. Furthermore, the intelligent network card DPU module also includes a digital front-end data adaptive following module, a measurement and control acceleration pooling module, an intelligent network card cloudification module, and a measurement and control algorithm acceleration module; The digital front-end data adaptive following module includes a digital phase-locked loop unit, which is used to track the UDP transmission frame frequency of the digital front-end device and send uplink data to the radio frequency front-end while receiving digital front-end data, using the restored data rate. The measurement and control acceleration pooling module includes an interconnection module, which connects multiple smart network cards. The smart network card cloudification module includes a CPU and cloud platform scheduler software. A corresponding agent program runs on the CPU of the smart network card cloudification module, and the agent program communicates with the cloud scheduling center. The FPGA chip and image repository of the measurement and control algorithm acceleration module utilize the FPGA chip to provide user algorithm space, allowing users to select the measurement and control algorithm library provided by the smart network card, create a standard algorithm library, and upload it to the image repository; The agent program reports smart network interface card information and executes tasks according to the scheduler's instructions; The cloud dispatch center enables the agent program to download and run software from the image repository, and also enables the software to configure parameters. Based on the status reported by the agent program, it can determine the health status of the smart network card. When a smart network card malfunctions and is detected by the cloud dispatch center, its services are migrated to other smart network cards for operation. The cloud scheduling center communicates with the CPU's agent program to download the algorithm to the smart network card; If the existing algorithm library cannot meet the user's needs, then the user can design their own algorithm according to the rules, compile it into an image, and publish it to the image repository.

2. The cloud computing measurement and control system based on a smart network interface card (NIC) DPU according to claim 1, characterized in that, This includes digital radio frequency front-end equipment used to send data to the smart network interface card (DPU) module.

3. The cloud computing measurement and control system based on a smart network interface card (NIC) DPU according to claim 1, characterized in that, The digital phase-locked loop unit includes a fully digital phase-locked loop unit (DPLL). While receiving downlink data, the DPLL unit is used to lock the frequency of the received data frame. Uplink data is then retrieved from the data buffer and sent according to this frequency, ensuring that the uplink and downlink rates are consistent.

4. The cloud computing measurement and control system based on a smart network interface card (NIC) DPU according to claim 1, characterized in that, The interconnection module uses the accelerator's 100G optical port for interconnection, enabling external network processing and data exchange with the external network, and internal network processing for collaboration between multiple cards, thus achieving interoperability between any smart network cards.

5. The cloud computing measurement and control system based on a smart network interface card (NIC) DPU according to claim 1, characterized in that, For low-speed data, the interconnect module communicates directly with the server via the PCIe bus.

6. A cloud computing measurement and control method based on a smart network interface card (NIC) DPU, characterized in that, Based on the system according to any one of claims 1 to 5, the system further includes the following steps: S1, upload the algorithm software of the measurement and control system to the image repository, and configure the cloud platform system to use the image repository; S2 uses a cloud scheduling system to pull server software and corresponding accelerator card software from the image repository for deployment. S3 sends data to the accelerator card through the digital front-end device and is configured to send data at different rates. The smart network card DPU module receives the data and, based on the tracking of the data rate, sends the uplink data from the server to the digital front-end at the same rate, and detects the reliability of the uplink data through the digital front-end. S4, after the smart network card DPU module receives downlink data, it performs calculations using the measurement and control acceleration algorithm loaded on the acceleration card, and then sends the calculation results to the CPU for further processing.

7. The cloud computing measurement and control method based on a smart network interface card (NIC) DPU according to claim 6, characterized in that, Following step S4, the following steps are also included: S5: When a smart network card is manually removed or malfunctions, the scheduling terminal detects the fault, selects an idle smart network card from the pool, loads the current task image into its agent program, and sets the IP address of the faulty network card through the agent program. The data from the digital front end is directly transmitted to the new smart network card, thereby realizing automatic and rapid migration of services and uninterrupted service.

8. The cloud computing measurement and control method based on a smart network interface card (NIC) DPU according to claim 6, characterized in that, In step S1, the algorithm software includes capture algorithm software, decoding algorithm software, and custom response algorithm software.

Images