Cloud PC server

Abstract

The present application relates to the field of computers. Disclosed is a cloud PC server. The cloud PC server is a 3U standard rack-mount server, and comprises compute blades, a fan module, power source modules, a network switching module, and a master control management module responsible for monitoring and managing operating states of system hardware, all of which are disposed in a chassis, wherein 24 compute blades are provided, and each compute blade can operate independently; the fan module uses a front-to-rear airflow heat dissipation mode; two power source modules are provided, allowing current sharing and hot standby; and the network switching module is provided with two networks which are independent of each other and physically isolated. The present application has good expandability and universality and high reliability.

Description

A cloud PC server

[0001] This application claims priority to Chinese Patent Application No. 202510230902.2, filed on February 28, 2025, entitled “A Cloud PC Server”, the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of computer technology, and in particular to a cloud PC server. Background Technology

[0003] Cloud PC servers are one of the most core components of a PC system. They centrally deploy traditional physical hosts in data centers as computing blades, providing users with a user experience almost identical to that of a local PC. Through network isolation and data security measures, they can significantly reduce the risks of network crosstalk and data leakage. However, with the increasing demand for high-performance computing in areas such as big data, artificial intelligence, graphic design, and cloud gaming, cloud PC servers need to be equipped with more powerful hardware resources, such as high-performance CPUs, GPUs, and large-capacity memory. Therefore, optimizing hardware design to meet the current technological requirements for high-performance computing, thermal management, storage and network optimization, energy efficiency, and security and reliability has become crucial for improving the overall performance of cloud PC servers. Summary of the Invention

[0004] The purpose of this application is to provide a cloud PC server that aims to solve at least one of the technical problems existing in the prior art and can effectively improve the overall performance of the machine.

[0005] To achieve the above objectives, this application provides the following solution:

[0006] This application provides a cloud PC server, which is a 3U standard rack-mount server, including: computing blades with a chassis, a fan module, a power module, a network switching module, and a main control management module responsible for monitoring and managing the operating status of the system hardware; the computing blades are provided with 24 blades, each of which can operate independently; the fan module adopts a front-to-back airflow cooling method; the power module has two units and allows for current sharing and hot backup; the network switching module has two independent and physically isolated networks.

[0007] In one embodiment, the chassis includes a base plate with a single-star structure, two switching boards serving as a service data network and a terminal data network respectively, two fan assemblies for blowing and exhausting air respectively, a front panel with a debugging network port and a power switch, and a rear panel with a pluggable power module interface, an optical fiber interface, and an Ethernet port; the base plate has 24 blade slots for connecting the computing blades and two switching board slots for connecting the switching boards.

[0008] In one embodiment, the computing blade is equipped with a computing unit that supports various sizes of M.2 SSD hard drives, various memory sizes and is compatible with various CPUs, an encoding unit for accessing audio and video and acting as a controller, and an interface unit for providing external interfaces for power supply, SERDES, I2C and slot I / O.

[0009] In one embodiment, the interface unit includes a network submodule and an IPMI submodule; the network submodule includes two independent PHY chips for service transmission and data management respectively; the IPMI submodule reads the temperature status of the computing unit in real time through the UART1 port, and can control the computing unit to power on and off and reset through control I / O, and read the power-on / off, temperature and power status of the entire board, and communicates with the encoding unit through the UART2 port.

[0010] In one embodiment, the computing blade is provided with an input interface submodule, a data processing submodule, and an output interface submodule; the input interface submodule is used for HDMI video data decoding; the data processing submodule is used for strip segmentation, pixel reconstruction, and video buffering; and the output interface submodule is used for HDMI video data encoding.

[0011] In one embodiment, the main control management module includes a health detection unit, a system information unit, a control unit, and a communication interface unit. The health detection unit connects to the power module via PMBus to monitor the overall power status, collects module temperature via an I2C temperature sensor, has a dedicated watchdog reset chip to improve system reliability, and uses a built-in ADC to monitor the real-time voltage of the 12V power supply. The system information unit stores or reads FRU information via EEPROM and uses a built-in RTC real-time clock in the MCU. The control unit adjusts the fan speed via PWM output, collects and monitors the fan speed in real time, detects whether the computing blades and network switching board are in place, and controls power-on / off.

[0012] In one embodiment, the main control management module uses the uCOS II system and has two threads, one for power button detection and the other for parsing and processing serial port messages received from the network switching module.

[0013] In one embodiment, the network switching module is provided with a service data network for network data transmission between PCs, and a terminal access network for transmitting media streams of screen images and audio data, as well as mouse and keyboard control signals.

[0014] In one embodiment, the network switching module includes a downlink network interface unit, a data switching unit, an uplink network interface unit, and an IPMC unit. The downlink network interface unit converts six QSGMII interfaces into 24 1000Base-X SerDes interfaces via six PHY chips and connects them to 24 slots on the baseboard. It also connects the baseboard's 12V power supply, slot number, and IPMI bus to the network switching module via an ADF connector. The data switching unit is used for Layer 2 switching of network data packets. The uplink network interface unit includes data interface units on the front and rear panels. The IPMC unit, as the management unit of the network switching board, communicates with the main control management module of the cloud PC server via an MCU, monitors the network switching module's temperature, voltage, and other health information in real time, implements power-on timing control via an FPGA, and parses the serial signals controlling the network port LEDs.

[0015] In one embodiment, the system also includes a cloud PC management platform for online upgrades. The cloud PC management platform sends an upgrade command to the network switching module. The network switching module controls the main control management module to reset and enter bootload mode. Then, the network switching module transmits firmware to the main control management module. After the firmware is received and burned, the program runs to complete the upgrade.

[0016] Based on the specific embodiments provided in this application, the following technical effects are disclosed:

[0017] The cloud PC server described in this application adopts a modular architecture design, supporting hot-swapping and flexible configuration of 24 computing blade units, giving the system excellent scalability and versatility. The device employs air cooling, with two sets of fans internally, one blowing air and the other exhausting it, forming a straight-through airflow channel within the device. It uses a common front-to-back airflow configuration to improve heat dissipation capacity. The front panel features power, operation, management, and service status indicators, providing a clear view of the device's operating status. The front panel also includes a debugging network port and a power switch, while the rear panel features a pluggable power module, fiber optic interface, and Ethernet port. The baseboard integrates only a small number of gate circuits for I / O expansion and isolation protection, resulting in very low power consumption, while providing sufficient current carrying capacity. Each computing blade is a pluggable, independent module, providing computing and data processing capabilities. Users can select different CPU, memory, and storage configurations as needed. Each computing blade operates independently, providing high-performance computing and data processing capabilities. The power supply module is designed with 1+1 current sharing and hot backup. There are two power supply modules, which work simultaneously and balance the load. In the 1+1 configuration, one of the power supplies can be replaced or maintained without affecting the normal operation of the server, thus improving the maintainability of the system. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 is a module connection diagram of a cloud PC server according to an embodiment of this application.

[0020] Figure 2 is a functional structure diagram of a cloud PC server according to an embodiment of this application.

[0021] Figure 3 is a hardware structure diagram of a cloud PC server according to an embodiment of this application.

[0022] Figure 4 is a diagram of the baseboard topology of a cloud PC server according to an embodiment of this application.

[0023] Figure 5 is a diagram of the baseboard interface structure of a cloud PC server according to an embodiment of this application.

[0024] Figure 6 is a hardware architecture diagram of a cloud PC computing blade of a cloud PC server according to an embodiment of this application.

[0025] Figure 7 is a logical structure block diagram of a cloud PC computing blade of a cloud PC server according to an embodiment of this application.

[0026] Figure 8 is a voltage module structure diagram of a cloud PC server according to an embodiment of this application.

[0027] Figure 9 is a structural diagram of the main control management module of the cloud PC server according to an embodiment of this application.

[0028] Figure 10 is a diagram of the dual-network isolation network connection of a cloud PC server according to an embodiment of this application.

[0029] Figure 11 is an understanding block diagram of the network switching module of a cloud PC server according to an embodiment of this application.

[0030] Figure 12 is a diagram of the external hardware interface of a cloud PC server according to an embodiment of this application.

[0031] Figure 13 is a structural diagram of the front panel of a cloud PC server according to an embodiment of this application.

[0032] Figure 14 is a structural diagram of the rear panel of a cloud PC server according to an embodiment of this application.

[0033] Figure 15 is a flowchart of the software upgrade process for a cloud PC server according to an embodiment of this application.

[0034] Reference numerals: Chassis - 100, Front panel - 110, Rear panel - 120, Base plate - 200, Computing blade - 300, Switching board - 400, Fan assembly - 500. Detailed Implementation

[0035] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0036] The following will provide a clear and complete description of the concept, specific structure, and resulting technical effects of this application in conjunction with embodiments and accompanying drawings, so as to fully understand the purpose, solution, and effects of this application. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0037] It should be noted that, unless otherwise specified, when a feature is referred to as "fixed" or "connected" to another feature, it can be directly fixed or connected to the other feature, or indirectly fixed or connected to the other feature. Furthermore, the descriptions of "upper," "lower," "left," "right," "top," and "bottom" used in this application are only relative to the relative positional relationships of the various components of this application in the accompanying drawings.

[0038] Furthermore, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of this application. The term "and / or" as used herein includes any combination of one or more of the associated listed items.

[0039] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various elements, these elements should not be limited to these terms. These terms are only used to distinguish elements of the same type from one another. For example, without departing from the scope of this disclosure, a first element may also be referred to as a second element, and similarly, a second element may also be referred to as a first element.

[0040] Referring to Figures 1 to 15, the cloud PC server of this application is a 3U standard rack-mount server, which includes computing blades 300 with a chassis 100, a fan module, a power supply module, a network switching module, and a main control management module responsible for monitoring and managing the operating status of the system hardware; there are 24 computing blades 300, each of which can operate independently; the fan module adopts a front-to-back airflow cooling method; there are two power supply modules, which allow for current sharing and hot backup; the network switching module has two independent and physically isolated networks.

[0041] Referring to Figures 1 and 2, the cloud PC management platform of the cloud PC server in this application is divided into an application layer, an interface layer, and a system layer, providing a complete chassis 100 management solution to ensure the stable operation, status monitoring, and anomaly handling of each functional module within the chassis 100, providing efficient and stable support for chassis 100 management in various business scenarios. The cloud PC server in this application is a 3U standard rack-mount server, adopting a modular design concept. The overall structure of the application layer is divided into a power module, a network switching module, a main control management (fan) module, a baseboard 200, and computing blades 300. This system has good scalability and versatility; a single cloud PC server can be configured with up to 24 computing blades 300 and supports server cascading expansion to meet the needs of larger user volumes. Furthermore, the computing blades 300 can be inserted into any 24 slots.

[0042] In some embodiments, referring to Figure 2, the application layer includes computing blades 300, a fan module, a network switching module, a main control management module, and a power supply module. The 24 computing blades 300 are used for slot number detection, power-on / off status detection, temperature detection, presence detection, power-on / off control, and resource usage detection. The fan module is used for fan speed control and fan speed detection. The network switching module is used for temperature detection and anomaly alarms. The main control management module is used for chassis ID reporting, power button detection, and online upgrade management. The power supply module is used for power status detection.

[0043] Specifically, the compute blade 300 (encoder board) sends its MAC address to the network switching module (terminal access network). The network switching module (terminal access network) matches the MAC address with the slot number of chassis 100 where the compute blade 300 (encoder board) is located and returns the result to the compute blade 300 (encoder board) to achieve slot number detection. In the power-on / off status detection, the power-on status of the compute blade 300 is read from the contents of the IO expansion chip register to determine whether the compute blade 300 is powered. If the level is low, it is powered on; otherwise, it is powered off. The MCU on the compute blade 300 reads the x86 / Phytium CPU temperature via serial port and sends it to the main control management module via the IPMB bus to complete temperature detection. It also checks whether the compute blade 300 is inserted in chassis slot 100 to complete presence detection. In the power-on / off control, the main control management module controls the IO expansion chip to output high and low levels to power on and off the compute blade 300. The MCU on the compute blade 300 obtains x86 / Phytium resources, including the number of CPU cores, memory capacity, and storage capacity, via serial port to achieve resource usage detection.

[0044] The MCU ID of the main control management module serves as the unique identifier of the chassis and is reported to the compute blades 300 (encoding board) and the cloud PC management platform to complete the chassis ID reporting. In the power button detection, when the cloud PC server chassis 100 is powered on normally, the first step is to check if the power button is pressed. If not pressed, only the main control management module powers on; other boards do not power on, and the fans do not run. After pressing, the network switching module powers on first, followed by the first row of eight compute blades 300, then the second and third rows, with the fans starting to run. After pressing the power button, all boards power on and the fans run. Pressing and holding the power button again initiates the normal power-off process, where all boards except the main control management module power off sequentially: the network switching module first, then the first row of eight compute blades 300, then the second and third rows, with the fans stopping. During online upgrades, the cloud PC management platform sends upgrade commands to the network switching module (terminal access network). The network switching module (terminal access network) controls the main control management module to reset and enter bootloader mode. Then, the network switching module (terminal access network) transmits firmware to the main control management module, and the firmware is then run to complete the upgrade. The system layer mainly includes the cloud PC management platform, which is primarily used for slot management, fan management, interface control, chassis information management, and online upgrades. The main control management module uses the uCOS II system and mainly operates two threads: one for power button detection and the other for parsing and processing serial port messages received from the network switching module (terminal access network).

[0045] The network switching module is used for temperature detection and anomaly warnings. The temperature detection configuration of the network management module is as follows: the MCU on the network switching module reads the temperature from the temperature sensor, and the MCU reports the temperature reading to the main control management module via the IPMB bus. The anomaly warning configuration of the network management module is as follows: the main control management module sends fan alarm, temperature alarm, and power alarm information to the network switching module (terminal access network) via serial port. If the computing blade 300 (encoding card) cannot ping the gateway, i.e., it is not connected to the network switching module, it then reports a network anomaly alarm to the network switching module (terminal access network) via serial port, and then reports to the cloud PC management platform. If the network switching module is also unreachable, the anomaly needs to be investigated via the debugging serial port. If there is a chassis malfunction, to quickly locate the chassis, the chassis alarm light will illuminate upon receiving a chassis 100 name call command from the cloud PC management platform of the network switching module (terminal access network).

[0046] Referring to Figure 3, the hardware of the cloud PC server in this application includes a server chassis 100, two fan assemblies 500, a fan board, a switching board 400, a base plate 200, and 24 computing blades 300. The number of computing blades 300 is configured according to the product form, and a maximum of 24 computing blades 300 are supported.

[0047] Specifically, referring to Figure 3, the cloud PC server of this application adopts a sheet metal chassis 100 design, consisting of a chassis 100 body, a front panel 110, a top cover, a removable cover plate, a handle, a fan assembly 500, and auxiliary materials. The use of a metal sheet metal frame reduces costs while maintaining ease of use, reliability, and maintainability. Specifically, the device adopts a standard 19-inch 3U height design, with plug-in internal cards, providing up to 24 service cards for simultaneous plug-in. The cards can be quickly opened for maintenance. The number of computing blades 300 contained in this application's 3U standard rack-mount cloud PC server is significantly higher than that of currently available 3U standard rack-mount servers. The device uses air cooling, with two sets of fans inside, one blowing air and the other exhausting air, forming a straight-through airflow channel within the device. The common front-to-back airflow method improves heat dissipation capacity. The front panel 110 is designed with power, operation, management and service indicators to intuitively reflect the equipment's operating status. The front panel 110 is designed with a debugging network port and a power switch, while the rear panel 120 is designed with a pluggable power module, fiber optic interface and Ethernet port.

[0048] In some embodiments, the baseboard 200 of the cloud PC server of this application is provided with 24 computing blade 300 slots and two network switch board 400 slots. The main control management module and the light board are connected by cables. The baseboard 200 adopts a single star topology. The signal topology diagram of the two switch boards 400 and the 24 computing blade 300 slots is shown in Figure 4. Specifically, the power interface for providing 12V power is connected to the two network switch board 400 slots, the 24 computing blade 300 slots, and the main control management unit. The 3.3V_STB power supply is connected to the two network switch board 400 slots, the 24 computing blade 300 slots, and the main control management unit. The SerDes network is connected to the 24 computing blade 300 slots from the second network switch board 400 slot through the terminal access network, and to the 24 computing blade 300 slots from the first network switch board 400 slot through the service data network. The IPMB signal passes through the two network switch board 400 slots and the 24 computing blade 300 slots. Two UART signals are connected to the main control management unit via two network switch board 400 slots respectively. Multiple PWR_EN&DET signals are connected to the main control management module from 24 blade slots respectively.

[0049] Figure 5 shows the external interface structure of the baseboard 200, including its 12V power supply, SerDes network, UART serial port, IPMB, and 3.3V_STB interface. Specifically, the 12V power supply module is directly connected to the power interface of the baseboard 200 via soldered terminals. The baseboard 200 has two sets of ADF connectors, used as SerDes network ports and UART interfaces for the two network switching modules, with each set containing three ADF connections. The baseboard 200 has 24 PCIe x4 connectors, used as blade interfaces to connect 24 computing blades 300, serving as interfaces for the two SerDes networks, IPMB signals, 3.3V_STB power supply, and 12V power supply, and also has 24 ADF connectors as power supply interfaces for the 24 computing blades 300. The baseboard 200 uses 6-pin and 8-pin connectors as IPMB & 3.3V interfaces and UART interfaces to connect to the main control management module.

[0050] To improve the reliability of the baseboard 200, only a small number of gate circuits for I / O expansion and isolation protection are integrated on the baseboard 200 of this application, resulting in very low power consumption. The main function of the baseboard 200 of this application is to provide a sufficiently large current carrying capacity. Furthermore, the connection between the power module and the baseboard 200 can be achieved by using a copper PCB soldered crimp nut, which has a 260A overcurrent capacity. Additionally, the baseboard 200 supplies power to the computing blade 300 through the ADF power connector, enabling a single ADF power connector to provide a power supply capacity of 40A.

[0051] In some embodiments, the computing blade 300 is a pluggable, independent module that provides computing and data processing capabilities. Users can select different CPU, memory, and storage configurations as needed. Furthermore, it may include CPUs (non-domestic Intel i3 / i5 / i7 and domestic Phytium D2000), memory capacity (8G / 16G / 32G), and hard disk capacity (512G / 1T / 2T), which can be flexibly configured according to user needs.

[0052] Each cloud PC server can be configured with up to 24 compute blades 300, each operating independently to provide high-performance computing and data processing capabilities. User operational data is stored within their respective compute blades 300, while backups of all user data are stored on a dedicated NAS storage server. Therefore, the cloud PC server itself does not handle backup data storage and can focus on providing real-time computing and processing capabilities. Specifically, the cloud PC server chassis 100 has 24 blade slots, supporting hot-swapping and flexible configuration of 24 compute blade 300 units, ensuring that plugging or unplugging one compute blade 300 does not affect the normal operation of other compute blades 300.

[0053] In one application embodiment, the computing blade 300 of the cloud PC server of this application adopts a PCIe interface and a blade-shaped structure. The airflow of the compatible chassis 100 adopts a front-to-back air intake and exhaust configuration. The computing blade 300 adopts a universal coding card base plate 200 design, and different blade modules can be selected and replaced as needed. Specifically, the cloud PC computing blade 300 is installed inside the server chassis 100, with 8 blades arranged in each row, for a total of 24 blades in 3 rows. Further, the external dimensions of the computing blade 300 can be 175.25×126.08×38mm. The external interface of the cloud PC computing blade 300 includes PCIe x4 gold fingers, power connectors, positioning pins, and mounting screws. The internal components of the cloud PC computing blade 300 mainly consist of a mounting bracket, a coding card PCBA, a coding card heatsink, a decorative cover, and either a purchased Phytium computing module or a purchased x86 module; the Phytium computing module and the x86 module can be configured separately.

[0054] In one application embodiment, the fan module supports intelligent fan speed adjustment. The main control management module monitors the system temperature in real time and adjusts the fan speed to optimize heat dissipation and energy efficiency, reducing noise and power consumption. The cloud PC server uses forced air cooling. With a maximum power of 2000W, the design employs two sets of fans: one for blowing air and one for exhausting air. The fans are high-speed, high-airflow axial fans (8038 type). The entire system uses a front-to-back airflow cooling method. It should be noted that, through structural stacking and thermal design analysis and verification using simulation software, this application demonstrates that it can handle the heat dissipation capacity of a single 60W × 24 blade 300 module. Under simulated conditions, the highest CPU case temperature is approximately 62 degrees Celsius, which meets operational requirements.

[0055] Specifically, the heatsink of the Cloud PC Computing Blade 300 encoding card is in contact with high-power chips such as the CPU and FPGA using a thermally conductive silicone pad, which conducts heat to the heatsink teeth and is then carried away by the airflow of the chassis 100. The Phytium computing module and the X86 computing module also use conduction to conduct heat to the matched heatsink, which is then carried away by the airflow of the chassis 100. Among them, the X86 module has high power consumption and uses a fin + VC heat dissipation plate design to effectively improve the heat dissipation capacity of the product.

[0056] In one application embodiment, the cloud PC computing blade 300 of this application is provided with a computing unit, an encoding unit, and an interface unit. The computing unit is the core of the system, including support for various specifications of M.2 SSD hard drives, various memory specifications, and compatibility with various CPUs. The encoding unit includes two interface chips for audio and video, as well as an RK3588 controller, which is used to implement functions such as compression encoding of I2S audio and various video interfaces, permission processing and transmission of USB signals, and to send the above data to the cloud PC user terminal through the network port, and an FPGA, which is used to implement video channel striping, pixel reconstruction, caching and synthesis. The interface unit provides external interfaces such as power supply, SERDES, I2C and slot IO to ensure the system's multifunctionality and scalability.

[0057] The network submodule of the interface unit contains two independent PHY chips, namely I210-IS and YT8521SH. I210-IS connects to the computing unit via a PCIe interface for service transmission; YT8521SH connects to the encoding unit via an RGMII interface for data management, thus achieving physical isolation between the two networks. In the IPMI submodule of the interface unit, the I2C and slot detection I / O interfaces have added TVS diode protection, and the ADF power connector has a maximum supply current of 14A. The IPMI module can read the computing unit's temperature status in real time via UART1 and can control the computing unit's power on / off and reset via control I / O. It can also read the overall board's power-on / off status, temperature, and power status, and then communicate with the encoding unit via UART2 to upload data to the cloud PC management platform.

[0058] In one application embodiment, the cloud PC computing blade 300 of this application is used to implement functions such as interface conversion, strip segmentation, pixel reconstruction, video data caching, and video compositing for two video channels. Referring to Figures 6 and 7, the two video channels reuse the same code module. Taking one video channel as an example, its overall logical design consists of three parts: an input interface submodule, a data processing submodule, and an output interface submodule. The input interface submodule mainly implements the function of HDMI video data decoding; the data processing submodule mainly implements functions such as strip segmentation, pixel reconstruction, and video caching; and the output interface submodule implements the function of HDMI video data encoding.

[0059] Specifically, the input interface submodule stores EDID information and generates a DDC channel for PC reading, as well as converting the input HDMI interface to an RGB_TTL interface. The data processing submodule slices the input full frame into four strip frames, for example, converting the input video format from 1920*1080@60fps to 1920*272@240fps, and adjusts the YUV component storage format based on the RK encoder NV12 encoding rules and the RK RX interface memory mapping method. It also features data buffering and clock domain isolation. The output interface submodule generates the header and body of the TMDS transmission embedded data island, generates the control signals embedded in the TMDS transmission, and integrates the data island, control signals, and video data content, converting it to HDMI output.

[0060] In some embodiments, the power module converts 220VAC power to 12V DC power, with a maximum power consumption of 2000W per module. The power module supports 1+1 current sharing hot backup, ensuring stable system power supply even in the event of a single power module failure, exhibiting high reliability and strong continuous operation capability. Furthermore, the power module can provide two 100-240VAC, 50 / 60Hz inputs and supports redundant backup.

[0061] Referring to Figure 8, the cloud PC server power supply module is designed with 1+1 current sharing and hot backup. Two power supply modules are configured, operating simultaneously and balancing the load. Through current sharing, the load is evenly distributed between the two power supply modules during normal operation, reducing the load pressure on any single module and extending its lifespan. Since both power supply modules are operating simultaneously, there is no need to wait for a switchover process in case of a failure; the system can seamlessly transition to the backup power supply, preventing server downtime due to power failure. In the 1+1 configuration, one power supply can be replaced or maintained without affecting normal server operation, improving system maintainability.

[0062] In some embodiments, the main control management module is responsible for monitoring and managing the operating status of system hardware. By monitoring key parameters such as temperature, voltage, and current in real time, it helps ensure system stability under various operating conditions. The main control management module of this application possesses comprehensive hardware monitoring and management capabilities, monitoring key parameters such as temperature and voltage in real time, supporting fault detection and event logging, and effectively preventing and responding to hardware failures. Combined with the communication function of the network switch board 400, the main control management module can provide remote power management of the server chassis 100 and control of the computing blades 300 power-on / off, enabling administrators to remotely maintain the system and troubleshoot faults.

[0063] In one application embodiment, referring to Figure 9, the main control management module includes a health detection unit, a system information unit, a control unit, and a communication interface unit. The health detection unit connects to the power module via a PMBus to monitor the overall power status of the cloud PC server. It collects module temperature via an I2C temperature sensor, incorporates a dedicated watchdog reset chip to improve system reliability, and uses a built-in ADC to monitor the real-time voltage of the 12V power supply. The system information unit uses an EEPROM to store or retrieve FRU information and a built-in RTC real-time clock in the MCU. The control unit adjusts the fan speed via PWM output, collects and monitors the fan speed in real time, and detects whether the computing blades 300 and the network switching board 400 are in place, controlling their power-on / off states.

[0064] In some embodiments, referring to Figure 10, the network switching module consists of two independent switching boards 400: a business data center network switching board 400 and a terminal access network switching board 400. These two networks are independent and physically isolated, ensuring efficient and secure data transmission. The network switching module implements Ethernet Layer 2 switching functionality, providing one 10 Gigabit optical port and one Gigabit electrical port. The cloud PC server internally supports 24 Gigabit SerDes interfaces, corresponding one-to-one with each of the 24 computing blades 300, ensuring high-speed data transmission and efficient communication between the computing nodes within the cloud PC server.

[0065] In one application embodiment, the business data center network switching board 400 and the terminal access network switching board 400 in the network switching module correspond to the business data network and the terminal access network, respectively. Referring to the two sets of network connection relationships in Figure 1, the business data network is used for network data transmission between PCs; the terminal access network is used to transmit media streams of screen images and audio data, as well as control signals such as mouse and keyboard.

[0066] It should be noted that the service data network switching board 400 and the terminal access network switching board 400 are completely identical in hardware design, possessing the same switching capabilities and interface configurations. Specifically, they are logically distinguished by the slot number on the baseboard 200. Specifically, the service data network switching board 400 is inserted into slot 1 of the baseboard 200, and the terminal access network switching board 400 is inserted into slot 2 of the baseboard 200. The system automatically identifies whether it serves as the service data network or the terminal access network based on the slot number.

[0067] In one application embodiment, the network switching module includes a downlink network interface unit, a data switching unit, an uplink network interface unit, and an IPMC unit. Referring to Figure 11, the downlink network interface unit converts six QSGMII interfaces into 24 1000Base-X SerDes interfaces via six PHY chips, connecting to 24 slots on the baseboard 200. The baseboard 200's 12V power supply, slot numbers, and IPMI bus are connected to the switching module via ADF connectors. The data switching unit is responsible for Layer 2 switching of network data packets, employing the NF5180 switching chip with an integrated A53 ARM dual-core processor. The uplink network interface unit includes data interface units on the front panel 110 and rear panel 120. A 10G fiber optic cable and a 1000M Ethernet port are connected to the rear panel 120, while the A53 ARM dual-core processor's 1000M maintenance Ethernet port and debugging serial port are connected to the front panel 110. The IPMC unit, serving as the management unit of the network switching board 400, communicates with the main control management module of the cloud PC server via the MCU. It monitors the network switching module's health information, such as temperature and voltage, in real time, and uses the FPGA to implement power-on timing control and parses the serial signals controlling the network port LEDs. Specifically, the network switching module interfaces are shown in Table 1 below.

[0068] Table 1 Network Switching Module Interface Information Table

[0069] The main chips of the network switching module in this application include a switching chip, an MCU chip, an FPGA chip, and six PHY chips. The power consumption of the entire module is about 15W, which can effectively achieve low power consumption.

[0070] Referring to Figure 12, the panel interface settings of the cloud PC server in this application are as follows:

[0071] Data network ports: 1 10 Gigabit optical port, 1 gigabit electrical port, physically isolated from the user network;

[0072] User network ports: 1 10 Gigabit optical port, 1 gigabit electrical port, physically isolated from the data network;

[0073] Debugging network ports: 2 gigabit Ethernet ports, with physical isolation between the two ports;

[0074] Debugging serial port: Supports 2 serial ports (RJ45 connector);

[0075] Power input: Supports 2 AC 220V inputs;

[0076] Indicator lights: 2 data network port status lights, 2 user network port status lights, 1 alarm light, and 1 running light;

[0077] Power switch: One power switch is installed on the front panel 110.

[0078] Specifically, referring to Figures 13 and 14, the front panel 110 is equipped with 4 debugging network ports, 6 indicator lights, and 1 power switch, while the rear panel 120 is equipped with 2 pluggable module power supplies, 2 service optical ports, 2 service network ports, and 1 grounding screw post. The power supply adopts a 1+1 backup design. The interface settings are shown in Table 2 below.

[0079] Table 2 Interface Information Table

[0080] In the cloud PC management platform of this application's cloud PC server, the computing blade 300 (encoding board) sends a slot acquisition request and its MAC address to the network switching module (terminal access network). The network switching module (terminal access network) matches the corresponding slot number based on the MAC address and returns it to the computing blade 300 (encoding board). The computing blade 300 (encoding board) then reports it to the cloud PC management platform. The cloud PC management platform can then display the card's position information and power-on / off status in the corresponding slot. In the default automatic mode, the main control management module collects the temperature data reported by the computing blade 300 (encoding board), controls the fan speed level, and sends request messages to the main control management module through the cloud PC management platform, including setting the power-on / off status of the computing blade 300 (encoding board) and setting the fan speed, thereby realizing the management and monitoring of the baseboard 200 slots, fans, and interfaces.

[0081] When the cloud PC management platform of this application needs to be upgraded online, as shown in Figure 15, the cloud PC management platform sends an upgrade command to the network switching module (terminal access network). The network switching module (terminal access network) controls the main control management module to reset and enter bootload. Then, the network switching module (terminal access network) transmits the firmware to the main control management module. After the firmware is received and burned, the program runs to complete the upgrade.

[0082] In the event of an abnormal power outage, since normal power-on is achieved by briefly pressing the button on chassis 100, if chassis 100 experiences an abnormal power outage, such as a server power failure in the server room, the status can be recorded via EPPROM or flash memory upon power restoration. If the power-on status is determined by a short press, all cards can be restored to power-on status; if the power-off status is determined by a long press, the cards will not be restored to power-on status. Furthermore, if the network switching module (terminal access network) chassis 100 management program exits abnormally, the daemon process can restart the chassis 100 management program.

[0083] The above description is merely a preferred embodiment of this application. This application is not limited to the above-described embodiments. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of this disclosure, as long as they achieve the same technical effects, should be included within the scope of protection of this disclosure and should fall under the protection scope of this application. Within the protection scope of this application, the technical solutions and / or implementation methods can have various modifications and variations.

Claims

1. A cloud PC server, characterized in that, The 3U standard rack-mount server includes: computing blades (300) with a chassis (100), a fan module, a power module, a network switching module, and a main control management module responsible for monitoring and managing the operating status of the system hardware; the computing blades (300) are configured with 24 blades, each of which can operate independently; the fan module adopts a front-to-back airflow cooling method; the power module has two units and allows for current sharing and hot backup; the network switching module has two independent and physically isolated networks.

2. The cloud PC server according to claim 1, characterized in that, The chassis (100) is provided with a base plate (200) with a single star structure, two switching boards (400) serving as the service data network and the terminal data network respectively, two fan assemblies (500) for blowing and exhausting respectively, a front panel (110) with a debugging network port and a power switch, and a rear panel (120) with a pluggable power module interface, an optical fiber interface and an Ethernet port; the base plate (200) is provided with 24 blade slots for connecting the computing blades (300) and two switching board (400) slots for connecting the switching boards (400).

3. The cloud PC server according to claim 1, characterized in that, The computing blade (300) is equipped with computing units that support various sizes of M.2 SSD hard drives, various memory sizes and are compatible with various CPUs, encoding units for accessing audio and video and acting as controllers, and interface units for providing external interfaces for power supply, SERDES, I2C and slot I / O.

4. The cloud PC server according to claim 3, characterized in that, The interface unit includes a network submodule and an IPMI submodule; the network submodule contains two independent PHY chips for service transmission and data management respectively; the IPMI submodule reads the temperature status of the computing unit in real time through the UART1 port, and can control the computing unit to power on and off and reset through control IO, and read the power-on, power-off, temperature and power status of the entire board, and communicates with the encoding unit through the UART2 port.

5. The cloud PC server according to claim 4, characterized in that, The computing blade (300) is provided with an input interface submodule, a data processing submodule, and an output interface submodule; the input interface submodule is used for HDMI video data decoding; the data processing submodule is used for strip segmentation, pixel reconstruction, and video buffering. The output interface submodule is used for HDMI video data encoding.

6. The cloud PC server according to claim 1, characterized in that, The main control management module is equipped with a health detection unit, a system information unit, a control unit, and a communication interface unit. The health detection unit connects to the power module via PMBus to monitor the overall power status, collects module temperature via an I2C temperature sensor, has a dedicated watchdog reset chip to improve system reliability, and uses a built-in ADC to monitor the real-time voltage of the 12V power supply. The system information unit stores or reads FRU information via EEPROM and uses a built-in RTC real-time clock in the MCU. The control unit adjusts the fan speed via PWM output, collects and monitors the fan speed in real time, detects whether the computing blade (300) and network switching board (400) are in place, and controls power-on / off.

7. The cloud PC server according to claim 1, characterized in that, The main control management module uses the uCOS II system and has two threads, one for power button detection and the other for parsing and processing serial port messages received from the network exchange module.

8. The cloud PC server according to claim 1, characterized in that, The network switching module is equipped with a service data network for network data transmission between PCs, and a terminal access network for transmitting media streams of screen images and audio data, as well as mouse and keyboard control signals.

9. The cloud PC server according to claim 1, characterized in that, The network switching module is equipped with a downlink network interface unit, a data switching unit, an uplink network interface unit, and an IPMC unit. The downlink network interface unit converts 6 QSGMII interfaces into 24 1000Base-X SerDes interfaces through 6 PHY chips and connects them to 24 slots on the baseboard (200). It also connects the 12V power supply, slot number, and IPMI bus of the baseboard (200) to the network switching module through an ADF connector. The data switching unit is used for Layer 2 switching of network data packets. The uplink network interface unit includes data interface units on the front panel (110) and the rear panel (120). The IPMC unit serves as the management unit of the network switching board (400). It communicates with the main control management module of the cloud PC server through the MCU and monitors the health information of the network switching module, such as temperature and voltage, in real time. It also implements power-on timing control through the FPGA and parses the serial signal of the network port LED control.

10. The cloud PC server according to claim 1, characterized in that, The cloud PC server also includes a cloud PC management platform for online upgrades. The cloud PC management platform sends an upgrade command to the network switching module. The network switching module controls the main control management module to reset and enter bootload mode. Then, the network switching module transmits firmware to the main control management module. After the firmware is received and burned, the program runs to complete the upgrade.

Images