On-chip computing apparatus, computer device, and cloud instance generation method and system

Abstract

The present disclosure discloses an on-chip computing apparatus, a computer device, and a cloud instance generation method and system. The on-chip computing apparatus comprises an interface terminal, at least one conversion chip, and at least one on-chip computing chip. The conversion chip is connected to the interface terminal, and is connected to at least one on-chip computing chip. The interface terminal is configured to be plugged into an interface slot of a processor in a computer device. The conversion chip is used to convert an operation mode of the on-chip computing chip into an operation mode required by the processor, and convert an interface type of the processor into an interface type supported by the on-chip computing chip, so that the computer device can virtualize a cloud instance on the basis of the on-chip computing chip. In the present disclosure, the on-chip computing apparatus can be connected to the computer device, and interacts with the processor of the computer device via the conversion chip, so that the on-chip computing chip can reuse power, hardware resources, network resources, and storage resources of the computer device, thereby providing a hardware foundation for the on-chip computing chip to access a cloud system.

Description

[0001]On-Chip Computing Device, Computer Equipment, Cloud Instance Generation Method, and System. This disclosure claims priority to Chinese patent application number 202410358485.5, filed with the China Patent Office on March 27, 2024, entitled "On-Chip Computing Device, Computer Equipment, Cloud Instance Generation Method, and System," the entire contents of which are incorporated herein by reference. Technical Field: This disclosure relates to the field of cloud technology, and more specifically, to an on-chip computing device, computer equipment, cloud instance generation method, and system. Background: Currently, mobile phones and some IoT devices use on-chip computing chips, which have the advantages of low power consumption, small size, and low cost. With the development of virtualization technology, user demand for virtualized on-chip computing chips has emerged. Related art provides a solution for providing users with cloud instances corresponding to on-chip computing chips. In this solution, cloud instances are virtualized based on the processor, memory, and network resources of a physical server, and the cloud instances are provided to users requesting on-chip computing resources. In the aforementioned related technologies, cloud instances exclusively occupy the computing, storage, and network resources of the physical server. However, the amount of on-chip computing resources required by users is much smaller than the resources available on the physical server. Directly using cloud instances virtualized from physical servers to meet user demands for on-chip computing resources results in wasted resources and high costs. SUMMARY OF THE INVENTION This disclosure addresses the technical issues of wasted resources and high costs associated with directly using cloud instances virtualized from physical servers to meet user demands for on-chip computing resources in the aforementioned related technologies. An on-chip computing device, computer equipment, and a cloud instance generation method and system are provided. A first aspect of the present disclosure provides an on-chip computing device comprising an interface terminal, at least one conversion chip, and at least one on-chip computing chip; the conversion chip is connected to the interface terminal, and the conversion chip is connected to the at least one on-chip computing chip; the interface terminal is configured to be plugged into an interface slot of a processor in a computer device; the conversion chip is configured to convert the operating mode of the on-chip computing chip to the operating mode required by the processor and convert the interface type of the processor to an interface type supported by the on-chip computing chip, enabling the on-chip computing chip to communicate with the processor, so that the computer device can virtualize a cloud instance based on the on-chip computing chip. A second aspect of the present disclosure provides a computer device comprising a processor, a network interface card, and the on-chip computing device described in the first aspect; the processor is respectively connected to the network interface card and the on-chip computing device. A third aspect of the present disclosure provides a cabinet comprising at least one computer device described in the second aspect.A fourth aspect of the present disclosure provides a data center, comprising a master control device and at least one cabinet as described in the third aspect; the master control device is connected to the computer devices in each cabinet. A fifth aspect of the present disclosure provides a cloud instance generation method, applied to the master control device included in the data center as described in the fourth aspect, the method comprising: receiving a resource usage request initiated by a client for an on-chip computing chip, the resource usage request carrying resource specification information required by the client; sending an instance creation request to a target computer device connected to a target on-chip computing chip that meets the resource specification information, so that the target computer device virtualizes the target on-chip computing chip to generate a cloud instance; receiving instance information of the cloud instance sent by the target computer device, and sending the instance information to the client. The sixth aspect of the present disclosure proposes a cloud instance generation method, which is applied to the computer device included in the data center described in the fourth aspect above, and the method includes: receiving an instance creation request sent by the main control device, the instance creation request including resource specification information required by the client and identification information of a target on-chip computing chip that meets the resource specification information; based on the resource specification information and the identification information, virtualizing the target on-chip computing chip in a locally connected on-chip computing device to generate a cloud instance corresponding to the target on-chip computing chip; and sending the instance information of the cloud instance to the main control device. A seventh aspect of the present disclosure proposes a cloud instance generation system, which is applied to the data center described in the fourth aspect, wherein the main control device is used to receive a resource usage request initiated by a client for the on-chip computing chip, and the resource usage request carries resource specification information required by the client; sends an instance creation request to a computer device connected to a target on-chip computing chip that meets the resource specification information; receives instance information of a cloud instance sent by the target computer device, and sends the instance information to the client; the target computer device is used to receive the instance creation request, and the instance creation request carries the resource specification information; determines the target on-chip computing chip that meets the resource specification information from each on-chip computing chip included in the locally connected on-chip computing device; virtualizes the target on-chip computing chip to generate a cloud instance corresponding to the target on-chip computing chip; and sends the instance information of the cloud instance to the main control device.An eighth aspect of the present disclosure provides a cloud instance generation system, applied to a cloud data center. The cloud data center includes a master control device and multiple servers, including a server connected to the on-chip computing device described in the first aspect. The master control device is configured to receive a resource usage request initiated by a client for an on-chip computing chip, the resource usage request carrying resource specification information required by the client. The master control device sends a cloud phone instance creation request to a server connected to a target on-chip computing chip that meets the resource specification information. The master control device receives instance information of the cloud phone instance from the server and sends the instance information to the client. The server receives the cloud phone instance creation request and, based on the resource specification information carried in the cloud phone instance creation request, determines a target on-chip computing chip in the target on-chip computing device connected to the server that meets the resource specification information. The server virtualizes the target on-chip computing chip to generate the cloud phone instance. The server sends the instance information of the cloud phone instance to the master control device. A ninth aspect of the present disclosure provides a computer-readable storage medium storing a computer program, the program being executed by a processor to implement the method described in the fifth or sixth aspect. The tenth aspect of the present disclosure provides a computer program product, including a computer program, which is executed by a processor to implement the method described in the fifth or sixth aspect. Based on the on-chip computing device described in the first aspect, the present disclosure has at least the following beneficial effects or advantages: In the embodiments of the present disclosure, the on-chip computing device can be plugged into a computer device via an interface terminal, similar to a hard drive. The on-chip computing chip in the on-chip computing device can interact with the processor of the computer device via a conversion chip. The on-chip computing chip can be connected to the computer device, enabling it to reuse the computer device's power, hardware resources, network resources, and storage resources. This provides the hardware foundation for the on-chip computing chip to connect to a cloud system. Connecting the on-chip computing chip to a cloud system enables cloud instance services to be provided to users directly based on the on-chip computing chip. Compared to related art solutions that rely on physical servers to provide cloud instance services for on-chip computing resources, the present disclosure can effectively reduce costs. The above description is merely an overview of the technical solutions of the present disclosure. To provide a clearer understanding of the technical solutions of the present disclosure, implementation is possible in accordance with the contents of this specification. To further enhance the understanding of the above and other objectives, features, and advantages of the present disclosure, specific embodiments of the present disclosure are described below. BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings described herein are provided to further enhance understanding of the present disclosure and constitute a part of the present disclosure. The illustrative embodiments of the present disclosure and their descriptions are provided for illustrative purposes only and are not intended to unduly limit the present disclosure.In the accompanying drawings: Figure 1 is a schematic diagram of the hardware structure of a physical server based on a cloud instance as shown in the related art; Figure 2 is a schematic diagram of the structure of a card array system as shown in the related art; Figure 3 is a schematic diagram of the structure of an on-chip computing device according to some exemplary embodiments of the present disclosure; Figure 4 is another schematic diagram of the structure of an on-chip computing device according to some exemplary embodiments of the present disclosure; Figure 5 is another schematic diagram of the structure of an on-chip computing device according to some exemplary embodiments of the present disclosure; Figure 6 is a schematic diagram of the structure of a computer device according to some exemplary embodiments of the present disclosure; Figure 7 is a schematic diagram of the structure of a cabinet according to some exemplary embodiments of the present disclosure; Figure 8 is a schematic diagram of the structure of a data center according to some exemplary embodiments of the present disclosure; Figure 9 is a flowchart of a cloud instance generation method according to some exemplary embodiments of the present disclosure; Figure 10 is a flowchart of another cloud instance generation method according to some exemplary embodiments of the present disclosure; Figure 11 is a schematic diagram of the hardware structure of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS Exemplary embodiments will be described in detail herein, with examples thereof illustrated in the accompanying drawings. In the following description, when referring to the drawings, unless otherwise indicated, identical numerals in different drawings represent identical or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Rather, they are merely examples of apparatuses and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims. The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting. As used in this disclosure and the appended claims, the singular forms "a," "an," "the," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and / or," as used herein, refers to and encompasses any and all possible combinations of one or more of the associated listed items. It should be understood that while the present disclosure may employ the terms "first," "second," and "third," etc., to describe various information, such information should not be limited to these terms. These terms are used solely to distinguish information of the same type from one another. For example, first information could be referred to as second information, and similarly, second information could be referred to as first information, without departing from the scope of this disclosure. Depending on the context, the term "if," as used herein, could be interpreted as meaning "when," "when," or "in response to determining," etc.It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, storage, and display) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties. The collection, use, and processing of the relevant data must comply with the relevant laws, regulations, and standards of the relevant countries and regions, and corresponding operation portals are provided for users to choose to authorize or refuse. The computing chips used in mobile phones and some IoT devices in the field are on-chip computing chips. On-chip computing chips have the characteristics of low power consumption, small size, and low cost. With the development of cloud technology and virtualization technology, cloud instances with computing resources similar to on-chip computing chips have emerged. Users can log in to cloud instances through a client to use computing resources similar to on-chip computing chips. This eliminates the need for users to purchase physical on-chip computing chips, resulting in lower usage costs and no maintenance for on-chip computing chips. Related art provides a solution for providing users with such cloud instances. In this solution, a physical server is virtualized into a cloud instance corresponding to the on-chip computing chip and provided to users. As shown in Figure 1, a physical server includes a core processor (CPU), a graphics processor, and a network interface card (NIC). The core processor typically uses an ARM (Advanced RICS Machines) architecture. Virtualization technology is used to virtualize the core processor, graphics processor, network interface card, and other hardware resources in the physical server to create a cloud instance. However, this cloud instance monopolizes the computing, storage, and network resources of the physical server, resulting in high costs. Furthermore, the resources required by the cloud instance for the on-chip computing chip required by the user are often smaller than those available on the physical server, resulting in significant idle and wasted resources. Related art also provides another solution for providing cloud instances. This solution uses a board array approach, as shown in Figure 2. Multiple on-chip computing chips are organized into a board array. Each on-chip computing chip is interconnected with an Ethernet switch chip, which is also connected to a USB (Universal Serial Bus) hub. The Ethernet switch chip is connected to a baseboard management controller via a PCIE (Peripheral Component Interconnect Express) interface, and the baseboard management controller is connected to the USB hub via a USB interface. Based on the board array shown in Figure 2, the computing resources of the on-chip computing chips are directly provided to users.In this solution, the card array is a self-contained system that cannot be integrated with the existing computing system. This requires additional maintenance of the card array system, resulting in high maintenance costs. Currently, cloud service providers typically provide cloud servers to users based on physical servers through virtualization technology. The first related technology solution allocates cloud server computing resources to users requesting on-chip computing resources, and users monopolize the resources of the physical Ethernet server, resulting in high costs and resource waste. In the second related technology solution, the card array system and the cloud system on which the cloud server relies are two independent systems, requiring the provider to maintain the card array system, resulting in high maintenance costs. The card array system cannot be integrated into the cloud system and cannot reuse cloud management, VPC (Virtual Private Cloud), and EBS (Elastic Block Store) capabilities. Furthermore, in existing cloud systems, the physical servers that power cloud servers require significant headroom for heat dissipation and power supply capabilities, as the maximum power consumption of physical servers differs significantly from their standard power consumption. This requires a significant headroom for network resources, as the maximum network bandwidth used by physical servers differs significantly from their standard bandwidth. Furthermore, after deploying multiple physical servers in a single cabinet in a data center, the remaining power consumption may not be sufficient to accommodate an additional physical server, resulting in a power headroom requirement. This demonstrates that existing cloud systems have a certain amount of resource redundancy, but these redundant resources are not fully utilized, resulting in resource waste. To address the problems existing in the aforementioned related technologies, embodiments of the present disclosure provide an on-chip computing device comprising at least one on-chip computing chip. This on-chip computing device can be plugged into a computer device like a hard drive, providing the hardware foundation for integrating on-chip computing resources into a cloud system. Within the cloud system, virtualization technology is used to virtualize the on-chip computing chip connected to the computer device into a cloud instance, ensuring that the computing resources provided to users match their needs and reducing resource waste. The on-chip computing device integrated into the cloud system can utilize redundant power consumption, network, and power resources within the cloud system, fully utilizing the existing redundant resources, improving resource utilization, and reducing costs. Furthermore, integrating the on-chip computing device into the cloud system allows for the reuse of hardware resources, cloud management resources, and block storage resources within the cloud system. Compared to the card array approach used in related technologies, this eliminates the need for additional system maintenance, requiring only the cloud system itself to be maintained, effectively reducing maintenance costs.The following describes in detail the technical solutions of the present disclosure and how they solve the aforementioned technical problems using specific embodiments. The enumerated specific embodiments may be combined with one another, and identical or similar concepts or processes may not be described in detail in certain embodiments. The embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings. Figure 3 is a schematic structural diagram of an on-chip computing device provided by an embodiment of the present disclosure. As shown in Figure 3, the on-chip computing device includes an interface terminal 1, at least one conversion chip 2, and at least one on-chip computing chip 3. Figure 3 schematically illustrates only one conversion chip 2 and two on-chip computing chips 3. In actual applications, the number of conversion chips and on-chip computing chips 3 can be adjusted based on actual needs. The conversion chip 2 is connected to the interface terminal 1 and is connected to at least one on-chip computing chip 3. The interface terminal 1 is configured to plug into an interface slot of a processor in a computer device. The conversion chip 2 is configured to convert the operating mode of the on-chip computing chip 3 to the operating mode required by the processor and convert the processor's interface type to an interface type supported by the on-chip computing chip 3, enabling communication between the on-chip computing chip 3 and the processor, allowing the computer device to virtualize a cloud instance based on the on-chip computing chip 3. The computer device may be a server, a desktop computer, or a mainframe computer. The processor of the computer device may be the core processor of a server, desktop computer, or other device. The interface terminal 1 is compatible with the interface slot of the processor in the computer device. The interface terminal 1 of the on-chip computing device can be customized based on the specifications and dimensions of the interface terminal 1 required by the processor's interface slot. The interface terminal 1 and power consumption of the on-chip computing device can all adopt the standard specifications of currently commonly used external devices for computer devices. Such external devices may include solid-state drives. For example, the physical form of the on-chip computing device can be designed based on the standard specifications of the U.2 or E3 interfaces commonly used by solid-state drives. This allows the on-chip computing device to be connected to the computer device without having to modify the processor's interface slot. This improves the reuse rate of the processor's existing interface slot, eliminates the need for hardware modification to the computer device, and reduces the cost of connecting the on-chip computing device to the computer device. After the on-chip computing device is inserted into a computer device, the conversion chip 2 is located between the processor of the computer device and the on-chip computing chip 3. The conversion chip 2 is used to convert the operating mode of the on-chip computing chip 3 into an operating mode that meets the requirements of the processor, and to convert the interface type of the processor into an interface type that can be used by the on-chip computing chip 3, thereby connecting the on-chip computing chip 3 to the processor of the computer device.In the embodiment of the present disclosure, the on-chip computing device may include one or more conversion chips 2. When multiple conversion chips 2 are included, each conversion chip 2 is connected to one or more on-chip computing chips 3. OThe conversion chip 2 may include any chip, such as an ASIC (Application Specific Integrated Circuit) chip or an FPGA (Field Programmable Gate Array) chip. The on-chip computing chip 3 may include a computing chip used in mobile phones or a computing chip used in the Internet of Things (IoT). The data of the on-chip computing chip 3 contained in the on-chip computing device may be determined based on the power consumption and size constraints of the entire on-chip computing device. In the above embodiment, the on-chip computing device can be plugged into the processor of the computer device via the interface terminal 1 of the on-chip computing device. The on-chip computing device includes at least one on-chip computing chip 3. After being plugged into the processor, the conversion chip 2 in the on-chip computing device enables communication between the on-chip computing chip 3 and the processor. This implements the connection of the on-chip computing chip 3 to the computer device at the hardware level, allowing the on-chip computing chip 3 to reuse the computer device's power, network resources, storage resources, etc., providing a hardware foundation for connecting the on-chip computing chip 3 to the cloud system. In some embodiments of the present disclosure, the on-chip computing chip 3 is connected to the conversion chip 2 via a first type of interface and a second type of interface, respectively. The conversion chip 2 is connected to the interface terminal 1 via the first type of interface. The first type of interface is used to transmit in-band data. The second type of interface is used to transmit out-band data. The first type of interface may include a common interface used by a computer processor to connect to peripherals such as a hard drive, such as a PCIe interface or an SCS I (Sma II Computer System I interface). The second type of interface may include an interface used by the on-chip computing chip 3 to transmit out-band data, such as a USB interface. The in-band data mentioned above includes service data generated by users using a cloud instance virtualized by the on-chip computing chip 3. The in-band data also includes in-band management data, including traffic statistics, billing, and other related data. The out-of-band data includes low-speed signals such as IIC (Inter-Integrated Circuit) and GPI0 (General Purpose Input / Output), which are used to implement functions such as power on and off, reset, and temperature detection of the on-chip computing chip 3.After the on-chip computing chip 3 is connected to a computer device, for it to function properly, it must undergo management operations such as powering on and off, resetting, and temperature monitoring. These management operations are implemented based on out-of-band data. The interface through which the on-chip computing chip receives or sends out-of-band data is the aforementioned second-type interface. The service data and in-band data generated by the on-chip computing chip 3 when providing services to users are generated by the on-chip computing chip 3. The interface through which the on-chip computing chip 3 receives or sends service data or in-band data is the aforementioned first-type interface. The interface through which the processor of the computer device sends out-of-band data, in-band data, and service data is the aforementioned first-type interface. Therefore, a direct connection between the on-chip computing chip 3 and the processor of the computer device cannot achieve the transmission of out-of-band data. In the disclosed embodiment, the conversion chip 2 is first connected to the processor via the first-type interface. The conversion chip 2 performs an interface conversion function, converting the interface into the first-type interface and the second-type interface. The converted interface is then connected to the on-chip computing chip 3 via the two converted interfaces, thereby achieving the transmission of out-of-band data, in-band data, and service data between the processor and the on-chip computing chip 3. As shown in FIG4 , the first type of interface is exemplified by a PC IE interface, the second type of interface is exemplified by a USB interface, and the conversion chip 2 is exemplified by an ASIC chip or an FPGA chip. FIG4 shows that the ASIC chip or FPGA chip is connected to the interface terminal 1 via the PC IE interface, transmitting OOB (Out of Band) signals and PC IE signals (i.e., in-band data) via the PC IE interface. The ASIC chip or FPGA chip is connected to the on-chip computing chip 3 via the PC IE interface and the USB interface, transmitting in-band data via the PC IE interface and out-of-band data via the USB interface. FIG4 uses two on-chip computing chips 3 as an example. FIG4 intuitively illustrates that the ASIC chip or FPGA chip expands the PCIE interface between the ASIC chip or FPGA chip and the computer device's processor into two PCIE interfaces between the two on-chip computing chips 3. Furthermore, the ASIC chip or FPGA chip converts the PCIE interface between the ASIC chip or FPGA chip and the computer device's processor into two USB interfaces between the two on-chip computing chips 3. Consequently, after the on-chip computing device shown in FIG4 is connected to the computer device, the computer device can implement in-band management, out-of-band management, and service data exchange for the two on-chip computing chips 3. In the disclosed embodiment, the conversion chip 2 expands the number of the first type of interfaces of the computer device, thereby enabling simultaneous access to multiple on-chip computing chips 3.O Furthermore, the conversion chip 2 can convert the computer device's interface into a second type of interface for transmitting out-of-band data. This allows in-band and out-of-band data transmission between the connected on-chip computing chip 3 and the computer device via the first and second types of interfaces. This ensures that the on-chip computing chip 3 can be connected to the computer device and achieves stable data transmission. In the disclosed embodiment, in addition to the aforementioned interface expansion and conversion functions, the conversion chip 2 must also be able to convert the operating mode of the on-chip computing chip 3 to the operating mode required by the computer device's processor. The operating mode characterizes the interconnection between devices. The operating mode of the on-chip computing chip 3 refers to the data transmission mode between the on-chip computing chip 3 and the computer device's processor. These operating modes include RC (Root Complex) mode and EP (Endpoint) mode. Currently, the most common operating mode for computer device processors is EP mode. Therefore, if the on-chip computing chip 3 is operating in EP mode, the conversion chip 2 does not need to convert the operating mode of the on-chip computing chip 3. After connection, the processor can directly transmit data with the on-chip computing chip 3. In actual applications, the on-chip computing chip 3 of EP mode can be directly customized OIn the case where the on-chip computing chip 3 is an EP mode chip, the conversion chip 2 needs to be able to realize the expansion of the number of PC IE interfaces and the conversion function between the PC IE interface and the USB interface, and support the transmission of out-of-band signals. If the working mode of the on-chip computing chip 3 is RC mode, the conversion chip 2 is required to convert the working mode of the on-chip computing chip 3. After the conversion, the on-chip computing chip 3 and the conversion chip 2 are equivalent to an EP mode device, so that data can be transmitted with the processor after access. The conversion of the working mode by the conversion chip 2 can be realized by the program burned in the conversion chip 2. The specific conversion logic can adopt the corresponding method in the relevant technology, which will not be repeated here. In the case where the working mode of the on-chip computing chip 3 is RC mode, the conversion chip 2 needs to have the following functions: (1) converting the working mode of the on-chip computing chip 3 from RC mode to EP mode, so that data can be transmitted between the on-chip computing chip 3 in RC mode and the processor; (2) realizing the expansion of the number of PC IE interfaces and the conversion function between the PC IE interface and the USB interface; (3) supporting the transmission of out-of-band signals. When the operating mode of the on-chip computing chip 3 is RC mode, the conversion chip 2 can be a chip having the three functions described above, such as an ASIC chip or an FPGA chip. In other embodiments, as shown in FIG5 , the conversion chip 2 may include a mode conversion chip 21 and an interface conversion chip 22. The mode conversion chip 21 and the interface conversion chip 22 are respectively connected to the interface terminal 1 and at least one on-chip computing chip 3. The mode conversion chip 21 is used to convert the operating mode of the on-chip computing chip 3 to the operating mode required by the processor; the interface conversion chip 22 is used to convert the processor's first type of interface to a second type of interface. The combination of the mode conversion chip 21 and the interface conversion chip 22 enables the on-chip computing chip 3 to be connected to the processor of a computer device, thereby diversifying the product structure of the on-chip computing device and helping to meet different user needs. In other embodiments of the present disclosure, the on-chip computing device also includes a housing, wherein the conversion chip 2 and the on-chip computing chip 3 are both disposed within the housing; the interface terminal 1 is disposed on the outer surface of the housing. The housing protects the conversion chip 2 and the on-chip computing chip 3. In the embodiment of the present disclosure, the on-chip computing device can be plugged into the computer device through the interface terminal like a hard disk, and the on-chip computing chip in the on-chip computing device can interact with the processor of the computer device through the conversion chip.An on-chip computing chip can be connected to a computer device, allowing it to reuse the computer device's power, hardware resources, network resources, and storage resources, providing a hardware foundation for the on-chip computing chip to connect to a cloud system. Some embodiments of the present disclosure also provide a computer device, as shown in FIG6 . The computer device includes a processor, a network interface card (NIC), and an on-chip computing device provided in any of the above embodiments. The processor is connected to the NIC and the on-chip computing device, respectively. As shown in FIG6 , the interfaces between the processor, the NIC, and the on-chip computing device can all be PCIe interfaces. In some embodiments of the present disclosure, the NIC can be a smart NIC with a processing chip. The processor's instruction set architecture can be, for example, an X86 architecture or an ARM architecture. Multiple applications can run on the on-chip computing chip. When the on-chip computing chip is used to provide user services, the computing resources of the computer device's processor are not occupied. The processor forwards data between the on-chip computing chip and the NIC. Applications on the on-chip computing chip forward data to be sent to other devices to the NIC via the processor. The network card (NIC) includes a processing chip that parses received data, determines its forwarding path, and then sends the data along the determined forwarding path. When the NIC receives data from an external device sent to an on-chip computing chip connected to the computer device to which the NIC belongs, the NIC's processing chip parses the data, determines which on-chip computing chip the data is destined for, and then forwards the data to the corresponding on-chip computing chip via the processor. This approach allows the on-chip computing chip to provide external services without occupying the computing resources of the computer device, thereby not impacting the computer device's existing service processing. Plugging the on-chip computing device into the computer device's processor allows the on-chip computing device to reuse the computer device's power, network resources, storage resources, and other resources, fully utilizing the computer device's resources and reducing resource waste. Furthermore, once the on-chip computing device is connected to the computer device, the on-chip computing device and the computer device can function as a single entity to provide services to users, enabling a more diverse range of services. This also lays the hardware foundation for connecting the on-chip computing chip to a cloud system, thereby providing cloud services to users based on the on-chip computing chip. Some embodiments of the present disclosure further provide a cabinet, as shown in FIG7 , which includes at least one computer device according to the aforementioned embodiments. When the cabinet includes multiple computer devices, all of the computer devices may be plugged into an on-chip computing device, or some of the computer devices may be plugged into an on-chip computing device.Because the cabinet contains computer devices connected to on-chip computing devices, if there is power headroom in the cabinet, the connected on-chip computing devices can utilize this headroom. For example, if the maximum power consumption of an Ethernet computer device is 500W, after installing 10 computers in the cabinet, the remaining power consumption is 50W, which is insufficient to meet the power consumption requirements of a single computer device, and thus this 50W of power is wasted. However, if the maximum power consumption of an on-chip computing device is 25W, connecting two on-chip computing devices to the cabinet can fully utilize the remaining 50W of power. The connected on-chip computing devices can also reuse the cabinet's power and network resources, fully utilizing the cabinet's redundant resources and reducing resource waste. Moreover, the computer devices connected to the on-chip computing devices in the cabinet can function as a whole to provide services to users, offering more diverse services. Such a cabinet also lays a solid hardware foundation for connecting on-chip computing chips to cloud systems and providing cloud services based on on-chip computing chips. Some embodiments of the present disclosure also provide a data center, as shown in FIG8 . The data center includes a master control device and at least one cabinet provided in the above embodiments. The master control device is connected to the computer devices in each cabinet. Among the computer devices in each cabinet, some are connected to an on-chip computing device. The master control device is responsible for managing the cabinets in the entire data center. The cabinets in the data center contain computer devices connected to on-chip computing devices, which effectively utilize the cabinet's power consumption margin. The connected on-chip computing devices can also reuse the cabinet's power and network resources, fully utilizing the cabinet's redundant resources and reducing resource waste in the data center. Furthermore, the computer devices in the cabinets connected to the on-chip computing devices can function as a whole to provide services to users, offering more diverse services. Such a data center integrates on-chip computing chips into a cloud system, enabling cloud services based on the on-chip computing chips. Compared to related art approaches that utilize a card array to provide on-chip computing resource services, the embodiments of the present disclosure only require maintenance of the data center, eliminating the need for additional maintenance of the card array system. This not only enables on-chip computing chips to connect to a cloud system to provide cloud services, but also reduces maintenance costs. Furthermore, providing cloud services based on the data center allows the on-chip computing chips to reuse cloud hardware resources, block storage resources, network resources, and other resources, improving resource utilization efficiency. Some embodiments of the present disclosure also provide a cloud instance generation method, which is applied to a master control device in the data center provided by any of the aforementioned embodiments. Referring to FIG. 9 , the method specifically includes the following operations: 101-103.Step 101: The master control device receives a resource usage request initiated by a client for an on-chip computing chip. This resource usage request carries the resource specification information required by the client. Step 102: The master control device sends an instance creation request to the target computer device connected to the target on-chip computing chip that meets the resource specification information, causing the target computer device to virtualize the target on-chip computing chip and generate a cloud instance. Step 103: The master control device receives the instance information of the cloud instance from the target computer device and sends this instance information to the client. The client can refer to the application on a user's mobile phone or computer device that accesses the cloud instance. The resource specification information includes the specifications of the computing resources, storage resources, and network resources required by the user. For example, the required computing resource specifications may be 4 cores, the required storage resources may be 8GB, and the required network resources may be 100MB. In addition to the resource specification information, the resource usage request may also include user identity information, device information, and other information. User identity information may include user account information, password information, and other information. Device information may include device model information, device identifier information, and device IP (Internet Protocol) address information. In some embodiments of the present disclosure, the master control device maintains the resources of all on-chip computing chips connected to the data center. Specifically, the master control device maintains a preset resource pool, which includes resource objects corresponding to all idle on-chip computing chips in the data center. Each resource object represents a single on-chip computing chip. A user sends a resource request through a client. The master control device receives the request and, based on the resource specification information carried in the request and the preset resource pool, determines whether a target on-chip computing chip that meets the resource specification information currently exists in the data center. If not, an error message is returned to the client. The client receives the error message and displays it to the user. The user can wait for an on-chip computing chip that meets their resource specification requirements to appear in the data center. Alternatively, the user can modify their resource specification requirements and send the resource request again. In other embodiments of the present disclosure, when the master control device determines that there is no on-chip computing chip that meets the user's needs in the data center, the master control device may also send the resource specifications supported by the current data center to the client, so that the user can select the resource specifications that he can use from the resource specifications supported by the current data center. This can shorten the user's waiting time and provide the user with a cloud instance that meets the user's needs as soon as possible.If, based on the client's resource usage request, the master control device determines that a target on-chip computing chip currently exists in the data center that meets the user's resource specification requirements, it sends an instance creation request to the target computer device connected to the target on-chip computing chip. After receiving the instance creation request, the target computer device virtualizes the target on-chip computing chip based on the instance creation request, obtaining a cloud instance that meets the resource specification requirements. The target computer device sends instance information of the cloud instance to the master control device, which forwards it to the client. This instance information includes at least the IP address of the cloud instance. After receiving this instance information, the client can log in to the cloud instance based on it. Cloud instance services are provided to users based on a data center that integrates on-chip computing chips with computer devices. The master control device in the data center determines a target on-chip computing chip that meets the user's resource specification requirements. The target on-chip computing chip is then virtualized by the target computer device connected to the target on-chip computing chip, and the resulting cloud instance is made available to the user. The target on-chip computing chip reuses the target computer device's power, network resources, storage resources, and other resources, improving resource utilization and reducing resource waste. Connecting an on-chip computing chip to a cloud system and providing cloud instance services to users based on the on-chip computing chip not only enables the on-chip computing chip to connect to the cloud system and provide cloud services, but also reduces maintenance costs. Furthermore, the on-chip computing chip can reuse cloud hardware resources, block storage resources, network resources, and other resources, thereby improving resource utilization efficiency. In some embodiments of the present disclosure, a master control device sends an instance creation request to a target computing device. Specifically, the master control device selects a target resource object that meets the resource specification information from a preset resource pool based on the resource specification information; determines, based on the resource deployment information included in the target resource object, the target computing device to which the target on-chip computing chip represented by the target resource object is connected; and sends the instance creation request to the target computing device. The preset resource pool includes resource objects corresponding to idle on-chip computing chips in the data center. The resource objects include resource specifications of the on-chip computing chip represented by the resource object and resource deployment information of the on-chip computing chip. The resource specifications of the on-chip computing chip include specifications of the computing resources, network resources, and storage resources of the on-chip computing chip. The resource deployment information of the on-chip computing chip includes the location information of the target cabinet where the on-chip computing chip is located, the location information of the computer device connected to the on-chip computing chip in the target cabinet, and the location or number information of the on-chip computing chip in the on-chip computing device to which it belongs.Based on the resource specification information carried in the client's resource usage request and the resource specifications of the on-chip computing chips represented by each resource object in the preset resource pool, the master control device selects a target resource object from the preset resource pool that meets the resource specification information. Then, based on the resource deployment information of the target on-chip computing chip represented by the target resource object, the master control device determines the location of the target on-chip computing chip in the data center, namely, the location of the target cabinet to which the target computing chip belongs, the location of the target computer device connected to the target on-chip computing chip within the target cabinet, and information such as the IP address or device identifier of the target computer device, as well as the location or number of the target on-chip computing chip within at least one on-chip computing chip included in the on-chip computing device plugged into the target computer device. After determining information such as the IP address or device identifier of the target computer device based on the resource deployment information of the target on-chip computing chip, the master control device sends an instance creation request to the target computer device. The instance creation request includes the resource specification information and the location or number of the target on-chip computing chip within the on-chip computing device. After receiving the instance creation request, the target computer device exchanges data with the target on-chip computing chip based on the location or number of the target on-chip computing chip in the on-chip computing device carried in the instance creation request. Based on the storage resource specifications included in the resource specification information in the instance creation request, the target computer device allocates block storage space to the target on-chip computing chip and sends the storage address of the allocated block storage space to the target on-chip computing chip. The target computer device also allocates a network port on the local network interface card that meets the network resource specifications in the resource specification to the target on-chip computing chip and sends the port number of the allocated network port to the target on-chip computing chip. The target computer device also uses a local virtualization system to virtualize the computing resources, allocated block storage space, and network resources of the target on-chip computing chip to obtain a cloud instance corresponding to the target on-chip computing chip. The target computer device sends instance information of the cloud instance to the master control device, which forwards the instance information to the client. The master control device maintains idle on-chip computing chips across the entire data center through a preset resource pool. Based on the user's resource specification information, it automatically selects a target on-chip computing chip that meets the user's needs from the numerous on-chip computing chips in the data center. Through interaction between the master control device and the target computer device connected to the target on-chip computing chip, a corresponding cloud instance is virtualized based on the target on-chip computing chip and provided to the user. This embodiment improves the accuracy and efficiency of allocating cloud instances to clients based on the client's resource specification information by centrally managing the on-chip computing chips in the data center through the master control device.In some embodiments of the present disclosure, the master control device may also periodically poll the performance data of each cabinet in the data center to determine whether any cabinets have redundant resources. For cabinets with redundant resources, on-chip computing devices may be deployed in these cabinets, thereby allocating the redundant cabinet resources to the on-chip computing chips in the on-chip computing devices, thereby improving resource utilization. The master control device may perform polling every five hours, once a day, or once a week. Specifically, during a polling cycle, after obtaining performance data from each cabinet in the data center, the master control device selects target cabinets from each cabinet that meet preset resource surplus conditions based on the performance data of each cabinet in the data center; determines computer devices with available interface slots from each computer device based on device information of each computer device in the target cabinet; and transmits deployment instruction information of the on-chip computing device to a racking terminal, so that the racking terminal instructs the target object to insert the interface terminal of the on-chip computing device into the aforementioned available interface slot based on the deployment instruction information. The preset resource surplus conditions include conditions for restricting the presence of redundant resources in a cabinet, such as the difference between the cabinet's reserved power consumption and its occupied power consumption not meeting the power consumption requirements of a computer device, or the cabinet's network port utilization being below a preset threshold. Each computer device deployed in a data center has its device information configured in the master control device. This information includes the computer device's device identification, resource specifications, identification of the cabinet in which it resides, and the device's IP address. After the master control device identifies a target cabinet with redundant resources based on the performance data of each cabinet, it then, based on the device information of each computer device in the target cabinet, selects a computer device with an interface backplane from the multiple computer devices in the target cabinet. The interface backplane includes multiple interface slots for the processor in the computer device. In one implementation, when an external device is connected to an interface slot on the computer device's interface backplane, the computer device may send the device information of the external device and the identification of the interface slot occupied by the external device to the master control device. The master control device stores device information for each external device connected to the computer device and erases identification information for the interfaces occupied by each external device. Thus, after the master control device determines a computer device in the target cabinet that has an interface backplane, in the manner described above, it can query locally stored information related to external devices connected to the computer device and, based on this information, determine whether there are any available interface slots on the interface backplane of the computer device. In another implementation, after the master control device determines a computer device in the target cabinet that has an interface backplane, it can send an interface query request to the computer device, inquiring whether there are any available interface slots on the interface backplane of the computer device.After receiving the interface query request, the computer device determines whether there are any available interface slots based on the connections to external devices in the interface slots on its interface backplane. If so, the master control device returns the identification information of the available interface slots to the master control device. After determining the computer device with the available interface slot in the target cabinet using the aforementioned method, the master control device sends deployment instruction information for the on-chip computing device to the racking terminal. This deployment instruction information includes the location information of the target cabinet in the data center, the location information of the computer device with the available interface slot in the target cabinet, and the identification information of the available interface slot. The identification information of the available interface slot can indicate the location of the available interface slot on the interface backplane of the computer device. The on-chip computing device has a certain maximum power consumption, such as 25W, 40W, or 75W. The redundant resources of the target cabinet selected by the master control device can meet the maximum power consumption requirement of the on-chip computing device. When the master control device sends the deployment instruction information to the racking terminal, it may also include information such as the model or maximum power consumption of the on-chip computing device to be racked. The target object may be a staff member in a data center responsible for maintaining or deploying the on-chip computing device, and the racking terminal may be the staff member's terminal. After receiving the deployment instruction information, the racking terminal displays the deployment instruction information. The staff member then inserts the on-chip computing device into a computer device with a vacant interface slot in the target cabinet based on the racking location indicated by the deployment instruction information. In other embodiments, the target object may be an automated deployment device such as a robotic arm or robot responsible for deploying the on-chip computing device in the data center, and the racking terminal may be a control terminal that communicates with the automated deployment device. After receiving the deployment instruction information, the racking terminal controls the automated deployment device to move to the location of the target cabinet based on the location information of the target cabinet included in the deployment instruction information. The automated deployment device then locates the computer device with a vacant interface slot in the target cabinet based on the location of the computer device with a vacant interface slot in the target cabinet included in the deployment instruction information. Finally, based on the identification information of the vacant interface slot, the vacant interface slot is located on the computer device's interface backplane, and the interface terminal of the on-chip computing device is aligned with the vacant interface slot and inserted. The above embodiment automatically selects target cabinets with redundant resources, automatically identifies computer devices capable of accessing the on-chip computing device and any available interface slots on the computer devices within the target cabinets, and sends deployment instructions to the racking terminal, which then completes the racking of the on-chip computing device. The entire process is fully automated, highly efficient, and low-cost.In some embodiments of the present disclosure, the process of a master control device selecting target cabinets that meet preset resource threshold conditions based on the performance data of each cabinet may include: the master control device, based on the actual power consumption and port utilization included in the performance data of each cabinet in the data center, selecting target cabinets whose port utilization is below a preset threshold and whose power consumption margin is within a preset power consumption range. The cabinet power consumption margin is calculated based on the cabinet's reserved power consumption and actual power consumption. The actual power consumption of a cabinet is the sum of the power consumption of all computer devices installed in the cabinet. Port utilization indicates the degree of utilization of the cabinet's network ports and is an indicator used to measure whether the network ports are being used efficiently. The preset threshold may be 70%, 80%, 85%, etc. The embodiments of the present disclosure do not limit the specific value of the preset threshold; it can be set according to actual application needs. The power consumption margin is the difference between the cabinet's reserved power consumption and its actual power consumption. The preset power consumption range may be greater than or equal to the maximum power consumption of an on-chip computing device. Based on the actual power consumption and port utilization of the cabinet, the master control device can accurately determine whether a cabinet has redundant resources and whether the redundant resources can meet the needs of a single on-chip computing device. Cabinets with redundant resources that can meet the needs of a single on-chip computing device are selected as target cabinets. This allows for the precise identification of target cabinets suitable for hosting an on-chip computing device, improving the accuracy of on-chip computing device deployment. Furthermore, deploying the on-chip computing device in the target cabinet allows the redundant resources of the target cabinet to be used by the on-chip computing chips in the on-chip computing device, improving resource utilization and reducing resource waste in the target cabinet. In some embodiments of the present disclosure, a master control device sends deployment instructions for an on-chip computing device to a mounting terminal. The mounting terminal then completes the insertion of the on-chip computing device into the corresponding computer device. The computer device then detects that the on-chip computing device is connected to a local interface slot and controls the power-on of the on-chip computing device. After power-on, the on-chip computing chip in the on-chip computing device obtains an initialization program for the on-chip computing chip from the network via the computer device's processor and network card, and executes the initialization program to complete the initialization of the on-chip computing device. The computer device then sends resource access information for each on-chip computing chip in the on-chip computing device to the master control device. The master control device receives the resource access information sent by the computer device after the on-chip computing device is connected. The resource access information describes the resource configuration of each on-chip computing chip in the on-chip computing device. Based on the resource access information, the master control device adds resource objects representing each on-chip computing chip to a preset resource pool. The resource access information includes identification information and resource configuration information for each on-chip computing chip. The resource configuration information includes the specifications of the computing resources of the on-chip computing chip.The above-described initialization operation for the on-chip computing chip prepares the operating system of the on-chip computing chip. After initialization is complete, the on-chip computing chip can provide external services and process data. At this point, the computer device sends resource access information for each on-chip computing chip in the connected on-chip computing device to the master control device. Based on the resource access information, the master control device adds the resource objects corresponding to each on-chip computing chip to the preset resource pool, thereby integrating the resources of the on-chip computing chip into the preset resource pool for unified management, so that the resources of the on-chip computing chip can be subsequently provided to users. After the computer device detects the locally connected on-chip computing device, it sends the resource access information of the on-chip computing device to the master control device. Based on this information, the master control device adds the resource objects corresponding to each on-chip computing chip in the on-chip computing device to the preset resource pool. In this way, these on-chip computing chips can be uniformly managed by the master control device. Subsequently, based on the preset resource pool, on-chip computing chips that meet the user's resource specifications can be efficiently and accurately allocated to users as cloud instances. In some embodiments of the present disclosure, after sending the instance information of a cloud instance to a client, when the user no longer needs to use the cloud instance, the client sends a deregistration request for the cloud instance to the master device through the client. This deregistration request carries user information and instance information of the cloud instance. User information may include the user account and the client's IP address. Instance information may include the IP address and identification information of the cloud instance. The master device receives the deregistration request for the cloud instance from the client and sends a deregistration instruction to the target computer device connected to the target on-chip computing chip corresponding to the cloud instance, enabling the target computer device to interact with the target on-chip computing chip to release the resources of the target on-chip computing chip. The master device then receives a deregistration completion notification from the target computer device and adds a resource object representing the target on-chip computing chip to a preset resource pool. In the aforementioned embodiments, after the master device assigns the cloud instance to the client, it may also store a mapping between the instance information of the cloud instance and the identification information of the corresponding target on-chip computing chip. When the master control device receives a deregistration request for a cloud instance, based on the instance information of the cloud instance included in the deregistration request, it queries the identification information of the target on-chip computing chip corresponding to the cloud instance from the pre-stored mapping relationship, and determines the target computer device to which the target computing chip is connected based on the identification information of the target on-chip computing chip and the relevant information of the connected on-chip computing chips in each computer device stored in the master control device.The master control device sends a deregistration instruction message for the cloud instance to the target computer device. The deregistration instruction message is used to instruct the target computer device to deregister the cloud instance corresponding to the target on-chip computing chip. The deregistration instruction message may include instance information of the cloud instance, specifications of the computing resources, storage resources, and network resources occupied by the cloud instance, and identification information of the target on-chip computing chip. The target computer device receives the deregistration instruction message and, based on the identification information of the target on-chip computing chip included in the deregistration instruction message, exchanges data with the target on-chip computing chip to release the computing resources of the target on-chip computing chip occupied by the cloud instance, release the block storage space occupied by the target on-chip computing chip, and release the network port occupied by the target on-chip computing chip. After releasing the resources of the target on-chip computing chip, the target computer device sends a deregistration completion notification to the master control device. The master control device then adds the resource object corresponding to the target on-chip computing chip to the preset resource pool to facilitate the subsequent allocation of the target on-chip computing chip to other users. The above-described method completes the process of providing a cloud instance for a user based on an on-chip computing chip, deregistering the cloud instance, and re-adding the resource objects corresponding to the on-chip computing chip to the preset resource pool. This entire process is a closed-loop process, ensuring a healthier and more orderly management of the on-chip computing chips in the entire data center, and ensuring that the data center can stably provide cloud services based on the on-chip computing chips. In the disclosed embodiments, an on-chip computing device can be connected to computer devices in the data center. Based on the data center's provision of cloud services to users, the data center's master control device determines a target on-chip computing chip that meets the user's resource specifications. The target on-chip computing chip is virtualized by the target computer device connected to the target on-chip computing chip, and the resulting cloud instance is provided to the user. The target on-chip computing chip reuses the target computer device's power, network resources, and storage resources, improving resource utilization and reducing resource waste. This not only enables the on-chip computing chip to connect to the cloud system to provide cloud services, but also reduces maintenance costs. Furthermore, the on-chip computing chip can reuse the cloud's hardware resources, block storage resources, network resources, and other resources, improving resource utilization efficiency. Some embodiments of the present disclosure also provide a cloud instance generation method, which is applied to a computer device in a data center provided in any of the above embodiments. As shown in Figure 10, the method specifically includes the following steps 201-203. Step 201: The computer device receives an instance creation request sent by a master control device. The instance creation request includes resource specification information required by a client and identification information of a target on-chip computing chip that meets the resource specification information.Step 202: Based on the resource specification information and identification information, the computer device virtualizes the target on-chip computing chip in the locally connected on-chip computing device and generates a cloud instance corresponding to the target on-chip computing chip. Step 203: The computer device sends the instance information of the cloud instance to the master control device. In the embodiments of the present disclosure, the operations of the master control device in the data center can be referred to as the operations of the master control device in the cloud instance generation method provided in the above embodiments and will not be further described here. The identification information may include the serial number of the target on-chip computing chip. After the on-chip computing device is connected to the computer device, each on-chip computing chip in the on-chip computing device sends its identification information to the computer device for storage. After receiving the instance creation request from the master control device, the computer device interacts with the target on-chip computing chip based on the identification information of the target on-chip computing chip carried in the instance creation request. Through this interaction, the computer device allocates block storage space with the storage resource specifications included in the resource specification information to the target on-chip computing chip, sends the storage address of the block storage space to the target on-chip computing chip, and allocates a network port of the local network card to the target on-chip computing chip and sends the port number of the network port to the target on-chip computing chip. The virtualization system on the computer device virtualizes the computing resources of the target on-chip computing chip, the allocated block storage space, the network port, and other resources, to obtain a cloud instance corresponding to the target on-chip computing chip. After the on-chip computing device is connected to the computer device, the computer device and the on-chip computing chip in the on-chip computing device can provide external services as a whole. When the master control device in the data center determines that the target on-chip computing chip meets the resource specifications required by the client, the target on-chip computing chip completes the virtualization processing of the target on-chip computing chip through interaction between the target on-chip computing chip and the connected computer device, and the resulting cloud instance is provided to the client for use. This allows on-chip computing chips to be connected to the cloud system, reusing the computer device's power, storage, and network resources, as well as cloud management and block storage resources, thereby improving resource utilization within the cloud data center. Furthermore, by directly providing cloud instances to users based on the on-chip computing chips, the resource specifications of the provided cloud instances can be aligned with user requirements, reducing resource waste. In some embodiments of the present disclosure, when a computer device detects that an on-chip computing device has been connected to an interface slot on a local interface backplane, it controls the connected on-chip computing device to power on; interacts with each on-chip computing chip in the connected on-chip computing device to initialize each on-chip computing chip; and transmits resource access information for each on-chip computing chip to a master control device, causing the master control device to add resource objects corresponding to each on-chip computing chip to a preset resource pool.The aforementioned initialization process requires the computer device to download an initialization program for the on-chip computing chip from the network, transfer the initialization program to the on-chip computing chip, and execute it to complete the initialization. After the computer device detects the locally connected on-chip computing device and completes the initialization of each on-chip computing chip within the on-chip computing device, it automatically sends resource access information for each on-chip computing chip to the master control device. The master control device then places the resource objects corresponding to each on-chip computing chip into a preset resource pool, allowing these on-chip computing chips to be subsequently provided to users as cloud instances. The computer device automatically detects the access of the on-chip computing chip and, through interaction with the master control device, automatically and flexibly manages the resource objects corresponding to the on-chip computing chip. In some embodiments of the present disclosure, after a cloud instance is assigned to a client, if the client no longer uses the cloud instance and initiates a deregistration request for the cloud instance, the computer device receives deregistration information for the cloud instance from the master control device. This deregistration information may include instance information for the cloud instance, specifications of the computing resources, storage resources, and network resources occupied by the cloud instance, and identification information of the target on-chip computing chip corresponding to the cloud instance. Based on the deregistration instruction, the computer device interacts with the target on-chip computing chip corresponding to the cloud instance to release the resources of the target on-chip computing chip. A deregistration completion notification is sent to the master control device, causing the master control device to add a resource object representing the target on-chip computing chip to the preset resource pool. During the cloud instance deregistration process, the computer device interacts with the target on-chip computing chip to release the resources and, after deregistration is complete, notifies the master control device to re-add the resource object corresponding to the target on-chip computing chip to the preset resource pool. This completes a closed-loop process, from providing a cloud instance based on the on-chip computing chip to deregistering the cloud instance and re-adding the resource object corresponding to the on-chip computing chip to the preset resource pool. This ensures a more robust and orderly management of the on-chip computing chips throughout the data center, ensuring the data center can stably provide cloud services based on the on-chip computing chips. In the disclosed embodiments, the on-chip computing chip can run multiple applications. This allows the on-chip computing chip to provide cloud instance services to users without occupying the computing resources of the computer's processor. The processor forwards data between the on-chip computing chip and the network interface card (NIC). Applications on the on-chip computing chip forward data to other devices to the NIC via the processor. The NIC has a processing chip that parses received data, determines a forwarding path for the data, and then sends the data along the determined forwarding path.When a network card receives data sent from an external device to an on-chip computing chip connected to the computer device to which the network card belongs, the network card uses its own processing chip to parse the data, determine which on-chip computing chip the data is intended for, and then forwards the data to the corresponding on-chip computing chip via the processor. This approach allows the on-chip computing chip to provide external services without occupying the computing resources of the computer device, thereby not impacting the computer device's existing service processing. In the disclosed embodiments, a computer device can be connected to an on-chip computing chip to provide cloud services to users. After receiving an instance creation request, the computer device interacts with a target on-chip computing chip that meets the user's resource specifications, completes virtualization of the target on-chip computing chip, and provides the resulting cloud instance to the user. The target on-chip computing chip reuses the computer device's power, network resources, storage resources, and other resources, improving resource utilization and reducing resource waste. This not only enables on-chip computing chips to access a cloud system and provide cloud services, but also reduces maintenance costs and enables on-chip computing chips to reuse cloud hardware resources, block storage resources, network resources, and other resources, thereby improving resource utilization efficiency. Some embodiments of the present disclosure also provide a cloud instance generation system, applicable to the data center provided in any of the aforementioned embodiments, as shown in FIG8 . A master control device is configured to receive a resource usage request initiated by a client for an on-chip computing chip, the resource usage request carrying resource specification information required by the client; send an instance creation request to a computer device connected to a target on-chip computing chip that meets the resource specification information; receive instance information of a cloud instance from the target computer device, and send the instance information to the client; the target computer device is configured to receive the instance creation request carrying resource specification information; determine a target on-chip computing chip that meets the resource specification information from each on-chip computing chip included in a locally connected on-chip computing device; virtualize the target on-chip computing chip to generate a cloud instance corresponding to the target on-chip computing chip; and send the instance information of the cloud instance to the master control device. The operational details of the master control device and target computer device in this embodiment can be referenced to those of the master control device and computer device in the previous embodiments and will not be repeated here. The cloud instance generation system and the cloud instance generation method provided in the embodiments of this disclosure are based on the same inventive concept and have the same beneficial effects as the methods employed, executed, or implemented therein. The solutions provided in each embodiment of this disclosure are applicable to any application scenario requiring cloud services to users based on an on-chip computing chip. For example, cloud phone scenarios and application scenarios providing cloud services such as image processing and artificial intelligence model inference based on an on-chip computing chip can be used.The cloud referred to in various embodiments of the present disclosure may be a public cloud, a private cloud, or a hybrid cloud. Taking a cloud phone scenario as an example, a cloud instance generation system is provided for use in a cloud data center. The cloud data center includes a master control device and multiple servers, including a server connected to an on-chip computing device. The master control device is configured to receive a resource usage request initiated by a client for an on-chip computing chip, the resource usage request carrying the resource specification information required by the client. The master control device sends a cloud phone instance creation request to a server connected to a target on-chip computing chip that meets the resource specification information. The master control device receives instance information of the cloud phone instance from the server and sends the instance information to the client. The server receives the cloud phone instance creation request and, based on the resource specification information carried in the cloud phone instance creation request, determines a target on-chip computing chip connected to the server that meets the resource specification information. The server virtualizes the target on-chip computing chip to generate a cloud phone instance. The server sends the instance information of the cloud phone instance to the master control device. A rack in the cloud data center contains a server connected to an on-chip computing device. The on-chip computing device can effectively utilize the rack's power consumption margin. The on-chip computing device can also reuse the cabinet's power and network resources, fully utilizing the cabinet's redundant resources and reducing resource waste in cloud data centers. Based on this cloud data center, cloud phone instances can be provided to users based on the on-chip computing chip, while also reusing the cloud data center's operation and maintenance resources, reducing maintenance costs. The on-chip computing chip reuses cloud hardware resources, block storage resources, network resources, and other resources, improving resource utilization efficiency. Corresponding to the aforementioned cloud instance generation method embodiment, the present disclosure also provides an embodiment of a cloud instance generation device. This device is configured to execute the operations performed by the master control device in the cloud instance generation method provided in any of the above-mentioned embodiments. The cloud instance generation device includes: a receiving module configured to receive a resource usage request initiated by a client for an on-chip computing chip, the resource usage request carrying resource specification information required by the client; a sending module configured to send an instance creation request to a target computer device connected to a target on-chip computing chip that meets the resource specification information, so that the target computer device virtualizes the target on-chip computing chip to generate a cloud instance; the receiving module is further configured to receive instance information of the cloud instance sent by the target computer device; and the sending module is further configured to send the instance information to the client. The sending module is specifically configured to filter target resource objects that meet the resource specification information from a preset resource pool based on the resource specification information; determine, based on resource deployment information included in the target resource objects, a target computer device connected to the target on-chip computing chip represented by the target resource objects; and send the instance creation request to the target computer device.The apparatus further includes: an on-chip computing device deployment module configured to, based on performance data of each cabinet in the data center, select target cabinets from each cabinet that meet preset resource surplus conditions; determine computer devices with available interface slots from each computer device based on device information of each computer device in the target cabinet; and transmit deployment instruction information of the on-chip computing device to a racking terminal, so that the racking terminal instructs the target object to insert the interface terminal of the on-chip computing device into the available interface slot based on the deployment instruction information. The on-chip computing device deployment module is specifically configured to, based on actual power consumption and port utilization included in the performance data of each cabinet in the data center, select target cabinets from each cabinet whose port utilization is below a preset threshold and whose power consumption margin is within a preset power consumption range, wherein the power consumption margin of the cabinet is calculated based on the reserved power consumption and actual power consumption of the cabinet. The receiving module is further configured to receive resource access information sent by a computer device after accessing the on-chip computing device. The resource access information describes the resource configuration of each on-chip computing chip in the on-chip computing device. The device also includes an adding module configured to add resource objects representing each on-chip computing chip to a preset resource pool based on the resource access information. The receiving module is further configured to receive a cloud instance deregistration request sent by a client. The sending module is further configured to send a deregistration instruction to a target computer device connected to a target on-chip computing chip corresponding to the cloud instance, so that the target computer device interacts with the target on-chip computing chip to release the resources of the target on-chip computing chip. The receiving module is further configured to receive a deregistration completion notification sent by the target computer device and add a resource object representing the target on-chip computing chip to the preset resource pool. The cloud instance generation apparatus provided in the embodiments of the present disclosure and the cloud instance generation method provided in the embodiments of the present disclosure are based on the same inventive concept and have the same beneficial effects as the methods employed, executed, or implemented therein. Corresponding to the aforementioned cloud instance generation method embodiment, the present disclosure also provides another embodiment of the cloud instance generation apparatus. The device is used to execute the operations performed by the computer device in the cloud instance generation method provided in any of the above embodiments. The cloud instance generation device includes: a receiving module, which is used to receive an instance creation request sent by a master control device, where the instance creation request includes resource specification information required by the client and identification information of a target on-chip computing chip that meets the resource specification information; a virtualization module, which is used to virtualize the target on-chip computing chip in a locally connected on-chip computing device based on the resource specification information and identification information, and generate a cloud instance corresponding to the target on-chip computing chip; and a sending module, which is used to send instance information of the cloud instance to the master control device.The apparatus further includes: a detection module for detecting that an on-chip computing device is connected to an interface slot on a local interface backplane and controlling the power-on of the connected on-chip computing device; an initialization module for interacting with each on-chip computing chip in the connected on-chip computing device to initialize each on-chip computing chip; and the aforementioned sending module for transmitting resource access information of each on-chip computing chip to a master control device, so that the master control device adds resource objects corresponding to each on-chip computing chip to a preset resource pool. The apparatus further includes: a cloud instance deregistration module for receiving deregistration instructions for a cloud instance from the master control device; interacting with the target on-chip computing chip corresponding to the cloud instance to release resources of the target on-chip computing chip; and transmitting a deregistration completion notification to the master control device, so that the master control device adds a resource object representing the target on-chip computing chip to the preset resource pool. The cloud instance generation apparatus provided in the embodiments of the present disclosure and the cloud instance generation method provided in the embodiments of the present disclosure are based on the same inventive concept and have the same beneficial effects as the methods employed, executed, or implemented therein. The implementation process of the functions and effects of each module in the above-mentioned apparatus is detailed in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here. Since the apparatus embodiments are generally similar to the method embodiments, the relevant details can be referred to the partial description of the method embodiments. The apparatus embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate, and the components illustrated as modules may or may not be physical modules, that is, they may be located in one place or distributed across multiple network elements. Some or all of these modules can be selected according to actual needs to achieve the objectives of the disclosed solutions. Persons of ordinary skill in the art can understand and implement these without inventive effort. Some embodiments of the present disclosure also provide an electronic device corresponding to the cloud instance generation method provided in the above-mentioned embodiments, for executing the aforementioned cloud instance generation method. Figure 11 is a hardware structure diagram of an electronic device according to an exemplary embodiment. The electronic device includes: a communication interface 601, a processor 602, a memory 603, and a bus 604. The communication interface 601, the processor 602, and the memory 603 communicate with each other via the bus 604. The processor 602 can execute the cloud instance generation method described above by reading and executing machine-executable instructions corresponding to the control logic of the cloud instance generation method in the memory 603. The details of this method are described in the above embodiments and are not further described here. The memory 603 mentioned in the embodiments of the present disclosure can be any electronic, magnetic, optical, or other physical storage device and can contain stored information, such as executable instructions, data, and the like.Specifically, the memory 603 may be RAM (Random Access Memory), flash memory, a storage drive (such as a hard disk drive), any type of storage disk (such as an optical disk, DVD, etc.), or similar storage media, or a combination thereof. The system network element and at least one other network element are connected via at least one communication interface 601 (which may be wired or wireless), and may utilize the Internet, a wide area network, a local area network, a metropolitan area network, etc. The bus 604 may be an ISA bus, a PCI bus, or an EISA bus. Such buses may be classified as address buses, data buses, and control buses. The memory 603 is used to store programs, and the processor 602 executes the programs upon receiving execution instructions. The processor 602 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method may be completed by hardware integrated logic circuits in the processor 602 or by software instructions. The processor 602 described above can be a general-purpose processor, including a network processor (NP), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware controls, etc. It can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of this disclosure. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in conjunction with the embodiments of this disclosure can be directly implemented and executed by a hardware decoding processor, or by a combination of hardware and software modules within the decoding processor. The electronic device provided in the embodiments of this disclosure and the cloud instance generation method provided in the embodiments of this disclosure are based on the same inventive concept and have the same beneficial effects as the methods employed, executed, or implemented therein. The embodiments of this disclosure also provide a computer-readable storage medium corresponding to the cloud instance generation method provided in the aforementioned embodiments, storing a computer program (i.e., a program product). When executed by a processor, the computer program executes the cloud instance generation method provided in any of the aforementioned embodiments.It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory, or other optical or magnetic storage media, and are not further described here. The computer-readable storage medium provided in the above-mentioned embodiments of the present disclosure and the cloud instance generation method provided in the embodiments of the present disclosure are based on the same inventive concept and have the same beneficial effects as the method used, executed, or implemented by the application program stored therein. The embodiments of the present disclosure also provide a computer program product corresponding to the cloud instance generation method provided in the above-mentioned embodiments. The computer program product includes a computer program that is executed by a processor to implement the cloud instance generation method provided in the above-mentioned embodiments. The computer program product provided in the above-mentioned embodiments of the present disclosure and the cloud instance generation method provided in the embodiments of the present disclosure are based on the same inventive concept and have the same beneficial effects as the method used, executed, or implemented by the application program stored therein. Those skilled in the art will readily recognize other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered illustrative only; the true scope and spirit of the present disclosure are indicated by the claims. It should also be noted that the terms "comprise," "include," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, product, or apparatus comprising a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, product, or apparatus. Without further limitation, an element specified by the phrase "comprising a..." does not preclude the presence of additional identical elements in the process, method, product, or apparatus comprising the recited element. The foregoing description is merely a preferred embodiment of the present disclosure and is not intended to limit the present disclosure. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present disclosure are intended to be within the scope of protection of the present disclosure.

Claims

Claims 1. A computing device on a chip, wherein: The computer comprises an interface terminal, at least one conversion chip, and at least one on-chip computing chip; the conversion chip is connected to the interface terminal, and the conversion chip is connected to at least one on-chip computing chip; the interface terminal is used to be plugged into the interface slot of the processor in the computer device; the conversion chip is used to convert the operating mode of the on-chip computing chip to the operating mode required by the processor, and convert the interface type of the processor to an interface type supported by the on-chip computing chip, so that the on-chip computing chip can communicate with the processor, so that the computer device can virtualize a cloud instance based on the on-chip computing chip.

2. The on-chip computing device according to claim 1, wherein: The on-chip computing chip is connected to the conversion chip through a first type of interface and a second type of interface respectively; the conversion chip is connected to the interface terminal through the first type of interface; the first type of interface is used to transmit in-band data; and the second type of interface is used to transmit out-of-band data.

3. The on-chip computing device according to claim 1 or 2, wherein: The conversion chip includes a mode conversion chip and an interface conversion chip; the mode conversion chip and the interface conversion chip are respectively connected to the interface terminal and at least one of the on-chip computing chips; the mode conversion chip is used to convert the operating mode of the on-chip computing chip into the operating mode required by the processor; the interface conversion chip is used to convert the first type of interface of the processor into a second type of interface.

4. The on-chip computing device according to any one of claims 1 to 3, wherein: The on-chip computing device further includes a housing, the conversion chip and the on-chip computing chip are arranged inside the housing; and the interface terminal is arranged on an outer surface of the housing.

5. A computer device, wherein: It comprises a processor, a network card and the on-chip computing device according to any one of claims 1 to 4; the processor is connected to the network card and the on-chip computing device respectively.

6. A cloud instance generation method, wherein: Applied to a master control device included in a data center, the data center also including the computer device described in claim 5, the method comprising: receiving a resource usage request initiated by a client for the on-chip computing chip, the resource usage request carrying resource specification information required by the client; sending an instance creation request to a target computer device connected to a target on-chip computing chip that meets the resource specification information, so that the target computer device virtualizes the target on-chip computing chip to generate a cloud instance; receiving instance information of the cloud instance sent by the target computer device, and sending the instance information to the client.

7. The cloud instance generation method according to claim 6, wherein: The sending of an instance creation request to a target computer device connected to a target on-chip computing chip that meets the resource specification information includes: based on the resource specification information, screening out a target resource object that meets the resource specification information from a preset resource pool; based on resource deployment information included in the target resource object, determining a target computer device connected to the target on-chip computing chip represented by the target resource object; and sending the instance creation request to the target computer device.

8. The cloud instance generation method according to claim 6 or 7, wherein: The method further comprises: Based on the performance data of each of the cabinets in the data center, a target cabinet that meets a preset resource surplus condition is screened out from each of the cabinets; based on the device information of each computer device in the target cabinet, a computer device with a vacant interface slot is determined from the computer devices; and deployment instruction information of the on-chip computing device is sent to a racking terminal, so that the racking terminal instructs a target object to insert the interface terminal of the on-chip computing device into the vacant interface slot based on the deployment instruction information.

9. The cloud instance generation method according to claim 8, wherein: The method of screening out target cabinets that meet preset resource surplus conditions from each cabinet based on the performance data of each cabinet in the data center includes: screening out target cabinets whose port utilization is lower than a preset threshold and whose power consumption margin is within a preset power consumption range from each cabinet based on the actual power consumption and port utilization included in the performance data of each cabinet in the data center, wherein the power consumption margin of the cabinet is calculated based on the reserved power consumption and actual power consumption of the cabinet.

10. The cloud instance generation method according to claim 8 or 9, wherein: After sending the deployment indication information of the on-chip computing device to the rack terminal, it also includes: receiving resource access information sent by the computer device after the on-chip computing device is accessed, the resource access information is used to describe the resource configuration of each on-chip computing chip in the on-chip computing device; based on the resource access information, adding resource objects used to represent each on-chip computing chip in the preset resource pool.

11. The cloud instance generation method according to any one of claims 6 to 10, wherein: After sending the instance information to the client, the method further includes: receiving a deregistration request for the cloud instance sent by the client; sending deregistration indication information to the target computer device to which the target on-chip computing chip corresponding to the cloud instance is connected, so that the target computer device interacts with the target on-chip computing chip to release the resources of the target on-chip computing chip; receiving a deregistration completion notification sent by the target computer device, and adding a resource object used to represent the target on-chip computing chip to a preset resource pool.

12. A cloud instance generation method, wherein: The computer device according to claim 5 is applied to a data center, wherein the data center also includes a master control device, and the method includes: receiving an instance creation request sent by the master control device, the instance creation request including resource specification information required by the client and identification information of a target on-chip computing chip that meets the resource specification information; based on the resource specification information and the identification information, virtualizing the target on-chip computing chip in a locally connected on-chip computing device to generate a cloud instance corresponding to the target on-chip computing chip; and sending the instance information of the cloud instance to the master control device.

13. The cloud instance generation method according to claim 12, wherein: Before receiving the instance creation request sent by the master control device, it also includes: detecting that the on-chip computing device is connected to the interface slot in the local interface backplane, and controlling the connected on-chip computing device to power on; interacting with each on-chip computing chip in the connected on-chip computing device, and initializing each on-chip computing chip; sending resource access information of each on-chip computing chip to the master control device, so that the master control device adds the resource objects corresponding to each on-chip computing chip in the preset resource pool.

14. The cloud instance generation method according to claim 12 or 13, wherein: After sending the instance information of the cloud instance to the master device, the method further includes: receiving deregistration instruction information of the cloud instance sent by the master device; Interact with the target on-chip computing chip corresponding to the cloud instance to release resources of the target on-chip computing chip; and send a deregistration completion notification to the master control device so that the master control device adds a resource object representing the target on-chip computing chip in a preset resource pool.

15. A cloud instance generation system, wherein: Applied to a data center, the data center comprising a main control device and the computer device according to claim 5; the main control device is configured to receive a resource usage request initiated by a client for the on-chip computing chip, the resource usage request carrying resource specification information required by the client; An instance creation request is sent to a computer device connected to a target on-chip computing chip that meets the resource specification information; instance information of a cloud instance is received from the target computer device, and the instance information is sent to the client; the target computer device is configured to receive the instance creation request, wherein the instance creation request carries the resource specification information; the target on-chip computing chip that meets the resource specification information is determined from among the on-chip computing chips included in the locally connected on-chip computing device; the target on-chip computing chip is virtualized to generate a cloud instance corresponding to the target on-chip computing chip; and the instance information of the cloud instance is sent to the main control device.

16. A cloud instance generation system, wherein: Applied to a cloud data center, the cloud data center includes a master control device and multiple servers, the multiple servers including a server connected to the on-chip computing device according to any one of claims 1 to 4; the master control device is used to receive a resource usage request initiated by a client for an on-chip computing chip, the resource usage request carrying resource specification information required by the client; send a cloud phone instance creation request to a server connected to a target on-chip computing chip that meets the resource specification information; receive instance information of the cloud phone instance sent by the server, and send the instance information to the client; the server is used to receive the cloud phone instance creation request, determine, based on the resource specification information carried in the cloud phone instance creation request, a target on-chip computing chip in the target on-chip computing device connected to the server that meets the resource specification information; virtualize the target on-chip computing chip to generate the cloud phone instance; The instance information of the cloud phone instance is sent to the main control device.

17. A computer-readable storage medium having a computer program stored thereon, wherein: The program is executed by a processor to implement the method according to any one of claims 6 to 1 or 12 to 14.

18. A computer program product, comprising a computer program, wherein: The computer program is executed by a processor to implement the method according to any one of claims 6 to 11 or 12 to 14.

Images