Cabinet and virtual server creation method
By introducing access layer switches and resource-providing device connection components in the data center, resource pooling is achieved using Ethernet and CXL protocols, the problem of low resource sharing and data transmission efficiency between servers is solved, and resource utilization and data are improved Processing efficiency reduces operating costs.
Patent Information
- Application Number
- PCT/CN2024/088874
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-10
- Filing Date
- 2024-04-19
- Publication Date
- 2024-11-14
AI Technical Summary
The inefficient resource sharing and data transmission between servers in existing data centers have led to insufficient utilization of computing power, memory and storage resources, increasing operating costs.
By introducing access layer switches and resource-providing device connection components in the cabinet, resource pooling is achieved using Ethernet and CXL protocols, and virtual servers are built to improve resource sharing and data processing efficiency and reduce operational costs.
Resource pooling is realized, resource utilization and data processing efficiency of data centers are improved, comprehensive operation costs are reduced, and data processing capabilities of load balancing are enhanced.
Smart Images

Figure CN2024088874_14112024_PF_FP_ABST
Abstract
Description
How to create a cabinet and virtual server
[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on May 10, 2023, with application number 2023105368324 and invention name “Cabinet and Virtual Server Creation Method”, the entire contents of which are incorporated by reference into this application. Technical Field
[0002] The embodiments of this specification relate to the field of computer technology, and in particular to a cabinet and a method for creating a virtual server. Background Art
[0003] Current data center architectures utilize multiple server models housed in server cabinets, with data transmitted between servers via Ethernet. While network bandwidth continues to increase, the computing power, memory, and storage resources within each server remain independent. Consequently, improvements are still needed in data transmission efficiency, resource sharing, and cache consistency.
[0004] Based on this, a new Internet protocol is provided. Not only can the servers still be interconnected through existing Ethernet connections, but the servers can also share resources within the server cabinet (such as memory sharing, computing power sharing, etc.) through this new Internet protocol.
[0005] However, since each server needs to deploy both an Ethernet card and a transmission card based on the new Internet protocol, the overall operating costs are greatly increased.
[0006] Summary of the Invention
[0007] In view of this, embodiments of this specification provide a cabinet. One or more embodiments of this specification also relate to a virtual server creation method, a computing device, a computer-readable storage medium, and a computer program to address technical deficiencies in related technologies.
[0008] According to the first aspect of the embodiments of this specification, a cabinet is provided, including an access layer switch, a resource provision device connecting component, and a resource provision device set, wherein the resource provision device set includes at least two types of resource provision devices, and each type of resource provision device includes at least two, wherein the access layer switch and the resource provision device connecting component are connected via Ethernet, and at least two resource provision devices of the same type are connected via the resource provision device connecting component to construct a resource pool for realizing resource sharing by at least two resource provision devices of the same type, and the resource provision device connecting component is used to construct a physical server according to the resource pool corresponding to each type of resource provision device, and create multiple virtual servers on the physical server according to current project requirements.
[0009] According to the second aspect of the embodiments of this specification, a virtual server creation method is provided, which is applied to a cabinet, wherein the cabinet includes an access layer switch, a resource provision device connecting component, and a resource provision device set, wherein the resource provision device set includes at least two types of resource provision devices, and each type of resource provision device includes at least two, wherein the access layer switch and the resource provision device connecting component are connected via Ethernet, and at least two resource provision devices of the same type are connected via the resource provision device connecting component to construct a resource pool for realizing resource sharing, wherein the method includes: constructing a physical server according to the resource pools corresponding to the various types of resource provision devices; and creating multiple virtual servers on the physical server according to the current project requirements, wherein the cabinet is the above-mentioned cabinet.
[0010] According to the third aspect of the embodiments of this specification, a computing device is provided, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, which, when executed by the processor, implement the steps of the above-mentioned virtual server creation method.
[0011] According to a fourth aspect of the embodiments of this specification, a computer-readable storage medium is provided, which stores computer-executable instructions, and when the instructions are executed by a processor, the steps of the above-mentioned virtual server creation method are implemented.
[0012] According to a fifth aspect of the embodiments of this specification, a computer program is provided, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned virtual server creation method. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] FIG1 is a schematic diagram of creating a virtual server based on the physical resources of a cabinet according to an embodiment of the present invention;
[0014] FIG2 is a schematic structural diagram of a cabinet provided by one embodiment of this specification;
[0015] FIG3 is a schematic structural diagram of a resource provision device connection component in a cabinet according to an embodiment of the present disclosure;
[0016] FIG4 is a schematic diagram of the location of a persistent log in a cabinet-based data link according to an embodiment of the present specification;
[0017] FIG5 is a schematic diagram of the structure of a cabinet CPU in a cabinet provided by one embodiment of this specification;
[0018] FIG6 is a schematic diagram of a management mechanism for persistent logs in a cabinet and a simplified processing diagram of a flash memory controller according to an embodiment of this specification;
[0019] FIG7 is a flowchart of a method for creating a virtual server provided by one embodiment of this specification;
[0020] FIG8 is a structural block diagram of a computing device provided by one embodiment of this specification. DETAILED DESCRIPTION
[0021] The following description sets forth many specific details to facilitate a thorough understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the scope of this specification. Therefore, this specification is not limited to the specific implementations disclosed below.
[0022] The terms used in one or more embodiments of this specification are for the purpose of describing specific embodiments only and are not intended to limit one or more embodiments of this specification. The singular forms "a," "the," and "the" used in one or more embodiments of this specification and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.
[0023] It should be understood that although the terms first, second, etc. may be used to describe various information in one or more embodiments of this specification, such information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".
[0024] In addition, 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, stored data, displayed data, etc.) involved in one or more embodiments of this specification are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.
[0025] First, the terms involved in one or more embodiments of this specification are explained.
[0026] CCI: Cache Coherence Interface, common communication interface.
[0027] Compute Express Link (CXL) is an open interconnect standard for high-speed connections between CPUs and devices, and between CPUs and DRAM (Dynamic Random Access Memory). CXL technology maintains memory coherence between the CPU memory space and the memory on connected devices.
[0028] ASW: Access Switch, access layer switch, used to connect servers with various applications installed and connect the servers to the network.
[0029] BMC: board management controller, is a dedicated service processor that uses sensors to monitor the status of a computer, network server, or other hardware-driven devices.
[0030] DDR: Double Data Rate, also known as Double Data Rate Synchronous Dynamic Random Access Memory, is a type of memory.
[0031] VM: Virtual Machine, refers to a complete computer system with complete hardware system functions simulated by software and running in a completely isolated environment.
[0032] CPU: Central Processing Unit, as the computing and control core of the computer system, is the final execution unit of information processing and program running.
[0033] PCle: PCI-Express (peripheral component interconnect express) is a high-speed serial computer expansion bus standard.
[0034] ECC: Error Correcting Code is a technology that can implement "error checking and correction". ECC memory is the memory that uses this technology. It is generally used in servers and graphics workstations to improve the stability of computer operation and increase reliability.
[0035] In this specification, a cabinet is provided. One or more embodiments of this specification also relate to a virtual server creation method, a computing device, a computer-readable storage medium, and a computer program, which are described in detail in the following embodiments.
[0036] Referring to FIG. 1 , FIG. 1 shows a specific schematic diagram of creating a virtual server based on physical resources of a cabinet according to an embodiment of the present specification.
[0037] The left side of Figure 1 shows the cabinet 102 (i.e., the server cabinet) and the physical resources within the cabinet 102, and the right side shows the abstract form of the physical resources within the cabinet 102, as well as multiple virtual servers created based on the physical resources within the cabinet 102; that is, the left side of Figure 1 shows the physical deployment form of the cabinet 102, and the right side shows the functional logical abstraction of the cabinet 102.
[0038] In one embodiment, as shown on the left side of FIG1 , cabinet 102 may be equipped with an ASW switch group, a CXL connection group, and multiple resource provisioning devices (such as one or more CPUs, one or more memory sticks, one or more accelerators, and one or more memories, as shown in FIG1 , where one or more memory sticks provide local and remote memory capabilities, the acceleration operators in one or more accelerators provide accelerated computing power, and one or more memories provide high-performance and large-capacity storage capabilities). The ASW switch group within cabinet 102 is connected to the CXL connection group (which includes an Ethernet port) via Ethernet. Other resource provisioning devices are connected via the CXL connection group, communicating using the CXL protocol to share different types of resources.
[0039] As shown on the right side of FIG1 , the embodiment of this specification can abstract the resources provided by multiple resource providing devices in the cabinet 102, which are connected via the CXL connection group and communicate using the CXL protocol, into resource pools corresponding to various types of resource providing devices, such as a CPU pool, an operator pool, a remote memory pool, a storage pool (such as a high-performance storage pool, a large-capacity storage pool), etc., wherein each CPU in the CPU pool contains its own exclusive local memory (for example, the memory close to the CPU can be used as its own exclusive local memory, and other memories as remote memories), and each operator in the operator pool also contains its own exclusive local memory (for example, the memory close to the accelerator can be used as its own exclusive local memory). Memory, other memory as remote memory), the remote memory pool is shared by all CXL devices (all devices connected through a CXL connection group are CXL devices), and the high-performance storage pool and large-capacity storage pool are also shared by all CXL devices, thereby configuring virtual machine resources with the required computing power, capacity, and throughput. In actual applications, high-performance storage has high cost and good storage performance, while large-capacity storage has low cost and relatively low storage performance. The two work together to ensure a balance between storage cost and storage performance, reducing total cost of ownership. At the same time, equipped with corresponding tiered / pooled storage can also greatly meet users' data access requirements and enhance their data processing experience.
[0040] In one application, the abstract resource pool of multiple resource providing devices in cabinet 102 can be compared to a physical server. Then, based on the resource pool of the physical server and the resource configuration required by the user, multiple virtual servers are created for external presentation. This allows subsequent users to use the conventional mode to connect to servers in other clusters or other cabinets without perceiving the difference in physical deployment forms within cabinet 102, thereby achieving load-balanced data processing without having to deploy various hardware devices in the server for resource sharing, greatly saving overall operating costs.
[0041] The embodiments of this specification provide a CXL connection group and cabinet-level server design that integrates Ethernet and the CXL protocol and abstracts and pools resources through virtualization. This design is embedded in existing data centers in a user-imperceptible service model, enabling data processing based on the abstracted computing and storage service capabilities generated by the unified Ethernet and CXL protocol to form a high-throughput connection. The CXL connection group provided in the embodiments of this specification is presented externally as multiple virtual servers connected to an ASW switch group that can meet the user's computing power, capacity, throughput, and other configurations. Internally, it eliminates physical deployment differences through the pooling of normalized CPUs, operators, remote memory, tiered storage, and other resources, achieving load-balanced data processing while significantly reducing hardware costs.
[0042] Referring to FIG. 2 , FIG. 2 shows a schematic structural diagram of a cabinet provided according to an embodiment of this specification.
[0043] In one embodiment, a cabinet includes an access layer switch 202, a resource provision device connection component 204, and a resource provision device collection 206, wherein the resource provision device collection 206 includes at least two types of resource provision devices, and each type of resource provision device includes at least two, wherein the access layer switch 202 is connected to the resource provision device connection component 204 via Ethernet, and at least two resource provision devices of the same type are connected via the resource provision device connection component 204 to construct a resource pool for realizing resource sharing by at least two resource provision devices of the same type, and the resource provision device connection component 204 is used to construct a physical server according to the resource pool corresponding to each type of resource provision device, and create multiple virtual servers on the physical server according to current project requirements.
[0044] Among them, the access layer switch 202 can be understood as the ASW in the above-mentioned embodiment, the resource providing device connection component 204 can be understood as the CXL in the above-mentioned embodiment, and the at least two types of resource providing devices included in the resource providing device set 206 can be understood as the CPU, memory bar, accelerator, memory, etc. in the above-mentioned embodiment, and each type of resource providing device includes at least two, for example, including two or more CPUs, two or more memory bars, two or more accelerators, two or more memories, etc.
[0045] Furthermore, in one or more embodiments of the present specification, in order to improve data processing security, the access layer switch 202 may be provided in multiple forms, for example, two access layer switches 202 may be provided to form an access layer switch group that can back up each other (such as the ASW switch group in the above embodiment); similarly, the resource provision device connection component 204 may also be provided in multiple forms, for example, two resource provision device connection components 204 may be provided to form a resource provision device connection component group that can back up each other (such as the CXL connection group in the above embodiment).
[0046] In one embodiment, the access layer switch 202 and the resource provision device connection component 204 provided with an Ethernet port are connected via Ethernet, that is, the access layer switch 202 and the resource provision device connection component 204 are connected via Ethernet, or the access layer switch group and the resource provision device connection group are connected via Ethernet; at least two resource provision devices of the same type are connected via the resource provision device connection component 204 to construct a resource pool for resource sharing by at least two resource provision devices of the same type, for example, two or more CPUs are communicatively connected via the resource provision device connection component 204 to construct a CPU pool for CPU resource sharing by two or more CPUs.
[0047] In addition, the resource provision device connection component 204 is used to construct a physical server based on the resource pool corresponding to each type of resource provision device, and to construct multiple virtual servers on the physical server according to the configuration required by the user, so as to adapt the current project requirements based on the multiple virtual servers. For example, the resource provision device connection component 204 compares the CPU pool, operator pool, remote memory pool, high-performance storage pool, large-capacity storage pool, etc. of the above-mentioned embodiment to a physical server, and then constructs multiple virtual servers on the physical server according to the configuration required by the user, so as to present the multiple virtual servers to the outside world and perform various data processing operations to adapt to the needs of the current project system. Among them, the project system can be understood as a project system in any scenario, such as a settlement project system in a financial scenario, a logistics project system in a shopping scenario, etc.
[0048] The cabinet provided in the embodiment of this specification connects the resource providing devices of various types in the resource providing device set through the resource providing device connecting component, so that the resource providing devices of various types can share resources through the corresponding type of resource pool, and regard the various types of resource pools in the cabinet as a physical server. Multiple virtual servers are created on the physical server, and then presented to the outside as multiple virtual servers, connecting to servers in other clusters or other cabinets to realize data transmission. Internally, the differences in physical deployment forms in the cabinet are eliminated through the normalized resource pools of various types, which greatly saves the overall operating costs.
[0049] Referring to FIG. 3 , FIG. 3 shows a schematic structural diagram of a resource provision device connection component in a cabinet according to an embodiment of the present disclosure.
[0050] In conjunction with Figure 3, in one or more embodiments of the present specification, the resource providing device connection component 204 includes a resource providing device connection switch 2042 and a cabinet CPU 2044, wherein the resource providing device connection switch 2042 is connected to the cabinet CPU 2044, and at least two resource providing devices of the same type are connected through the resource providing device connection switch 2042 to construct a resource pool for resource sharing by at least two resource providing devices of the same type. The cabinet CPU 2044 is used to obtain the resource pool corresponding to each type of resource providing device according to the resource providing device connection switch 2042, construct a physical server according to the resource pool corresponding to each type of resource providing device, and create multiple virtual servers on the physical server according to current project requirements.
[0051] In one embodiment, the resource provision device connection switch 2042 can be understood as a CXL switch. The CXL switch serves as a data transmission hub and connects to various CXL devices (such as remote memory, high-performance storage (such as SCM / SSD), large-capacity storage (such as HBA / Expander card), CPU, operators, network memory, etc. in Figure 3) via a CCI interface to implement the multi-configuration virtual server in Figure 1. The cabinet CPU 2044 runs a configuration management program. Resource allocation, data flow, operation management, etc. of each CXL device connected to the CXL switch are all implemented in real time by the configuration management program running on the cabinet CPU. The cabinet CPU 2044 is connected to the CXL switch, but does not provide computing power externally as a resource. Instead, it is used to implement subsequent functions such as protocol conversion, persistent logging, virtualization management, and BMC out-of-band monitoring.
[0052] In one or more embodiments of the present specification, the resource providing device connection switch 2042 is communicatively connected to the cabinet CPU 2044 via the CCI interface, and the resource providing device switch 2042 is communicatively connected to each resource providing device in the resource providing device set 206 via the CCI interface, that is, multiple resource providing devices of the same type are interconnected through the CCI interface of the resource providing device switch 2042, and a resource pool is constructed to enable resource sharing by at least two resource providing devices of the same type, such as the CPU pool mentioned above.
[0053] The configuration management program running in the cabinet CPU 2044 can access the resource pools corresponding to various types of resource provision devices based on the connected resource provision devices connected to the switch 2042. Based on these resource pools, it constructs a physical server and creates multiple virtual servers within the physical server based on current project requirements, enabling subsequent processing of current project requirements based on these multiple virtual servers. Specifically, the configuration management program running in the cabinet CPU 2044 organizes the configuration of multiple virtual servers within the cabinet and maps them to actual physical resources. This configuration management program schedules data and configures the corresponding computing power, throughput, and capacity to match the specifications of the abstracted virtual servers.
[0054] The cabinet provided in the embodiments of this specification has a resource provision device connection component 204 including a resource provision device connection switch 2042 and a cabinet CPU 2044. The resource provision device connection switch 2042 can connect to each CXL device via a CCI interface to implement the multi-configuration virtual server shown in Figure 1 for subsequent load balancing data processing. The cabinet CPU 2044 can serve as a management hub to provide a complete service system including virtualization configuration, protocol conversion, traffic scheduling, monitoring and operation and maintenance, achieving flexible mapping and stable scheduling between abstract capabilities and physical components.
[0055] Still referring to FIG3 , the resource providing device connection component 204 further includes a network card 2046 and a network memory 2048. The network card 2046 is connected to the access layer switch 202, the network memory 2048, and the cabinet CPU 2044. The network memory 2048 is connected to the cabinet CPU 2044 and the resource providing device connection switch 2042.
[0056] The network card 2046 is used to receive the data packet sent by the access layer switch 202, perform protocol conversion on the received data packet, obtain the target data after the protocol conversion, and perform flow marking on the target data, determine the flow label of the target data, and store the target data and the flow label of the target data in the network memory 2048; or the network card 2046 is also used to receive the data packet sent by the access layer switch 202 and send the data packet to the cabinet CPU 2044; the cabinet CPU 2044 is also used to perform protocol conversion on the received data packet, obtain the target data after protocol conversion, and perform traffic tagging on the target data, determine the traffic label of the target data, and store the target data and the traffic label of the target data in the network memory 2048; the resource providing device is connected to the switch 2042, and is also used to obtain the target data and the traffic label of the target data from the network memory 2048, and send the target data to the corresponding virtual machine server for data processing according to the traffic label of the target data, wherein the label is a user-defined field added to the configurable part of the CXL data packet to indicate which virtual server the CXL data packet is processed through.
[0057] Among them, the network card 2046 can be understood as an enhanced network card (such as a network card that enhances network signals), and the network memory 2048 can be understood as a storage device on a network device and a server.
[0058] In one or more embodiments of this specification, there are two ways for the resource provision device connection component 204 to perform protocol conversion on the data packets sent by the access layer switch 202 and to perform traffic marking on the target data after the protocol conversion. One way is to implement the protocol conversion and traffic marking in the enhanced network card by hardware offloading. One implementation method is as follows:
[0059] The network card 2046 is used to receive data packets sent by the access layer switch 202, and then perform protocol conversion on the received data packets to obtain the target data after the protocol conversion, and perform traffic tagging on the target data, determine the traffic label of the target data, and store the target data and the traffic label of the target data in the network memory 2048; such as enhancing the network card matching ASW throughput, presenting multiple virtual servers built by multiple IP address matching to the outside world, and realizing the conversion from Ethernet transmission protocol to CXL protocol, that is, the network card 2046 will unpack the data packets received from ASW based on the Ethernet protocol and repack them according to the CXL protocol The target data is obtained by splicing. At the same time, the encapsulated CXL data (i.e., target data) is tagged (i.e., traffic packaging) based on the virtualization settings and stored in the CXL memory device, Network Memory 2048. The subsequent CXL switch can transmit the data to the corresponding CPU, operator, remote memory, storage and other CXL devices based on the label protocol, thereby realizing data transmission from the virtual server network card to the corresponding CPU, operator, remote memory, storage and other CXL devices. The label is a user-defined field added to the configurable part of the CXL data packet to indicate which virtual server the CXL data packet will be processed by.
[0060] Another method is to implement protocol conversion and traffic labeling by running software on the cabinet CPU 2044. One implementation method is as follows: the network card 2046 is further configured to receive data packets sent by the access layer switch 202 and send the data packets to the cabinet CPU 2044. The cabinet CPU 2044 is further configured to perform protocol conversion on the received data packets to obtain target data after protocol conversion, perform traffic labeling on the target data, determine the traffic label of the target data, and store the target data and the traffic label of the target data in the network memory 2048. If the cabinet CPU 2044 responds to the data packet, it uses several of its CPU cores to convert between Ethernet and CXL, and according to the virtualization configuration, after marking the incoming traffic, stores it in the network memory 2048, a CXL memory device. Subsequently, the CXL switch can transmit the target data to the corresponding CXL device, such as the CPU, operator, remote memory, and storage, based on the label protocol, thereby implementing data transmission from the virtual server network card to the corresponding CXL device, such as the CPU, operator, remote memory, and storage.
[0061] After the target data is stored in the network memory 2048, the resource providing device connected to the switch 2042 can obtain the target data from the network memory 2048 and perform specific data processing. One implementation method is as follows:
[0062] The resource provisioning device, connected to switch 2042, is also configured to retrieve target data and its traffic label from network memory and, based on the target data's traffic label, transmit the target data to the corresponding virtual machine server for processing. For example, a CXL switch can determine the target virtual server for processing the target data based on the target data's label and then transmit the target data to the target virtual server's corresponding CXL devices, such as the CPU, operator, remote memory, and storage, enabling data transmission from the virtual server's network interface card to the corresponding CXL devices, such as the CPU, operator, remote memory, and storage.
[0063] The cabinet provided in the embodiment of this specification, its resource providing device connection component 204 also includes a network card 2046 and a network memory 2048. Through the collaboration of the network card 2046 and the network memory 2048, the high-speed interconnection within the cabinet is simplified, and it is seamlessly integrated with the existing data center network between cabinets to adapt to the existing application system, greatly improving the effective implementation of real-time resource capacity scheduling; at the same time, protocol conversion and traffic labeling can be performed through two implementation methods: the network card 2046 and the cabinet CPU 2044, greatly improving the performance of the cabinet.
[0064] Still referring to Figure 3, the resource providing device connection component 204 also includes a log storage device 2050, and the log storage device 2050 is respectively connected to the network memory 2048 and the cabinet CPU 2044, wherein the cabinet CPU 2044 is also used to store the to-be-processed data in the network memory 2048 in the form of a log file to the log storage device 2050 in the event of data processing abnormality.
[0065] Among them, data processing abnormalities can be understood as power failure, file system damage, etc.
[0066] In one embodiment, in the event of a data processing anomaly, the cabinet CPU 2044 can store the pending data in the network memory 2048 in the form of a log file to the log storage device 2050, forming a persistent log. This allows the unprocessed data to be retrieved from the persistent log for continued processing when data processing returns to normal, thus avoiding data loss. The persistent log can be understood as a non-volatile write cache designed at the junction of the ASW and CXL connection groups, shortening the path and achieving low-latency, high-bandwidth performance.
[0067] Referring to FIG. 4 , FIG. 4 shows a schematic diagram of the location of a persistent log in a cabinet-based data link according to an embodiment of this specification.
[0068] FIG4 shows a schematic diagram of a data link when a virtual server is processing data and is constructed based on a resource pool abstracted by a cabinet-based resource provider.
[0069] In one implementation, a VM can be understood as a virtual server constructed by the source resource pool of the source end (a cabinet). The source CXL connection group and the source ASW group constitute external data transmission. The source PSW group is a layer of switch above the ASW, and data can be sent to the terminal ASW group through the source PSW group. Similarly, data is forwarded to the corresponding resource unit within the cabinet through the terminal ASW group, the terminal CXL connection group (including enhanced network card, network memory), etc. of the terminal (another cabinet).
[0070] This data link uses a persistent log in the network memory to significantly shorten the path from source to destination. This means that as soon as data enters the network memory, a successful write is reported back to the source, reducing write latency. Once data enters the network memory, the CXL switch, in conjunction with configuration management and control procedures, transmits the data to the CPU or operator. After processing by the CPU or operator, the data is written to the components of the CXL storage pool for long-term non-volatile storage.
[0071] When a data processing exception occurs, the valid data in the network memory (that is, the data that is not successfully written to the next link) is quickly written to a persistent medium (such as flash memory) in the form of a log to form a persistent log; after the system recovers, the cabinet CPU can read the data from the persistent log and restore it to the network memory. After recovering from the abnormal point, components such as the CXL switch can connect to the endpoint and process data according to the normal template.
[0072] The cabinet provided in the embodiment of this specification includes a log storage device. By setting a persistent log in the log storage device, when data enters the network memory, a write success can be fed back to the source end, shortening the data write path and alleviating the performance requirements during high-throughput reading and writing.
[0073] Still referring to Figure 3, the cabinet also includes a device status monitor 208, which is respectively connected to the cabinet CPU 2044 and the various types of resource providing devices, wherein the device status monitor 208 is used to monitor the operating status of the various types of resource providing devices and send the monitoring results to the cabinet CPU 2044.
[0074] The device status monitor 208 may be understood as BMC out-of-band monitoring.
[0075] In one embodiment, each resource provision device in the cabinet is configured as a device status monitor 208. The device status monitor 208 is used to monitor the operating status of each resource provision device (i.e., each CXL device in the cabinet) and send the monitoring results to the cabinet CPU 2044. In other words, each CXL device in the cabinet is configured with BMC out-of-band monitoring to monitor its operating status in real time to improve online operation and maintenance and ensure cabinet reliability and stability.
[0076] In one or more embodiments of this specification, the CPU cores in the cabinet CPU can be divided into four parts, such as the first part being the first CPU core, the second part being the second CPU core, the third part being the third CPU core, and the fourth part being the fourth CPU core. The first CPU core, the second CPU core, the third CPU core, and the fourth CPU core are respectively used to handle different division of labor operations, thereby achieving highly efficient data processing. If the CPU cores of the cabinet CPU include the first CPU core, the second CPU core, the third CPU core, and the fourth CPU core, an implementation method for each part of the CPU cores is as follows:
[0077] In one or more embodiments of the present specification, the cabinet CPU includes a first CPU core, which is used to connect the switch according to the resource providing device, obtain the resource pool corresponding to the various types of resource providing devices, build a physical server according to the resource pool corresponding to the various types of resource providing devices, and create multiple virtual servers on the physical server according to current project requirements.
[0078] The first CPU core includes one or more CPU cores.
[0079] In one embodiment, the first CPU core of the cabinet CPU is used for the entire virtualization implementation. According to the resource providing device, the switch is connected to obtain the resource pool corresponding to each type of resource providing device. According to the resource pool corresponding to each type of resource providing device, a physical server is built, and multiple virtual servers are created on the physical server according to the current project requirements, so that the current project requirements can be processed according to the multiple virtual servers in the future; that is, the first CPU core presents multiple virtual servers customized based on the user's configuration specifications to the outside, and realizes internal scheduling of each virtual server and scheduling between virtual servers in the same cabinet internally.
[0080] In one or more embodiments of the present specification, the cabinet CPU includes a second CPU core, which is used to respond to the data processing instruction, trigger the network card to perform protocol conversion on the received data packet, obtain the target data after the protocol conversion, and perform traffic tagging on the target data, determine the traffic label of the target data, and store the target data and the traffic label of the target data in the network memory.
[0081] The second CPU core includes one or more CPU cores.
[0082] In one embodiment, the second CPU core of the cabinet CPU is used for conversion between Ethernet and CXL protocols. In response to the data processing instruction of the enhanced network card, the enhanced network card is triggered to perform protocol conversion on the received data packet, obtain the target data after the protocol conversion, and perform traffic labeling on the target data, determine the traffic label of the target data, and store the target data and the traffic label of the target data in the network memory; that is, after the second CPU core triggers the enhanced network card to receive the data packet from the ASW, it unpacks it based on the Ethernet protocol and repackages and splices it according to the CXL protocol to generate the target data. At the same time, based on the virtualization configuration, the encapsulated CXL data (that is, the target data) is labeled and stored in the network memory, which is a CXL memory device; the outgoing traffic can also be distributed to the corresponding Ethernet network port.
[0083] In one or more embodiments of the present specification, the cabinet CPU includes a third CPU core, and the third CPU core is used to store the to-be-processed data in the network memory in the form of a log file to the log storage device in the event of abnormal data processing.
[0084] The third CPU core includes one or more CPU cores.
[0085] In one embodiment, the third CPU core of the cabinet CPU is used to store pending data in the network memory as a log file to a log storage device in the event of a data processing anomaly. Specifically, the third CPU core and the log storage device (i.e., a persistent log module, including a controller, media, and power-off protection) collaborate to form a persistent log of write traffic, thereby shortening the write path and alleviating the real-time performance requirements for throughput reads and writes.
[0086] In one or more embodiments of this specification, the cabinet CPU includes a fourth CPU core, and the device status monitor is further used to monitor the operating status of the various types of resource providing devices and send the monitoring results to the fourth CPU core.
[0087] The fourth CPU core includes one or more CPU cores.
[0088] In one embodiment, the fourth CPU core of the cabinet CPU is configured to promptly process monitoring results of the operating status of various resource-providing devices sent by the device status monitor. Specifically, the fourth CPU core is configured to collaborate with the device status monitor (i.e., BMC out-of-band monitoring) to achieve efficient operation and maintenance, respond to out-of-band monitoring requests, and so on.
[0089] Referring to FIG. 5 , FIG. 5 shows a schematic structural diagram of a cabinet CPU in a cabinet provided according to an embodiment of this specification.
[0090] The CPU core of the cabinet CPU in FIG5 is divided into four parts, such as the first CPU core, the second CPU core, the third CPU core, and the fourth CPU core in the above embodiment. The first CPU core is used to obtain resources from each resource pool based on the CXL switch and present virtual server capabilities based on the resources of each resource pool. In other words, it provides virtual server capabilities based on the resources of each resource pool and multiple virtual servers customized according to user configuration specifications. Internally, it implements internal scheduling of each virtual server and scheduling between virtual servers in the same cabinet through configuration management programs. The second CPU core is used to trigger the enhanced network card to depacketize the data packet based on the Ethernet protocol after determining that the enhanced network card has received the data packet from the ASW, and repack and splice the data packet based on the CXL protocol to generate target data. At the same time, the encapsulated CXL data (i.e., target data) is tagged based on the virtualization configuration and stored in the network memory, a CXL memory device. At the same time, it can also distribute the outgoing traffic to the corresponding Ethernet network port.
[0091] The third CPU core is used to collaborate with the log storage device (i.e., a device including a controller, media, and power-off protection, such as the simplified flash memory controller in Figure 5) to write valid data in the network memory to a persistent log in the event of data processing abnormalities to achieve data persistence, and collaborate with the simplified flash memory controller to implement power-off protection for the data processing process, wherein the simplified flash memory controller includes multiple flash memory particles; the fourth CPU core is used to collaborate with the device status monitor (i.e., BMC out-of-band monitoring) to achieve efficient operation and maintenance, respond to the out-of-band monitoring request, etc.
[0092] In one or more embodiments of this specification, the log storage device includes a flash memory controller, which is connected to the network memory and the cabinet CPU respectively, wherein:
[0093] The cabinet CPU is used to store the pending data in the network memory in the form of a log file to the flash memory controller in the event of data processing abnormality, and to read the pending data from the log file of the flash memory controller and restore the pending data to the network memory when data processing returns to normal.
[0094] Among them, the specific explanation of data processing abnormality can be found in the above embodiment and will not be repeated here; and data recovery can be understood as system recovery; for example, data processing abnormality is understood as power off, and data recovery can be understood as power on.
[0095] In one embodiment, when the cabinet CPU includes a first CPU core, a second CPU core, a third CPU core, and a fourth CPU core, in the embodiment of this specification, the third CPU core is used to store the pending data (i.e., valid data) in the network memory in the form of a log file to the flash memory controller in the event of data processing abnormality, that is, to form a persistent log in the flash memory controller; and when it is determined that the data has returned to normal and the system has recovered, the pending data is read from the persistent log of the flash memory controller, and the pending data is restored to the network memory for subsequent data processing.
[0096] When a data processing anomaly occurs, the cabinet CPU quickly writes all valid data in network memory to persistent media (i.e., the flash controller) in the form of a log. Upon system recovery, the cabinet CPU reads the data from the persistent log file on this persistent media and restores it to network memory. After recovery from the anomaly, components such as the CXL switch pick up where the anomaly left off and continue operating normally, preventing data loss.
[0097] In one or more embodiments of the present specification, the flash memory controller is also connected to the local memory in the resource provision device set, wherein the cabinet CPU is further used to send a data exception trigger instruction to the flash memory controller in the event of data processing exception; the flash memory controller is used to read the local data in the local memory and the data to be processed in the network memory in response to the data exception trigger instruction, and store the local data and the data to be processed after data processing; and the cabinet CPU is further used to send a data recovery trigger instruction to the flash memory controller in the event of data processing returning to normal; the flash memory controller is further used to read the stored local data and the data to be processed after data processing in response to the data recovery trigger instruction to restore the data to be processed in the network memory.
[0098] 6 , which shows a schematic diagram of a management and control mechanism for a persistent log in a cabinet and a simplified processing of a flash memory controller according to an embodiment of this specification.
[0099] As shown in Figure 6 (a), under normal operating conditions, the cabinet CPU performs data flow tagging, path matching, and memory space reclamation during high-throughput interaction between the network memory, the enhanced network, and the CXL switch. Therefore, the cabinet CPU maintains a dirty page bitmap in local memory in real time based on the data received and sent by the network memory. This bitmap indicates the validity of the data at the corresponding memory address (e.g., cache line) using bits. This dirty page bitmap is modified and stored in local memory, and the space containing the dirty page bitmap is also made read-only to the simplified flash controller via PCIe mapping. The dirty page bitmap indicates data that has been written to the CXL switch but not yet written to the corresponding memory or storage component. While the data in the dirty page bitmap appears to have been written to the corresponding virtual server's memory or storage component to the interconnected cluster, it is actually still being restored to the network card memory and requires subsequent operations. The dirty page bitmap distinguishes this data from other data.
[0100] When an exception occurs, the simplified flash controller can immediately read the valid memory data in the dirty page bitmap from the local memory and save it in its own memory. The data space of the network memory also provides read-only permissions to the simplified flash controller through PCIe mapping. At the same time, the simplified flash controller can quickly read the valid data from the network memory through the PCIe connection based on the valid data bitmap (i.e., valid memory data) obtained from the local memory when the exception occurs, aggregate it in its own cache, and after splicing it into large blocks of data (such as 1MB), write it to the flash media in the form of a persistent log in a high-concurrency mode, as shown in Figure 6 (b). The dirty page bitmap is also written to the flash media to form persistent storage.
[0101] As shown in Figure 6(b), during system recovery, the simplified flash controller reads the corresponding flush data (i.e., data written from the network memory to the persistent log) and the dirty page bitmap from the flash media via the flash interface in a fixed write order. It then restores the flush data and dirty page bitmap to the network memory via the PCIe interface, thereby restoring the network memory state and resuming the data. The fixed write order is to read from the memory in the order of the memory address space index and then write it to the persistent log. Reading also follows the same order, so the memory address space index corresponding to the read content is known. The simplified flash controller only performs fixed-mode operations such as sequential writes and reads, eliminating the need to handle random reads and writes, garbage collection, and wear leveling. Furthermore, because the actual time data remains in flash memory is the duration required for fault recovery (e.g., hours), which is much shorter than the effective data retention period in conventional storage (e.g., months), and because there are no other read and write operations, the requirements for data retention, read disturb, write disturb, and data refresh are significantly reduced. Therefore, the simplified flash memory controller needs to provide data error correction protection, realize the PCIe interface and the media interface, and use its own data cache to achieve sequential writing and sequential reading of data, which simplifies the system design.
[0102] In addition, to ensure data accuracy, the simplified flash memory controller will encode the data through the ECC encoder when aggregating and caching the data into the memory. When reading the data, the simplified flash memory controller can also correct errors in reading data from the flash memory based on the ECC decoder it has set. For example, the simplified flash memory controller reads D1-D6 from the flash memory through the flash memory interface, then decodes and corrects it through the ECC decoder, and then restores it to the network memory through the PCIe interface.
[0103] The cabinet provided in the embodiment of this specification has a flash memory controller that is also connected to the local memory in the resource provision device set. Through the connection between the flash memory controller and the network memory and local memory, data persistence and high-throughput storage based on read-only access and simplified flash memory controller design are designed in abnormal situations.
[0104] Referring to Figure 7, Figure 7 shows a flowchart of a virtual server creation method provided according to an embodiment of this specification, wherein the virtual server creation method is applied to a cabinet, the cabinet includes an access layer switch, a resource provision device connection component, and a resource provision device set, the resource provision device set includes at least two types of resource provision devices, and each type of resource provision device includes at least two, wherein the access layer switch and the resource provision device connection component are connected via Ethernet, and at least two resource provision devices of the same type are connected via the resource provision device connection component to construct a resource pool for resource sharing. In one embodiment, the virtual server creation method includes the following steps.
[0105] Step 702: Construct a physical server according to the resource pool corresponding to each type of resource providing device.
[0106] Step 704: Create multiple virtual servers on the physical server according to current project requirements.
[0107] The cabinet is any one of the cabinets described in the above embodiments.
[0108] In one implementation, the virtual server creation method is the same as the resource provision device connection component of the cabinet and the creation method of the virtual server in the above embodiment. For details, please refer to the above implementation details.
[0109] The above is a schematic diagram of a virtual server creation method according to this embodiment. It should be noted that the technical solution of this virtual server creation method is based on the same concept as the technical solution for creating a virtual server using the resource provision device connection component in the cabinet described above. For details not described in detail in the technical solution of the virtual server creation method, please refer to the description of the technical solution for creating a virtual server using the resource provision device connection component in the cabinet described above.
[0110] One embodiment of the present specification implements a cabinet, including an access layer switch, a resource provision device connecting component, and a resource provision device set, wherein the resource provision device set includes at least two types of resource provision devices, and each type of resource provision device includes at least two, wherein the access layer switch and the resource provision device connecting component are connected via Ethernet, and at least two resource provision devices of the same type are connected via the resource provision device connecting component to construct a resource pool for realizing resource sharing by at least two resource provision devices of the same type, and the resource provision device connecting component is used to construct a physical server according to the resource pool corresponding to each type of resource provision device, and create multiple virtual servers on the physical server according to current project requirements. In one embodiment, the cabinet connects the resource providing devices of various types in the resource providing device set through the resource providing device connection component, so that the resource providing devices of various types can share resources through the corresponding type of resource pool, and the various types of resource pools in the cabinet are regarded as a physical server. Multiple virtual servers are created on the physical server, and then presented to the outside as multiple virtual servers, connecting to servers in other clusters or other cabinets to realize data transmission. Internally, the normalized resource pools of various types eliminate the differences in physical deployment forms in the cabinet, which greatly saves the overall operating costs.
[0111] Figure 8 shows a block diagram of a computing device 800 according to one embodiment of this specification. Components of the computing device 800 include, but are not limited to, a memory 810 and a processor 820. The processor 820 is connected to the memory 810 via a bus 830, and a database 850 is used to store data.
[0112] The computing device 800 also includes an access device 840 that enables the computing device 800 to communicate via one or more networks 860. Examples of such networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. The access device 840 may include one or more of any type of network interface (e.g., a network interface controller (NIC)) whether wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a Worldwide Interoperability for Microwave Access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, a near field communication (NFC) interface, and the like.
[0113] In one embodiment of the present specification, the aforementioned components of computing device 800 and other components not shown in FIG8 may also be connected to each other, for example, via a bus. It should be understood that the computing device structure block diagram shown in FIG8 is for illustrative purposes only and does not limit the scope of this specification. Those skilled in the art may add or replace other components as needed.
[0114] Computing device 800 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, personal digital assistant, laptop computer, notebook computer, netbook computer, etc.), a mobile phone (e.g., a smartphone), a wearable computing device (e.g., a smartwatch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or personal computer (PC). Computing device 800 may also be a mobile or stationary server.
[0115] The processor 820 is configured to execute the following computer-executable instructions, which, when executed by the processor, implement the steps of the above-mentioned virtual server creation method.
[0116] The above is a schematic diagram of a computing device according to this embodiment. It should be noted that the technical solution of the computing device and the technical solution of the above-mentioned virtual server creation method are based on the same concept. For details not described in detail in the technical solution of the computing device, please refer to the description of the technical solution of the above-mentioned virtual server creation method.
[0117] An embodiment of the present specification further provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the above-mentioned virtual server creation method.
[0118] The above is a schematic diagram of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of this storage medium and the technical solution of the virtual server creation method described above are based on the same concept. For details not described in detail in the technical solution of the storage medium, please refer to the description of the technical solution of the virtual server creation method described above.
[0119] An embodiment of the present specification further provides a computer program, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned virtual server creation method.
[0120] The above is a schematic solution of a computer program of this embodiment. It should be noted that the technical solution of this computer program and the technical solution of the above-mentioned virtual server creation method are based on the same concept. For details not described in detail in the technical solution of the computer program, please refer to the description of the technical solution of the above-mentioned virtual server creation method.
[0121] The foregoing description of this specification describes specific embodiments. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different from that described in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order shown or the sequential order to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0122] The computer instructions include computer program code, which may be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal, and software distribution medium. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of patent practice. For example, in some regions, according to patent practice, computer-readable media does not include electric carrier signals and telecommunication signals.
[0123] It should be noted that for the aforementioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the embodiments of this specification are not limited by the order of the actions described, because according to the embodiments of this specification, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the embodiments of this specification.
[0124] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.
[0125] The preferred embodiments disclosed above are intended only to help illustrate this specification. The optional embodiments do not exhaustively describe all details, nor do they limit this application to the specific embodiments described. Obviously, many modifications and variations can be made based on the content of the embodiments of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the embodiments of this specification, so that those skilled in the art can better understand and utilize this specification. This specification is limited only by the claims and their full scope and equivalents.
Claims
1. A cabinet comprising an access layer switch, a resource providing device connecting component, and a resource providing device set, wherein the resource providing device set includes at least two types of resource providing devices, and each type of resource providing device includes at least two, in, The access layer switch is connected to the resource providing device connection component via Ethernet, and at least two resource providing devices of the same type are connected via the resource providing device connection component to construct a resource pool for resource sharing by at least two resource providing devices of the same type. The resource providing device connection component is used to build a physical server according to the resource pools corresponding to the various types of resource providing devices, and to create multiple virtual servers on the physical server according to current project requirements.
2. The cabinet according to claim 1, wherein the resource providing device connection component comprises a resource providing device connection switch and a cabinet CPU, in, The resource providing device connection switch is connected to the cabinet CPU, and the at least two resource providing devices of the same type are connected through the resource providing device connection switch to build a resource pool for resource sharing by the at least two resource providing devices of the same type. The cabinet CPU is used to obtain the resource pool corresponding to each type of resource providing device according to the switch connected to the resource providing device, A physical server is constructed according to the resource pools corresponding to the various types of resource providing devices, and a plurality of virtual servers are created on the physical server according to current project requirements.
3. The cabinet according to claim 2, wherein the cabinet CPU comprises a first CPU core, The first CPU core is used to obtain the resource pool corresponding to each type of resource providing device according to the switch connected to the resource providing device, A physical server is constructed according to the resource pools corresponding to the various types of resource providing devices, and a plurality of virtual servers are created on the physical server according to current project requirements.
4. The cabinet according to claim 2, wherein the resource providing device connection component further comprises a network card and a network memory, wherein the network card is connected to the access layer switch, the network memory, and the cabinet CPU, and the network memory is connected to the cabinet CPU and the resource providing device connection switch, in, The network card is used to receive the data packet sent by the access layer switch, perform protocol conversion on the received data packet, obtain target data after protocol conversion, perform flow marking on the target data, determine the flow label of the target data, and store the target data and the flow label of the target data in the network memory; or The network card is further used to receive a data packet sent by the access layer switch and send the data packet to the cabinet CPU; The cabinet CPU is further used to perform protocol conversion on the received data packet, obtain target data after protocol conversion, and perform flow tagging on the target data, determine the flow label of the target data, and store the target data and the flow label of the target data in the network memory; The resource providing device is connected to the switch, and is also used to obtain the target data and the flow label of the target data from the network memory, and send the target data to the corresponding virtual machine server for data processing according to the flow label of the target data.
5. The cabinet according to claim 4, wherein the cabinet CPU comprises a second CPU core, The second CPU core is used to trigger the network card to process the received data in response to the data processing instruction. The packet is converted into a protocol to obtain target data after the protocol conversion, and the target data is flow-labeled to determine the flow label of the target data, and the target data and the flow label of the target data are stored in the network memory.
6. The cabinet according to claim 2, wherein the resource providing device connection component further comprises a log storage device, wherein the log storage device is connected to the network memory and the cabinet CPU respectively. in, The cabinet CPU is further configured to store the to-be-processed data in the network memory in the form of a log file to the log storage device in the event of abnormal data processing.
7. The cabinet according to claim 6, wherein the cabinet CPU comprises a third CPU core, The third CPU core is used to store the to-be-processed data in the network memory in the form of a log file to the log storage device in the event of abnormal data processing.
8. The cabinet according to claim 6 or 7, wherein the log storage device comprises a flash memory controller, and the flash memory controller is connected to the network memory and the cabinet CPU respectively. in, The cabinet CPU is used to store the to-be-processed data in the network memory in the form of a log file to the flash memory controller in the event of abnormal data processing, and When the data processing returns to normal, the data to be processed is read from the log file of the flash memory controller, and the data to be processed is restored to the network memory.
9. The cabinet according to claim 8, wherein the flash memory controller is further connected to a local memory in the resource providing device set. in, The cabinet CPU is further configured to send a data abnormality trigger instruction to the flash memory controller in the event of data processing abnormality; The flash memory controller is used to read the local data in the local memory and the data to be processed in the network memory in response to the data exception trigger instruction, and store the local data and the data to be processed after data processing; as well as The cabinet CPU is further configured to send a data recovery trigger instruction to the flash memory controller when data processing returns to normal; The flash memory controller is further used to respond to the data recovery trigger instruction to read the stored local data after data processing and the data to be processed to restore the data to be processed in the network memory.
10. The cabinet according to claim 2, further comprising a device status monitor, wherein the device status monitor is connected to the cabinet CPU and each type of resource providing device respectively. in, The device status monitor is used to monitor the operating status of the various types of resource providing devices and send the monitoring result to the cabinet CPU.
11. The rack according to claim 10, wherein the rack CPU comprises a fourth CPU core, The device status monitor is further used to monitor the operating status of the various types of resource providing devices and send the monitoring results to the fourth CPU core.
12. A virtual server creation method, applied to a cabinet, the cabinet comprising an access layer switch, a resource provision device connection component, and a resource provision device set, the resource provision device set comprising at least two types of resource provision devices, and each type of resource provision device comprises at least two, in, The access layer switch is connected to the resource providing device connection component via Ethernet, at least two resource providing devices of the same type are connected via the resource providing device connection component to build a resource pool for resource sharing, and the method includes: Building a physical server according to the resource pools corresponding to the resource providing devices of each type; Create multiple virtual servers on the physical server according to the current project requirements. Wherein, the cabinet is any one of the cabinets in claims 1-11.
13. A computing device, include: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the virtual server creation method according to claim 12 are implemented.
14. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the virtual server creation method according to claim 12.
Citation Information
Patent Citations
System, method and device for managing memory resources
CN113553170A
Systems and methods for scalable and coherent memory devices
CN113742256A
System with cache-coherent memory and server-linking switch
CN113746762A
Data resource scheduling method and system, electronic equipment and storage medium
CN115543617A
Resource sharing device, resource management device, and resource management method
CN115586964A
