Write node determining method, data writing method, distributed system, metadata node in distributed system, and client

Abstract

Embodiments of the present application provide a write node determining method, a data writing method, a distributed system, a metadata node in the distributed system, and a client. The write node determining method is applied to a metadata node in a distributed system, and comprises: receiving a data write request sent by a client, wherein the data write request carries access frequency information; on the basis of the access frequency information and working parameters of at least two working nodes in the distributed system, determining a target write node from among the at least two working nodes; and feeding back the target write node to the client, wherein the working parameters of the target write node are adapted to the access frequency information, and the target write node is used for the client to write data to be written. A metadata node can sense the access frequency of data to be written, and on the basis of the access frequency, determine a working node having adapted working parameters as a target write node; thus, the performance of the working node is fully utilized, and the service quality of the distributed system is improved.

Description

[0001]This disclosure claims priority to Chinese Patent Application No. 202410508959.X, filed with the Chinese Patent Office on April 25, 2024, entitled "Write Node Determination Method, Data Writing Method, Distributed System and Metadata Node in Distributed System, Client," the entire contents of which are incorporated herein by reference. Technical Field: Embodiments of this disclosure relate to the field of computer technology, and particularly to a write node determination method, a data writing method, a distributed system, a metadata node in a distributed system, a client in a distributed system, a computing device, a computer-readable storage medium, and a computer program product. Background Art: With the rapid development of computer technology, the demand for data storage has exploded, and the application of distributed systems is becoming increasingly widespread. Distributed systems are often composed of metadata nodes and data nodes. Metadata nodes mainly record metadata information in the distributed system, while user data in the distributed system is stored in data nodes in the form of files or data blocks. These data nodes are also called worker nodes. Currently, storing data blocks on worker nodes often involves evenly distributing all data blocks across worker nodes in a distributed system. This storage scheme fails to fully utilize the performance of each worker node. Therefore, there is an urgent need for a data writing scheme that can fully utilize the performance of each worker node. In view of this, the present disclosure provides a write node determination method. One or more embodiments of the present disclosure also relate to a data writing method, a distributed system, a metadata node in a distributed system, a client in a distributed system, a computing device, a computer-readable storage medium, and a computer program product, to address the technical deficiencies in the prior art. According to a first aspect of the present disclosure, a write node determination method is provided, applied to a metadata node in a distributed system. The method includes: receiving a data write request sent by a client, wherein the data write request carries access frequency information; determining a target write node from the at least two worker nodes based on the access frequency information and the working parameters of at least two worker nodes in the distributed system; and feeding back the target write node to the client, wherein the working parameters of the target write node are adapted to the access frequency information for the client to write the data to be written.According to a second aspect of the present disclosure, a data writing method is provided, applied to a client in a distributed system, comprising: receiving a data creation instruction, wherein the data creation instruction carries data to be written and access frequency information; sending a data writing request carrying the access frequency information to a metadata node in the distributed system; receiving a target writing node returned by the metadata node, and writing the data to be written to the target writing node, wherein the target writing node is determined by the metadata node based on the access frequency information and the working parameters of at least two working nodes in the distributed system, and the working parameters of the target writing node are adapted to the access frequency information. According to a third aspect of the present disclosure, a distributed system is provided, including a client, a metadata node, and at least two worker nodes; the client is configured to receive a data creation instruction, wherein the data creation instruction carries data to be written and access frequency information; and to send a data write request carrying the access frequency information to the metadata node; the metadata node is configured to determine a target write node from the at least two worker nodes based on the access frequency information and the working parameters of the at least two worker nodes, and to feed back the target write node to the client, wherein the working parameters of the target write node are adapted to the access frequency information; the client is further configured to write the data to be written to the target write node. According to a fourth aspect of the present disclosure, a metadata node in a distributed system is provided, comprising: a first receiving module configured to receive a data write request sent by a client, wherein the data write request carries access frequency information; a determining module configured to determine a target write node from at least two working nodes based on the access frequency information and working parameters of at least two working nodes in the distributed system; and a feedback module configured to feed back the target write node to the client, wherein the working parameters of the target write node are adapted to the access frequency information for the client to write data to be written.According to a fifth aspect of the present disclosure, a client in a distributed system is provided, comprising: a second receiving module configured to receive a data creation instruction, wherein the data creation instruction carries data to be written and access frequency information; a sending module configured to send a data write request carrying the access frequency information to a metadata node in the distributed system; and a writing module configured to receive a target writing node returned by the metadata node and write the data to be written to the target writing node, wherein the target writing node is determined by the metadata node based on the access frequency information and working parameters of at least two working nodes in the distributed system, and the working parameters of the target writing node are adapted to the access frequency information. According to a sixth aspect of the present disclosure, a computing device is provided, comprising: a memory and a processor; the memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions, wherein the computer programs / instructions, when executed by the processor, implement the steps of the above-described writing node determination method or data writing method. According to a seventh aspect of the present disclosure, a computer-readable storage medium is provided, which stores a computer program / instructions that, when executed by a processor, implement the steps of the above-described write node determination method or data writing method. According to an eighth aspect of the present disclosure, a computer program product is provided, including a computer program / instructions that, when executed by a processor, implement the steps of the above-described write node determination method or data writing method. One embodiment of the present disclosure provides a write node determination method applied to a metadata node in a distributed system, which receives a data write request sent by a client, wherein the data write request carries access frequency information; determines a target write node from the at least two working nodes based on the access frequency information and the working parameters of at least two working nodes in the distributed system; and feeds back the target write node to the client, wherein the working parameters of the target write node are adapted to the access frequency information for the client to write data to be written. One embodiment of this disclosure implements a method whereby, when a data write request is initiated, access frequency information of the data to be written is specified. The metadata node determines a target write node whose working parameters match the access frequency information from at least two working nodes in the distributed system based on the access frequency information and the working parameters of at least two working nodes. The determined target write node is then fed back to the client, enabling the client to write the data to be written to the target write node.In this way, the metadata node can sense the access frequency of the data to be written and determine the working node with suitable working parameters as the target writing node based on the access frequency, making full use of the performance of the working node and improving the service quality of the distributed system. Figure 1 is a schematic diagram of an application scenario of a writing node determination method provided in an embodiment of this disclosure; Figure 2 is a flowchart of a writing node determination method provided in an embodiment of this disclosure; Figure 3 is a flowchart of a data writing method provided in an embodiment of this disclosure; Figure 4 is a structural schematic diagram of a distributed system provided in an embodiment of this disclosure; Figure 5 is a flowchart of the processing procedure of a data writing method provided in an embodiment of this disclosure; Figure 6 is a structural schematic diagram of a metadata node in a distributed system provided in an embodiment of this disclosure; Figure 7 is a structural schematic diagram of a client in a distributed system provided in an embodiment of this disclosure; Figure 8 is a structural block diagram of a computing device provided in an embodiment of this disclosure. Detailed embodiments are set forth in the following description to provide a thorough understanding of this disclosure. However, this disclosure can be implemented in many other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this disclosure. Therefore, this disclosure is not limited to the specific embodiments disclosed below. The terminology used in one or more embodiments of this disclosure is for the purpose of describing particular embodiments only and is not intended to limit the scope of one or more embodiments of this disclosure. The singular forms “a” and “the” as used in one or more embodiments of this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in one or more embodiments of this disclosure refers to and includes any or all possible combinations of one or more associated listed items. 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 disclosure, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, first may also be referred to as second, and similarly, second may also be referred to as first, without departing from the scope of one or more embodiments of this disclosure. Depending on the context, the word “if” as used herein may be interpreted as “when…” or “in response to a determination”.Furthermore, 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 disclosure are all information and data authorized by the user or fully authorized by all parties. The collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse. First, the terms and concepts involved in one or more embodiments of this disclosure are explained. Data Block (Chunk): A distributed system divides files stored on it into a series of data blocks for storage. Each data block is called a Chunk. Replica: A distributed system replicates a Chunk to generate identical copies of the data, which exist on different data nodes. This replica is called a Replica. Metadata Node: The node in a distributed system where the metadata of a file is stored. It is typically used to store file status information and Chunk location information, etc. Data Node: The node in a distributed system where data blocks are stored. It is typically responsible for writing, storing, reading, and deleting data block replicas, and is also called a worker node. Machine Model: The model of the machine in the distributed system. Machines of the same model have the same hardware configuration, such as disk size, CPU (Central Processing Unit) model, and memory size. Distributed System: Includes metadata nodes and data nodes. User data in a distributed system is stored in data nodes as files or data blocks. Additionally, data blocks (chunks) can be replicated to create identical Replicas and stored on different data nodes. It should be noted that current distributed systems often consist of metadata nodes and data nodes. One possible implementation distributes all user data blocks evenly across all worker nodes, similar to HDFS (Hadoop Distributed File System). Because worker nodes in a distributed system often cannot guarantee consistent performance, meaning different worker nodes may have different machine models—for example, differences in disk size and CPU performance—they may vary. Taking disk size as an example, different machine types A and B have the same number of disks on their working nodes, but the HDD (hard disk drive) on the working node of type A is larger than that of type B.Since HDD disk read performance is independent of disk size, the read QPS (Query Per Second) remains consistent. However, because high-capacity Type A worker nodes have larger disks, they contain more user data, causing an overall QPS skew. Type A worker nodes receive a higher QPS, impacting the overall service capacity of the distributed system. With all user data blocks evenly distributed across all worker nodes, the distributed system cannot fully utilize the processing power and capacity of each worker node when multiple worker nodes have inconsistent processing capabilities or storage space. Since users perform read and write operations on data stored in the distributed system, but the frequency of access varies for different types of data, one embodiment of this disclosure provides a data placement strategy for heterogeneous scenarios with multiple worker node models in a distributed system. Based on the frequency of user data access, user data is divided into hot and cold data. Hot data refers to data that users frequently read and write, such as personal data and documents; cold data refers to data that users rarely read or write after writing, such as backup and archive data. By sensing the frequency of hot and cold access to user data, the working parameters of worker nodes are considered when writing user data into a distributed system to ensure that the data written to the worker nodes is adapted to their working parameters, thus fully utilizing the performance of each worker node. This disclosure provides a method for determining write nodes, and also relates to a data writing method, a distributed system, a metadata node in the distributed system, a client in the distributed system, a computing device, a computer-readable storage medium, and a computer program product, which are described in detail in the following embodiments. Referring to Figure 1, Figure 1 illustrates an application scenario of a write node determination method according to an embodiment of this disclosure. Figure 1 includes a client 102, a metadata node 104, and at least two worker nodes 106. The worker nodes can be terminal devices, storage nodes, disks, etc. In specific implementation, as shown in Figure 1, the user issues a data creation command through the client 102, which carries the data to be written and access frequency information; the client 102 sends a data write request carrying the access frequency information to the metadata node 104.Metadata node 104 determines the target write node 1062 from at least two working nodes 106 based on access frequency information and working parameters of at least two working nodes 106, and feeds back the target write node 1062 to client 102. The working parameters of the target write node 1062 are adapted to the access frequency information. Client 102 writes the data to be written to the target write node 1062. Client 102 may include a browser, an application (APP), or a web application such as an H5 (Hypertext Markup Languages) application, or a lightweight application (also known as a mini-program), or a cloud application, etc. The end-side device can be developed based on the software development kit (SDK) provided by the server, such as based on a real-time communication (RTC) SDK. The end-side device can be deployed in an electronic device and depends on the device's operation or certain APPs within the device to run. Electronic devices can have a display screen and support information browsing, such as personal mobile terminals like mobile phones, tablets, and personal computers. Various other types of applications can also be configured in electronic devices, such as data storage and backup applications, human-computer interaction applications, model training applications, text processing applications, web browser applications, shopping applications, search applications, instant messaging tools, email clients, and social media platform software. Referring to Figure 2, Figure 2 shows a flowchart of a write node determination method according to an embodiment of this disclosure, applied to metadata nodes in a distributed system, specifically including the following steps 202-206: Step 202: Receiving a data write request sent by a client, wherein the data write request carries access frequency information. In practice, a distributed system includes a client, a metadata node, and at least two worker nodes. Users can initiate data write requests through the client to write data to the worker nodes of the distributed system. Upon receiving the user's data creation instruction, the client can send a data write request to the metadata node, including access frequency information for the data to be written. This allows the metadata node to allocate worker nodes with appropriate parameters based on the access frequency information. The data to be written is the data intended to be stored in the distributed system, such as photos, documents, downloaded videos, etc.Access frequency information indicates the frequency of access to the data to be written. This information can be an access frequency identifier, such as high-frequency access, low-frequency access, etc.; or it can be the access frequency itself, such as multiple accesses per day, once per day, once per week, X accesses within a unit of time, etc.; or it can directly identify the frequency type, such as hot data type, cold data type, etc.; or it can be the data attribute information of the data to be written, using the data attribute information to indicate its access frequency. For example, the data attribute can be backup photos, work documents, backup videos, etc. Generally, work documents may have a higher access frequency, while backup photos and backup videos may have a lower access frequency. In practice, this access frequency information can be specified by the user when writing data to the distributed system through the client, or it can be obtained by the client analyzing the attributes of the data to be written. It should be noted that the data write request indicates that the data to be written can be one or at least two. If it is at least two, the data write request needs to carry the identifier of each data to be written, as well as the corresponding access frequency information. In this embodiment of the disclosure, when the client initiates a data write request to the metadata node, it can carry access frequency information. This access frequency information can serve as a reference for the metadata node to allocate write nodes, so that the metadata node can perceive the hot and cold access frequency of user data and consider the working parameters of the worker nodes when writing user data into the distributed system, so that the data written by the worker nodes is adapted to its working parameters, and the performance of each worker node is fully utilized. Step 204: Based on the access frequency information and the working parameters of at least two worker nodes in the distributed system, determine the target write node from the at least two worker nodes. Different machine models have different working parameters for their worker nodes. These parameters refer to the worker node's performance parameters and stored data information. Performance parameters indicate various performance aspects of the worker node, such as read / write performance and storage capacity. Stored data information indicates relevant information about the data currently stored on the worker node, such as the amount of hot data, the amount of cold data, the ratio of hot to cold data, and the total data volume. It's important to note that the metadata node in the distributed system can obtain the working parameters of at least two worker nodes. Combining access frequency information with these parameters, it selects the target write node from these two worker nodes. When determining the target write node, access frequency and the worker node's working parameters are considered to ensure that the data written to the worker node matches its working parameters.In an optional implementation of this embodiment, the metadata node can also read the working parameters of each worker node in the distributed system. That is, the method for determining the write node further includes: reading the working parameters of at least two worker nodes every preset time interval. The working parameters include performance parameters and storage data information. The performance parameters include node read / write performance and storage capacity. The storage data information includes the amount of hot data stored and the total amount of data. The preset time interval is the time interval configured in the metadata node for updating the working parameters of the worker nodes; for example, the preset time interval can be 12 hours, 1 day, 1 week, 1 month, etc. In practical applications, because worker nodes continuously write data, their operating parameters change over time. Therefore, the metadata node can periodically read the operating parameters of each worker node in the distributed system at preset intervals. These operating parameters can include performance parameters and storage data information. Performance parameters include node read / write performance and storage capacity. Storage data information includes the amount of hot data and the total amount of data. Combining node read / write performance and the amount of hot data can determine whether the worker node's access capability is suitable for storing data with high access frequency. Combining storage capacity and the total amount of data can determine whether the worker node's space utilization is suitable for storing data with low access frequency. In addition to the metadata node periodically reading the operating parameters of at least two worker nodes, worker nodes in the distributed system can also periodically and proactively report their own operating parameters to the metadata node, or report their own operating parameters to the metadata node when their operating parameters change. This disclosure does not restrict the method by which the metadata node obtains the operating parameters of at least two worker nodes. In this embodiment, the metadata node can periodically acquire the working parameters of each worker node in the distributed system, thereby monitoring the usage of each worker node in the distributed system. When data needs to be written, it can refer to the current usage of each worker node to allocate a suitable write node and make full use of the worker node's performance. In an optional implementation of this embodiment, determining the target write node from at least two worker nodes based on access frequency information and the working parameters of at least two worker nodes in the distributed system includes: determining the target frequency type indicated by the access frequency information, and determining the target selection strategy corresponding to the target frequency type, wherein the target selection strategy is a write node filtering rule configured for the target frequency type; determining the target write node from at least two worker nodes based on the target selection strategy and the target working parameters of at least two worker nodes, wherein the target selection strategy is configured with a corresponding target parameter type, and the target working parameters are the working parameters of the target parameter type.In practice, frequency types can be categorized based on access frequency. For example, frequency types can be divided into high-frequency access, medium-frequency access, and low-frequency access, or into hot data types and cold data types. Of course, in real-world scenarios, frequency types can be further subdivided into many other categories based on requirements. It's important to note that access frequency information indicates the frequency of the data to be written. Therefore, the metadata node can determine the target frequency type indicated by the received access frequency information—that is, whether the data the user wants to write is of high-frequency or low-frequency access type. Based on the target selection strategy corresponding to the target frequency type, the target write node is selected. In this embodiment, different frequency types have different access characteristics and storage requirements, and different factors need to be considered when selecting write nodes. Therefore, the metadata node is configured with multiple selection strategies corresponding to different frequency types. After determining the corresponding target frequency type based on the received access frequency information, the target selection strategy corresponding to the target frequency type is adopted to analyze and judge the target working parameters of the at least two working nodes. Based on the parameter dimensions required by the target selection strategy, the target write node is determined from the at least two working nodes. In this way, for different frequency types, different selection strategies can be used to filter target write nodes to adapt to the characteristics of the corresponding frequency type, analyze from the corresponding dimensions, maximize the full utilization of the performance of working nodes, and store data with different access frequencies. In an optional implementation of this embodiment, the access frequency information is data attribute information; determining the target frequency type indicated by the access frequency information includes: determining the target frequency type corresponding to the data attribute information according to the attribute and frequency type correspondence, wherein the attribute and frequency type correspondence is pre-configured by the metadata node. It should be noted that access frequency information can be data attribute information. The metadata node, based on this data attribute information, queries the correspondence between attributes and frequency types to determine the target frequency type corresponding to the data attribute information. This data attribute information can include data type, data identifier, data title, data remarks, and other attribute-related information, which can, to a certain extent, indicate the frequency of data access. In actual implementation, the metadata node can be configured with the correspondence between attributes and frequency types, pre-storing the correspondence between different attributes and frequency types. This correspondence can be configured in the metadata node by staff based on experience or statistics to determine the access situation corresponding to different attributes, or it can be read by the metadata node from other platforms or databases. As an example, the correspondence between attributes and frequency types can be shown in Table 1 below. Table 1: Correspondence between Attributes and Frequency TypesIn this embodiment, the access frequency information can be data attribute information. Based on the data attribute information, the metadata node automatically analyzes and obtains the corresponding target frequency type. When a user writes data to be written through the client, they do not need to manually specify the access frequency of the data to be written, simplifying user operations and improving the user experience. In another possible implementation, if the access frequency information is an access frequency identifier, the metadata node directly uses the access frequency identifier as the corresponding target frequency type. For example, if the access frequency identifier received by the metadata node is high-frequency access, then the corresponding target frequency type is determined to be a high-frequency access type. Alternatively, the access frequency identifier can be converted into other corresponding type representations. For example, if the access frequency identifier received by the metadata node is high-frequency access, then the corresponding target frequency type is converted to a hot data type. In yet another possible implementation, if the access frequency information directly identifies the frequency type, the metadata node can directly use the access frequency information carried in the received data write request as the determined target frequency type. For example, if the access frequency information carried in the data write request is a hot data type, then the metadata node can determine the target frequency type to be a hot data type. In another possible implementation, if the access frequency information is the number of accesses, the metadata node can be configured with a correspondence between frequency range and frequency type, so that the metadata node can determine the corresponding target frequency type based on this correspondence. For example, if the access frequency received by the metadata node is X accesses per unit time, the correspondence between frequency range and frequency type in the metadata node is: less than Y accesses per unit time corresponds to cold data type, and greater than or equal to Y accesses per unit time corresponds to hot data type. In this disclosure, the access frequency information can be in various forms, either manually specified by the user or automatically analyzed by the metadata node. The method of identifying data access frequency is flexible and can adapt to various application scenarios and meet diverse user needs. In an optional implementation of this embodiment, determining the target selection strategy corresponding to the target frequency type includes: when the target frequency type is hot data type, determining the target selection strategy as selecting the target write node based on the unit access capacity of at least two working nodes; when the target frequency type is cold data type, determining the target selection strategy as selecting the target write node based on the space utilization rate of at least two working nodes. The unit access capacity is a numerical value. The smaller the value, the lower the amount of hot data stored per unit QPS, and the more suitable the worker node is for storing hot data.It should be noted that, taking the frequency-based classification of data types into hot and cold data types as an example, hot data types are accessed more frequently and require higher read / write capabilities. Therefore, for hot data types, the target selection strategy can be to select target write nodes based on the unit access capability of at least two working nodes. For cold data types, they are rarely accessed after storage and only require maximizing space utilization. Therefore, for cold data types, the target selection strategy can be to select target write nodes based on the space utilization of at least two working nodes. In this embodiment, for hot data types, target write nodes are selected based on the unit access capability of working nodes, ensuring that the working nodes for writing hot data are matched with their unit access capabilities, and the amount of hot data placed on different working nodes remains basically consistent under the unit access capabilities. For cold data types, target write nodes are selected based on the space utilization of working nodes, ensuring that the working nodes for writing cold data are matched with their space utilization, and the cold data is placed in the corresponding target write nodes, ensuring that the space utilization on different working nodes remains basically consistent. For hot and cold data, different dimensions are used for analysis, making fuller use of and balancing the performance of different working nodes. It should be noted that the above selection strategies for hot and cold data types are merely examples. In real-world scenarios, frequency types can be further subdivided into other more refined types, and corresponding selection strategies can be configured based on the characteristics of the corresponding data types. For example, for regular data types, the target selection strategy is to randomly select from idle worker nodes, or it can be to select the target write node based on the unit access capacity and space utilization of at least two worker nodes. In addition to the selection strategies for hot and cold data types mentioned above, metadata nodes can also configure other selection strategies for hot and cold data types. For example, for hot data types, the target selection strategy can be to select the target write node based on the unit access capacity of at least two worker nodes and the ratio of hot to cold data volume; for cold data types, the target selection strategy can be to select the target write node based on the space utilization of at least two worker nodes and the amount of hot and cold data already stored, etc. This disclosure does not impose any limitations on these aspects.In an optional implementation of this embodiment, when the target selection strategy is to select the target write node based on the unit access capability of at least two working nodes, the target write node is determined from at least two working nodes according to the target selection strategy and the target working parameters of at least two working nodes. This includes: determining the unit access capability of each working node based on the node read / write performance and the amount of hot data stored, wherein the unit access capability of each working node is the amount of hot data corresponding to the unit query rate of that working node; and determining the target write node based on the unit access capability of each working node. It should be noted that for hot data types, the target parameter type corresponding to the target selection strategy is node read / write performance and the amount of hot data stored. The target write node needs to be selected based on the unit access capability of at least two working nodes. Specifically, the unit access capability of each working node can be calculated based on the node read / write performance and the amount of hot data stored, where the unit access capability is the amount of hot data corresponding to the unit query rate of the working node, and then the target write node is determined based on the unit access capability of each working node. In practice, for any worker node, the amount of hot data stored on that node can be divided by its unit query rate to obtain the amount of hot data corresponding to the unit query rate. The amount of hot data corresponding to the unit query rate represents the unit access capability of that worker node, and the target write node is determined based on the unit access capability of each worker node. Specifically, IOPS (Input / Output Operations Per Second) can be used to test the QPS (Query Per Second) of a worker node. IOPS is a measurement method used for performance testing of computer storage devices (such as hard disk drives (HDDs), solid-state drives (SSDs), or storage area networks (SANs), and can be considered as measuring the number of read and write operations per second. For example, the unit query rate can be the QPS of a worker node. If a worker node has a QPS of 100 and stores 10TB of hot data, then the amount of hot data stored per unit QPS is 0.1TB, meaning the unit access capacity is 0. In this embodiment of ITBo, the distributed system metadata node analyzes the QPS of each worker node and the amount of hot data stored on that worker node to obtain the unit access capacity of each worker node. Based on the unit access capacity, the target write node is selected to ensure that the amount of hot data stored per unit QPS on different worker nodes is basically the same, thus making full use of the access capacity of different worker nodes to achieve high-frequency access to hot data.In an optional implementation of this embodiment, determining the target write node based on the unit access capability of each worker node includes: determining whether the unit access capability of each worker node meets a first access constraint; if the first access constraint is met, determining the worker node with the lowest unit access capability as the target write node; if the first access constraint is not met, selecting the target write node based on set security constraints. The first access constraint can be pre-configured on the metadata node, representing the conditions that must be met to select the worker node with the lowest unit access capability as the target write node. For example, the first access constraint can be configured as a large difference in unit access capability among the worker nodes, or inconsistent unit access capabilities among the worker nodes. If the first access constraint is met, the worker node with the lowest unit access capability can be selected as the target write node. A large difference in unit access capability can mean the difference between the maximum and minimum is greater than a set threshold, or the difference in unit access capability between any two worker nodes is greater than a set threshold. Furthermore, the set security constraints are rules based on access security settings without considering performance. These security constraints can be default rules of the distributed system or rules manually configured by the user, such as requiring each replica to be placed on a worker node of a different machine type, or random selection. As an example, the distributed system metadata node analyzes the QPS (Queries Per Second) and the amount of hot data stored on each worker node. For hot data, if the amount of hot data per unit QPS varies significantly across worker nodes, the metadata node will prioritize the worker node with the lowest amount of hot data per unit QPS, thus selecting it as the target write node. If the amount of hot data per unit QPS is roughly the same across worker nodes, the metadata node can randomly select a worker node while meeting security requirements, i.e., selecting the target write node based on set security constraints.In this embodiment, the metadata node can determine whether the unit access capability of each working node meets the first access constraint condition. If it does, the working node with the lowest unit access capability is selected as the target write node. If it does not meet the constraint, a target write node is selected based on the set security constraints. This ensures that the amount of hot data placed per QPS on different working nodes is basically consistent, thereby making full use of the access capabilities of different working nodes. This avoids excessive hot data placed per QPS on a certain working node, leading to excessive pressure during high-frequency access and affecting access services, while other working nodes have insufficient hot data placed per QPS, resulting in idle access capabilities that cannot be fully utilized. In an optional implementation of this embodiment, when the target selection strategy is to select the target write node based on the space utilization rate of at least two working nodes, the target write node is determined from at least two working nodes according to the target selection strategy and the target working parameters of at least two working nodes. This includes: determining the space utilization rate of each working node based on the storage space capacity and total storage data of at least two working nodes; and determining the target write node based on the space utilization rate of each working node. It should be noted that for cold data types, the target parameter types corresponding to the target selection strategy are storage space capacity and total data volume. The target write node needs to be selected based on the space utilization rate of at least two working nodes. Specifically, the space utilization rate of each working node can be calculated based on the storage space capacity and total data volume of at least two working nodes, and then the target write node is determined based on the space utilization rate of each working node. In actual implementation, for any working node, the space utilization rate of that working node can be obtained by dividing the total data volume stored on that working node by its storage space capacity. In this embodiment, selecting the target write node for storing cold data based on the space utilization rate of each working node can avoid excessive space utilization on a certain working node, leading to excessive storage pressure and affecting access services, while other working nodes have excessively low space utilization, resulting in a large amount of idle storage space that cannot be fully utilized. In one optional implementation of this embodiment, determining the target write node based on the space utilization rate of each working node includes: determining whether the space utilization rate of each working node meets the second access constraint condition; if the second access constraint condition is met, determining the working node with the lowest space utilization rate as the target write node; if the second access constraint condition is not met, selecting the target write node based on the set security constraints.The second access constraint can be pre-configured on the metadata node to represent the conditions that must be met to select the worker node with the lowest space utilization as the target write node. For example, the second access constraint could be configured as a large difference in space utilization among the worker nodes, or inconsistent space utilization among the worker nodes. If the second access constraint is met, the worker node with the lowest space utilization can be selected as the target write node. A large difference in space utilization could mean the difference between the maximum and minimum space utilization exceeds a set threshold, or the difference between the space utilization of any two worker nodes exceeds a set threshold. As an example, for cold data, if the space utilization of the worker nodes differs significantly, the distributed system metadata node will prioritize placing the cold data block on the worker node with the lowest space utilization, i.e., selecting the worker node with the lowest space utilization as the determined target write node. If the space utilization of the worker nodes is basically consistent, the metadata node can randomly select a worker node while meeting security requirements, i.e., selecting the target write node based on set security constraints. In this embodiment, the metadata node can determine whether the space utilization rate of each working node meets the second access constraint condition. If it does, the working node with the lowest space utilization rate is selected as the target write node. If it does not, a target write node is selected based on the set security constraints to ensure that the space utilization rate of different working nodes is basically consistent, thereby making full use of the storage capacity of different working nodes and avoiding excessive storage pressure on a certain working node, which affects access services, while other working nodes have a large amount of idle storage space that cannot be fully utilized. In an optional implementation of this embodiment, the stored data information also includes the ratio of hot and cold data. After determining the target write node from at least two working nodes based on access frequency information and the working parameters of at least two working nodes in the distributed system, the method further includes: if the target write node is determined based on unit access capacity, obtaining the ratio of hot and cold data of the target write node; determining whether the unit access capacity and the ratio of hot and cold data match; if not, updating the target write node.It should be noted that the metadata node can also include the hot / cold data ratio in the working parameters it obtains from the worker node. After the metadata node filters out the target write node, for cases where the data to be written is a hot data type (i.e., the target write node is determined based on unit access capability), it can also determine whether the unit access capability and the hot / cold data ratio of the target write node are compatible. If they are incompatible, it means that the selected target write node is not suitable for writing the data to be written and cannot fully utilize the performance of the worker node. Therefore, the target write node can be updated. The specific update process can be to return and re-execute the above selection process to reselect the target write node, or new working parameters can be added to reselect the target write node, such as adding the hot / cold data ratio. In practical applications, determining whether the unit access capability and the hot / cold data ratio match can be based on the set matching rules. For example, the set matching rule can be: if the proportion of hot data and cold data exceeds the ratio threshold, then it is determined that the unit access capability and the hot / cold data ratio do not match. In other words, if the ratio of hot data to cold data exceeds the threshold, it means that the amount of hot data stored on that worker node far exceeds the amount of cold data. Even if the unit access capacity of that worker node is small, continuing to write hot data to that worker node may still lead to excessive access pressure. In this case, we can return to selecting another worker node with a smaller unit access capacity as the target write node. For example, assuming the amount of hot data per unit QPS is small at 0.1TB, corresponding to worker node A, worker node A is selected as the target write node. Assuming the hot-cold data ratio of worker node A is 100:1, exceeding the ratio threshold of 10:1, we can determine that the hot-cold data ratio of worker node A does not match its unit access capacity. We then return to re-select worker node B with a hot data volume of 0.5TB per unit QPS as the target write node. In this embodiment, the ratio of hot and cold data on the target write node can be viewed to determine whether the hot data placed on the node matches the unit access capacity. This ensures that the ratio of hot and cold data on different working nodes is adapted to the unit access capacity, avoiding excessive access pressure caused by too much hot data on a certain working node. Step 206: Feed back the target write node to the client. The working parameters of the target write node are adapted to the access frequency information, which is used by the client to write the data to be written. Applying the scheme of this embodiment, after the metadata node determines the target write node, it feeds it back to the client, and the client writes the data to be written to the target write node.When different worker nodes in a distributed system have different read capabilities and disk block sizes, for hot data, the unit access capability of the worker node can be referenced to place the hot data on the worker nodes of the distributed system, so that user data read and write requests can be evenly distributed to each worker node according to the unit access capability of the worker node; for cold data, the space utilization rate of the worker node can be referenced to place the cold data on the worker nodes of the distributed system, so as to ensure that the data volume of each worker node can match the total data capacity of the node. Through the above strategy, the distributed system can ensure that the access capability and storage space of the worker nodes are fully utilized. Referring to Figure 3, Figure 3 shows a flowchart of a data writing method according to an embodiment of the present disclosure, applied to a client in a distributed system, specifically including the following steps 302-306: Step 302: Receive a data creation instruction, wherein the data creation instruction carries the data to be written and access frequency information. Step 304: Send a data write request carrying access frequency information to the metadata node in the distributed system. Step 306: Receive the target write node returned by the metadata node, and write the data to be written to the target write node. The target write node is determined by the metadata node based on access frequency information and the working parameters of at least two worker nodes in the distributed system. The working parameters of the target write node are compatible with the access frequency information. In this embodiment, the user initiates a data creation command through a client in the distributed system. The client sends a data write request carrying access frequency information to the metadata node in the distributed system. The metadata node determines the target write node whose working parameters are compatible with the access frequency information and feeds it back to the client. The client then writes the data to be written to the target write node. In this way, the metadata node can sense the access frequency of the data to be written and determine the worker node with compatible working parameters as the target write node based on the access frequency. The client writes data based on the target write node selected by the metadata node, fully utilizing the performance of the worker nodes and improving the service quality of the distributed system. The above is an illustrative scheme of a data writing method according to this embodiment. It should be noted that the technical solution of this data writing method and the technical solution of the above-mentioned write node determination method belong to the same concept. For details not described in detail in the technical solution of the data writing method, please refer to the description of the technical solution of the above-mentioned write node determination method.Corresponding to the above method embodiments, this disclosure also provides a distributed system embodiment. Referring to Figure 4, Figure 4 shows a schematic diagram of the structure of a distributed system according to an embodiment of this disclosure. As shown in Figure 4, the distributed system includes a client 402, a metadata node 404, and at least two worker nodes 406. The client 402 is configured to receive a data creation instruction, wherein the data creation instruction carries data to be written and access frequency information; and to send a data write request carrying access frequency information to the metadata node 404. The metadata node 404 is configured to determine a target write node 4062 from the at least two worker nodes 406 based on the access frequency information and the working parameters of the at least two worker nodes 406, and to feed back the target write node 4062 to the client 402, wherein the working parameters of the target write node 4062 are adapted to the access frequency information. The client 402 is further configured to write the data to be written to the target write node 4062. In this embodiment of the disclosure, the distributed system includes a client, a metadata node, and at least two worker nodes. When a client wants to write data to a worker node in a distributed system, the metadata node senses the access frequency of the data to be written and determines the worker node with suitable working parameters as the target writing node based on the access frequency. This information is then fed back to the client, who then writes the data to the target writing node. In this way, the metadata node selects the worker node corresponding to the suitable working parameters based on the access frequency in the distributed system, fully utilizing the performance of the worker nodes and improving the service quality of the distributed system. The above is an illustrative scheme of a distributed system according to this embodiment. It should be noted that the technical solution of this distributed system and the technical solution of the writing node determination method described above belong to the same concept. Details not described in detail in the technical solution of the distributed system can be found in the description of the technical solution of the writing node determination method described above. The following, in conjunction with Figure 5, takes the application of the data writing method provided in this disclosure in a user's personal computer scenario as an example to further illustrate the data writing method. Figure 5 shows a flowchart of the processing procedure of a data writing method provided in an embodiment of this disclosure, specifically including the following steps. Step 502: The metadata node in the distributed system periodically obtains the working parameters of each worker node in the distributed system. These working parameters include performance parameters and storage data information. The performance parameters include node read / write performance and storage space capacity. The storage data information includes the amount of hot data stored and the total amount of data.Step 504: The user initiates a data creation command on their personal computer. This command instructs the writing of a "working document" and the download of a video to the distributed system. The access frequency information for the "working document" is daily, and the access information for the "downloaded video" is once. Step 506: The personal computer sends a data write request to the metadata node in the distributed system. This request carries the "working document" and its corresponding access frequency information "daily," and the "downloaded video" and its corresponding access frequency information "once." Step 508: The metadata node determines the target frequency type as a "hot" data type based on "daily access" and as a "cold" data type based on "once access." Step 510: For hot data types, the metadata node calculates the amount of hot data stored per unit QPS based on the QPS and the amount of hot data stored in each worker node, and selects the worker node with the smallest amount of hot data stored per unit QPS as worker node A as the target write node corresponding to "Work Document"; For cold data types, the metadata node obtains the space utilization rate of each worker node by dividing the total amount of data stored in each worker node by the corresponding storage space capacity, and selects the worker node B with the smallest space utilization rate as the target write node corresponding to "Download Movie". Step 512: The metadata node feeds back to the personal computer the worker node A corresponding to "Work Document" and the worker node Bo corresponding to "Download Movie". Step 514: The personal computer writes the specific data corresponding to "Work Document" to worker node A and the specific data corresponding to "Download Movie" to worker node Bo. In this embodiment of the disclosure, the user initiates a data creation command through the personal computer. The personal computer sends a data write request carrying the data identifier of each data to be written and the corresponding access frequency information to the metadata node in the distributed system. The metadata node determines the suitable target write node based on the selection strategy corresponding to the access frequency information and feeds it back to the personal computer. The personal computer writes the data to be written to the corresponding target write node. The metadata node allocates target write nodes based on the access frequency of the data to be written, making full use of the performance of each worker node in the distributed system and improving the service quality of the distributed system. Corresponding to the above method embodiment, this disclosure also provides an embodiment of a metadata node in a distributed system. Figure 6 shows a schematic diagram of the structure of a metadata node in a distributed system provided by an embodiment of this disclosure.As shown in Figure 6, the metadata node includes: a first receiving module 602, configured to receive data write requests sent by a client, wherein the data write request carries access frequency information; a determining module 604, configured to determine a target write node from at least two working nodes based on the access frequency information and the working parameters of at least two working nodes in the distributed system; and a feedback module 606, configured to feed back the target write node to the client, wherein the working parameters of the target write node are adapted to the access frequency information for the client to write the data to be written. Optionally, the determining module 604 is further configured to: determine the target frequency type indicated by the access frequency information, and determine the target selection strategy corresponding to the target frequency type, wherein the target selection strategy is a write node filtering rule configured for the target frequency type; and determine the target write node from at least two working nodes based on the target selection strategy and the target working parameters of at least two working nodes, wherein the target selection strategy is configured with a corresponding target parameter type, and the target working parameters are the working parameters of the target parameter type. Optionally, the access frequency information is data attribute information; the determining module 604 is further configured to: determine the target frequency type corresponding to the data attribute information according to the attribute-frequency type correspondence, wherein the attribute-frequency type correspondence is pre-configured by the metadata node. Optionally, the determining module 604 is further configured to: when the target frequency type is a hot data type, determine the target selection strategy as selecting the target write node based on the unit access capability of at least two working nodes; when the target frequency type is a cold data type, determine the target selection strategy as selecting the target write node based on the space utilization rate of at least two working nodes. Optionally, when the target selection strategy is to select the target write node based on the unit access capability of at least two working nodes, the determining module 604 is further configured to: determine the unit access capability of each working node according to the node read / write performance and the amount of hot data stored of at least two working nodes, wherein the unit access capability of each working node is the amount of hot data corresponding to the unit query rate of that working node; and determine the target write node according to the unit access capability of each working node. Optionally, the determining module 604 is further configured to: determine whether the unit access capability of each working node meets the first access constraint; if the first access constraint is met, determine the working node with the lowest unit access capability as the target write node; if the first access constraint is not met, select the target write node based on the set security constraints.Optionally, when the target selection strategy is to select the target write node based on the space utilization rate of at least two working nodes, the determining module 604 is further configured to: determine the space utilization rate of each working node based on the storage space capacity and total data volume of at least two working nodes; and determine the target write node based on the space utilization rate of each working node. Optionally, the determining module 604 is further configured to: determine whether the space utilization rate of each working node meets the second access constraint condition; if the second access constraint condition is met, determine the working node with the lowest space utilization rate as the target write node; if the second access constraint condition is not met, select the target write node based on the set security constraint. Optionally, the metadata node also includes a reading module, configured to: read the working parameters of at least two working nodes at preset intervals; wherein, the working parameters include performance parameters and storage data information, the performance parameters include node read / write performance and storage space capacity, and the storage data information includes the amount of hot data and the total data volume. Optionally, the stored data information also includes a hot and cold data ratio; the metadata node also includes an update module, configured to: if the target write node is determined based on unit access capability, obtain the hot and cold data ratio of the target write node; determine whether the unit access capability and the hot and cold data ratio match; if not, update the target write node. The metadata node in the distributed system provided in this embodiment includes a first receiving module, a determining module, and a feedback module. Utilizing the interaction and cooperation of the first receiving module, the determining module, and the feedback module, the access frequency of the data to be written is sensed, and based on the access frequency, a working node with suitable working parameters is determined as the target write node, enabling the client to write the data to be written to the target write node, fully utilizing the performance of each working node in the distributed system, and improving the service quality of the distributed system. The above is an illustrative scheme of a metadata node in a distributed system according to this embodiment. It should be noted that the technical solution of the metadata node in this distributed system and the technical solution of the write node determination method described above belong to the same concept. Details not described in detail in the technical solution of the metadata node in the distributed system can be found in the description of the technical solution of the write node determination method described above. Corresponding to the above method embodiments, this disclosure also provides a client embodiment in a distributed system. Figure 7 shows a schematic diagram of the structure of a client in a distributed system provided by an embodiment of this disclosure.As shown in Figure 7, the client includes: a second receiving module 702, configured to receive a data creation instruction, wherein the data creation instruction carries the data to be written and access frequency information; a sending module 704, configured to send a data write request carrying access frequency information to the metadata node in the distributed system; and a writing module 706, configured to receive the target writing node returned by the metadata node and write the data to be written to the target writing node, wherein the target writing node is determined by the metadata node based on the access frequency information and the working parameters of at least two working nodes in the distributed system, and the working parameters of the target writing node are adapted to the access frequency information. The client in the distributed system provided in this embodiment includes a second receiving module, a sending module, and a writing module. Through the interaction of the second receiving module, the sending module, and the writing module, data to be written is written to a target writing node. This target writing node is a working node obtained by the writing module and fed back by the metadata node. The metadata node determines the target writing node by sensing the access frequency of the data to be written and performing adaptation analysis with the working parameters of the working node, thereby fully utilizing the performance of each working node in the distributed system and improving the service quality of the distributed system. The above is an illustrative scheme of a client in a distributed system according to this embodiment. It should be noted that the technical solution of the client in this distributed system and the technical solution of the above-described data writing method belong to the same concept. Details not described in detail in the technical solution of the client in the distributed system can be found in the description of the technical solution of the above-described data writing method. Figure 8 shows a structural block diagram of a computing device according to an embodiment of this disclosure. The components of the computing device 800 include, but are not limited to, a memory 810 and a processor 820. The processor 820 and the memory 810 are connected via a bus 830, and a database 850 is used to store data. The computing device 800 also includes an access device 840, which 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.Access device 840 may include one or more of any type of wired or wireless network interface (e.g., a network interface card (NIC)), such as an IEEE 802.11 Wireless Local Area Network (WLAN) interface, a Wi-MAX (Worldwide Interoperability for Microwave Access) interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, or a Near Field Communication (NFC) interface. In one embodiment of this disclosure, the aforementioned components of computing device 800, as well as other components not shown in FIG. 8, may also be interconnected, for example, via a bus. It should be understood that the computing device block diagram shown in FIG. 8 is merely for illustrative purposes and is not intended to limit the scope of this disclosure. Those skilled in the art can add or replace other components as needed. The computing device 800 can be any type of stationary or mobile computing device, including mobile computers or mobile computing devices (e.g., tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (e.g., smartphones), wearable computing devices (e.g., smartwatches, smart glasses, etc.) or other types of mobile devices, or stationary computing devices such as desktop computers or personal computers (PCs). The computing device 800 can also be a mobile or stationary server. The processor 820 is configured to execute a computer program / instruction that, when executed by the processor, implements the steps of the above-described write node determination method or data writing method. The various embodiments in this disclosure are described in a progressive manner; similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for the computing device embodiments, since they are basically similar to the write node determination method or data writing method embodiments, the description is relatively simple; relevant parts can be referred to in the description of the write node determination method or data writing method embodiments.This disclosure also provides a computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implement the steps of the write node determination method or data writing method described above. The various embodiments in this disclosure are described in a progressive manner; similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the computer-readable storage medium embodiment is basically similar to the write node determination method or data writing method embodiment, so the description is relatively simple; relevant parts can be referred to the descriptions of the write node determination method or data writing method embodiment. This disclosure also provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the write node determination method or data writing method described above. The above is an illustrative scheme of a computer program product according to this embodiment. It should be noted that the technical solution of this computer program product belongs to the same concept as the technical solution of the write node determination method or data writing method described above. Details not described in detail in the technical solution of the computer program product can be referred to the description of the technical solution of the write node determination method or data writing method described above. Specific embodiments of this disclosure have been described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims may be performed in a different order than those shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous. Computer instructions include computer program code, which may be in the form of source code, object code, executable files, or certain intermediate forms. Computer-readable media may include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in computer-readable media may be appropriately added or removed according to the requirements of patent practice. For example, in some regions, according to patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of this disclosure are not limited to the described order of actions, because according to the embodiments of this disclosure, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments of this disclosure. In the above embodiments, the descriptions of each embodiment have different focuses; for parts not described in detail in a certain embodiment, refer to the relevant descriptions of other embodiments. The preferred embodiments of this disclosure disclosed above are only used to help illustrate this disclosure. Optional embodiments do not describe all details exhaustively, nor do they limit the invention to only the specific implementations described. Obviously, many modifications and variations can be made based on the content of the embodiments of this disclosure. This disclosure selects and specifically describes these embodiments to better explain the principles and practical applications of the embodiments of this disclosure, so that those skilled in the art can well understand and utilize this disclosure. This disclosure is only limited by the claims and their full scope and equivalents.

Claims

Claims 1. A method for determining write nodes, applied to metadata nodes in a distributed system, the method comprising: The system receives a data write request sent by a client, wherein the data write request carries access frequency information; based on the access frequency information and the working parameters of at least two working nodes in the distributed system, a target write node is determined from the at least two working nodes; the target write node is fed back to the client, wherein the working parameters of the target write node are adapted to the access frequency information, and are used by the client to write the data to be written.

2. The write node determination method according to claim 1, wherein determining the target write node from the at least two working nodes based on the access frequency information and the working parameters of at least two working nodes in the distributed system includes: The target frequency type indicated by the access frequency information is determined, and the target selection strategy corresponding to the target frequency type is determined, wherein the target selection strategy is a write node filtering rule configured for the target frequency type; the target write node is determined from the at least two working nodes according to the target selection strategy and the target working parameters of the at least two working nodes, wherein the target selection strategy is configured with a corresponding target parameter type, and the target working parameter is the working parameter of the target parameter type.

3. The write node determination method according to claim 2, wherein the access frequency information is data attribute information; determining the target frequency type indicated by the access frequency information includes: Based on the correspondence between attributes and frequency types, the target frequency type corresponding to the data attribute information is determined, wherein the correspondence between attributes and frequency types is pre-configured by the metadata node.

4. The write node determination method according to claim 2 or 3, wherein determining the target selection strategy corresponding to the target frequency type includes: When the target frequency type is a hot data type, the target selection strategy is determined to be to select the target write node based on the unit access capability of the at least two working nodes; When the target frequency type is a cold data type, the target selection strategy is determined to be to select the target write node based on the space utilization of the at least two working nodes.

5. The write node determination method according to claim 4, wherein when the target selection strategy is to select the target write node based on the unit access capability of the at least two working nodes, the step of determining the target write node from the at least two working nodes according to the target selection strategy and the target working parameters of the at least two working nodes includes: Based on the node read / write performance and the amount of hot data stored of the at least two working nodes, the unit access capability of each working node is determined, wherein the unit access capability of each working node is the amount of hot data corresponding to the unit query rate of the working node; the target write node is determined based on the unit access capability of each working node.

6. The write node determination method according to claim 5, wherein determining the target write node based on the unit access capability of each working node includes: Determine whether the unit access capability of each working node meets the first access constraint condition; If the first access constraint is met, the working node with the lowest unit access capability is determined as the [node name missing]. Target write node; if the first access constraint is not met, the target write node is selected based on the set security constraints.

7. The write node determination method according to claim 4, wherein when the target selection strategy is to select the target write node based on the space utilization rate of the at least two working nodes, the step of determining the target write node from the at least two working nodes according to the target selection strategy and the target working parameters of the at least two working nodes includes: Based on the storage space capacity and total data volume of the at least two working nodes, determine the space utilization rate of each working node; The target write node is determined based on the space utilization rate of each working node.

8. The write node determination method according to claim 7, wherein determining the target write node based on the space utilization rate of each working node includes: Determine whether the space utilization rate of each working node meets the second access constraint condition; If the second access constraint is met, the working node with the lowest space utilization rate is determined as the target write node; If the second access constraint is not met, the target write node is selected based on the set security constraints.

9. The write node determination method according to any one of claims 1-8, the method further comprising: The working parameters of the at least two working nodes are read at preset intervals; wherein, the working parameters include performance parameters and storage data information, the performance parameters include node read / write performance and storage space capacity, and the storage data information includes the amount of hot data stored and the total amount of data.

10. The write node determination method according to claim 9, wherein the stored data information further includes a hot-cold data ratio; after determining the target write node from the at least two working nodes based on the access frequency information and the working parameters of at least two working nodes in the distributed system, the method further includes: If the target write node is determined based on unit access capacity, obtain the ratio of hot and cold data of the target write node; Determine whether the unit access capacity and the ratio of hot and cold data match. If not, update the target write node.

11. A data writing method, applied to a client in a distributed system, comprising: The system receives a data creation instruction, wherein the data creation instruction carries data to be written and access frequency information; sends a data write request carrying the access frequency information to the metadata node in the distributed system; receives a target write node returned by the metadata node, and writes the data to be written to the target write node, wherein the target write node is determined by the metadata node based on the access frequency information and the working parameters of at least two working nodes in the distributed system, and the working parameters of the target write node are adapted to the access frequency information.

12. A distributed system, comprising a client, a metadata node, and at least two worker nodes; the client is configured to receive data creation instructions, wherein, The data creation instruction carries the data to be written and access frequency information; Send a data write request carrying the access frequency information to the metadata node; The metadata node is configured to determine a target write node from the at least two working nodes based on the access frequency information and the working parameters of the at least two working nodes, and to feed back the target write node to the client, wherein the working parameters of the target write node are adapted to the access frequency information; The client is further configured to write the data to be written to the target write node.

13. A metadata node in a distributed system, comprising: A first receiving module is configured to receive a data write request sent by a client, wherein the data write request carries access frequency information; a determining module is configured to determine a target write node from at least two working nodes based on the access frequency information and the working parameters of at least two working nodes in the distributed system; a feedback module is configured to feed back the target write node to the client, wherein the working parameters of the target write node are adapted to the access frequency information for the client to write the data to be written.

14. A client in a distributed system, comprising: The second receiving module is configured to receive a data creation instruction, wherein the data creation instruction carries data to be written and access frequency information; the sending module is configured to send a data write request carrying the access frequency information to the metadata node in the distributed system; the writing module is configured to receive a target writing node returned by the metadata node, and write the data to be written to the target writing node, wherein the target writing node is determined by the metadata node based on the access frequency information and the working parameters of at least two working nodes in the distributed system, and the working parameters of the target writing node are adapted to the access frequency information.

15. A computing device, comprising: Memory and processor; The memory is used to store computer programs / instructions, and the processor is used to execute the computer programs / instructions. When the computer programs / instructions are executed by the processor, they implement the steps of the write node determination method according to any one of claims 1 to 10 or the data writing method according to claim 11.

16. A computer-readable storage medium storing a computer program / instructions that, when executed by a processor, implement the steps of the write node determination method of any one of claims 1 to 10 or the data writing method of claim 11.

17. A computer program product comprising a computer program / instructions that, when executed by a processor, implement the steps of the write node determination method of any one of claims 1 to 10 or the data writing method of claim 11.

Images