WO2025124250A1 - Cluster shared traffic threshold-based rate limiting method and apparatus - Google Patents

Abstract

The present application relates to a cluster shared traffic threshold-based rate limiting method and apparatus, which are used for achieving effective bandwidth management and allocation among a plurality of nodes in a cluster and among a plurality of network card queues in the nodes. The method comprises: S1, a cluster manager issuing a cluster rate limiting threshold of a specified data stream to each node in a network cluster, and adding the forwarding process of each node into a multicast to share traffic information between the nodes, the cluster rate limiting threshold being acquired in the forwarding process of the node; S2, a node controller collecting statistics about the traffic of the specified data stream flowing into a network card per second, performing marking, sending out, by means of multicasting, a unique identifier of the node and network card traffic received per second to be shared to other nodes in the cluster, and on the basis of a multicast packet received from the cluster, calculating a node rate limiting quota threshold of the present node; S3, on the basis of each service core for processing traffic of queues in the forwarding process of the node and the bandwidth proportion of the total traffic of the nodes, a bandwidth allocation scheduler allocating a core rate limiting quota threshold of each node; and S4, on the basis of real-time data of the processed traffic of each queue in the nodes and the proportion of the total bandwidth of the nodes, a token bucket manager performing periodical and dynamical adjustment by means of a token bucket algorithm to obtain the proportion of each queue in a cluster rate limiting bandwidth, so as to limit the rate of the traffic passing through the queue.

Description

Cluster shared traffic threshold speed limiting method and device

Related Applications

This application claims priority to Chinese patent application number 202311721740.X, filed on December 14, 2023, and entitled “A speed limiting method and device for cluster shared traffic thresholds”, the entire text of which is hereby incorporated by reference.

Technical Field

The present application belongs to the field of network technology, and in particular, relates to a speed limiting method and device for a cluster shared traffic threshold.

Background Art

In modern computer networks, cluster systems are usually composed of multiple nodes. Each node processes data packets using multiple queues and binds multiple CPU cores to data independently and without locks to improve packet processing performance. In this distributed environment, it is a key challenge to effectively implement a certain type of traffic speed limit in a cluster and ensure the quality of network traffic and the effective use of resources. Traditional QOS methods are difficult to provide flexible and efficient traffic speed limit in this multi-node, multi-queue, and multi-core environment, resulting in network congestion, performance degradation, and waste of resources. It is also impossible to set bandwidth speed limit for the entire cluster.

Summary of the invention

In view of the above shortcomings of the prior art, this application aims to propose a technology based on cluster shared traffic speed limit thresholds to achieve effective traffic management between each node, queue and CPU core in the cluster. The technical solution of this application can dynamically allocate traffic speed limit resources in a distributed environment to ensure that each node and queue can operate within a reasonable speed limit range. At the same time, this technology can achieve fair distribution of traffic within the cluster and avoid a node or queue occupying too many traffic resources.

In a first aspect of the present application, a method for limiting the speed of a cluster shared traffic threshold is proposed, the method comprising:

S1, the cluster manager sends the cluster speed limit threshold of the specified data flow to each node in the network cluster, and adds the forwarding process of each node to the multicast to share traffic information between nodes, and the forwarding process in the node obtains the cluster speed limit threshold;

S2: The node controller counts the traffic of the specified data flow flowing into the network card per second and marks it. It sends the node unique identifier and the network card traffic received per second through multicast to share with other nodes in the cluster. It calculates the node speed limit quota threshold of the node based on the multicast message received from the cluster.

S3, the bandwidth allocation scheduler allocates the core speed limit quota threshold of each node according to the bandwidth ratio of each service core processing queue traffic in the node forwarding process to the total traffic of the node;

S4: The token bucket manager uses the token bucket algorithm to periodically and dynamically adjust the real-time data of the proportion of traffic processed by each queue in the node to the total bandwidth of the node, obtain the proportion of each queue in the cluster's speed-limited bandwidth, and limit the speed of the traffic passing through the queue.

In some of the embodiments, when a node receives a multicast message sent by other nodes, the node unique identifier in the multicast message is compared with the node unique identifier of the node itself. If the comparison result is equal, the multicast message is ignored. If the comparison result is unequal, the multicast message is recorded and the node speed limit quota threshold of the node is calculated.

In some embodiments, the calculation formula of the node speed limit quota threshold is as follows:
speed _{node i} ＝speed _cluster *traffic _{node i} /∑ traffic _{node i}

Where speed _cluster is the cluster speed limit threshold of the specified data flow, traffic _{node i} is the traffic volume of the specified data flow flowing into the node network card of node i within one second, i is the node number, ranging from 1 to N, and N is the number of cluster nodes.

In some embodiments, in step S3, the calculation formula of the core speed limit quota threshold is as follows:
speed _{node(i)_lcore(n)} =speed _node(i) *traffic _{node(i)_lcore(n)} /∑traffic _{node(i)_lcore(n)}

Among them, speed _{node(i)_lcore(n)} is the traffic volume of the specified data flow flowing from node i to service core n within this second, i is the node number, ranging from 1 to N, N is the maximum number of nodes in the cluster, n is the service core number, ranging from 1 to M, M is the maximum number of service cores.

In a second aspect of the present application, a speed limiting device for a cluster shared traffic threshold is proposed, the device is used to implement the above method, the device includes a cluster manager and multiple nodes located in the same network cluster;

The cluster manager is connected to multiple nodes respectively, and is used to coordinate and manage the entire multi-node cluster, and set, adjust and issue cluster speed limit thresholds according to speed limit requirements;

A node controller, which is arranged on each of the plurality of nodes and serves as a control main thread of the node forwarding process, and is used to monitor the local bandwidth utilization and flow;

The bandwidth allocation scheduler is set in the business core of the service process of each node to ensure fair allocation and utilization efficiency of bandwidth resources;

The token bucket manager is set on the thread bound to each queue served by each node. It is used to limit the speed of the traffic passing through the queue using the token bucket algorithm to achieve speed limit for the entire cluster.

In some of the embodiments, the node controller counts the local traffic data flowing into the network card every second, and shares the local traffic data via multicast.

In some embodiments, sharing the local traffic data includes: the node controller sends the local traffic data to each node in the cluster via multicast, the node receives traffic data from other nodes in the cluster, and after excluding its own traffic data, calculates the node's own node speed limit quota threshold based on the traffic ratio.

In some of the embodiments, ensuring fair allocation and efficient utilization of bandwidth resources specifically includes: the bandwidth allocation scheduler determines the node speed limit quota threshold according to the traffic bandwidth ratio of the nodes in the cluster, and allocates the core speed limit quota threshold of each node according to the bandwidth ratio of each business core processing queue traffic in the node forwarding process to the total traffic of the node.

In some of the embodiments, the token bucket algorithm is used to limit the speed of the traffic passing through the queue, specifically including: the token bucket manager periodically and dynamically adjusts according to the real-time data of the ratio of the traffic processed by each queue in the node to the total bandwidth of the node, to obtain the ratio of each queue in the cluster speed limit bandwidth.

According to a third aspect of the present application, an electronic device is provided, comprising: a memory unit and a processor unit, wherein a computer program is stored in the memory unit, and the processor unit implements the above method when executing the program.

The beneficial effects of the present application are as follows: In the prior art, the traffic speed limit of the cluster is often limited only by the node, without dynamic adjustment according to the traffic, and it is often difficult to achieve accurate speed limit of the entire cluster. The present application monitors the traffic bandwidth ratio in real time, effectively and quickly dynamically adjusts the speed limit quota between different cores (threads) of the same node process service and different nodes of the same cluster, determines the token bucket generation rate of each node and each queue to effectively limit the speed, regards the entire cluster as a logical token bucket, and implements it in a way that each core of each node has an independent token bucket, so as to achieve lock-free speed limit of the cluster without affecting the performance of the cause traffic processing.

BRIEF DESCRIPTION OF THE DRAWINGS

The accompanying drawings are only used to illustrate specific embodiments and are not considered to limit the present application. In the entire drawings, the same reference symbols represent the same components. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

FIG1 is a schematic diagram of the structure of a cluster controller according to an embodiment of the present application.

FIG. 2 is a schematic diagram of the structure of a node controller according to an embodiment of the present application.

FIG3 is a schematic diagram of a processing flow of a bandwidth allocation scheduler and a token bucket manager according to an embodiment of the present application.

DETAILED DESCRIPTION

In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present application, the technical solutions of the present application will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. It should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present application. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in the field without making creative work should fall within the scope of protection of the present application.

Furthermore, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessarily confusing the concepts disclosed in the present application.

In the description of the present application, it should be noted that, unless otherwise expressly specified and limited, the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first", "second", and "third" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance. The terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be a connection between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of methods and systems consistent with some aspects of the present application as detailed in the appended claims.

The following is an explanation of the meanings of the technical terms used in the technical solution:

Cluster: refers to an architecture that connects multiple computers (also called nodes or servers) together to complete computing tasks together.

Node: A server node in a cluster refers to each independent computer or server in the cluster. Each node is an independent entity.

Quality of Service (QoS) refers to a network's ability to use various basic technologies to provide better service capabilities for specified network communications. It is a network security mechanism and a technology used to solve problems such as network delays and congestion.

Forwarding process: refers to the main service process running in the cluster node, which is responsible for forwarding data packets from one network node to another network node. Here, it is the host process implemented by the speed limiting method of this application.

In a first aspect, the present application proposes a rate limiting device for cluster shared traffic threshold based on a token bucket algorithm, as shown in FIG1 , the device includes a cluster manager and multiple nodes located in the same network cluster;

The cluster manager is connected to multiple nodes respectively, and is used to coordinate and manage the entire multi-node cluster, and set, adjust and issue cluster speed limit thresholds according to speed limit requirements;

A node controller, which is arranged on each of the plurality of nodes and serves as a control main thread of the node forwarding process, and is used to monitor the local bandwidth utilization and flow;

The bandwidth allocation scheduler is set in the business core of the service process of each node to ensure fair allocation and utilization efficiency of bandwidth resources;

The token bucket manager is set on the thread bound to each queue served by each node. It is used to limit the speed of the traffic passing through the queue using the token bucket algorithm to achieve speed limit for the entire cluster.

The node controller counts the local traffic data flowing into the network card every second, and shares the local traffic data by multicast;

Sharing the local traffic data includes: the node controller sends the local traffic data to each node in the cluster by multicasting, the node receives the traffic data from other nodes in the cluster, and after excluding its own traffic data, calculates the node speed limit quota threshold of the node itself according to the traffic ratio;

The ensuring of fair allocation and utilization efficiency of bandwidth resources specifically includes: the bandwidth allocation scheduler determines the node speed limit quota threshold according to the traffic bandwidth ratio of the nodes in the cluster, and allocates the core speed limit quota threshold of each node according to the bandwidth ratio of each service core processing queue traffic in the node forwarding process to the total traffic of the node;

The token bucket algorithm is used to limit the speed of the traffic passing through the queue, specifically including: the token bucket manager periodically and dynamically adjusts the proportion of the traffic processed by each queue in the node to the total bandwidth of the node according to the real-time data, and obtains the proportion of each queue in the cluster speed limit bandwidth;

In a second aspect, the present application proposes, in one embodiment, a method for cluster shared traffic threshold based on a token bucket algorithm, the method comprising the following steps S1 to S4.

S1: The cluster manager sends a cluster speed limit threshold of a specified data flow to each node in the network cluster, and adds the forwarding process of each node to the multicast to share traffic information between nodes. The forwarding process in the node obtains the cluster speed limit threshold.

In the above steps, the cluster manager issues certain traffic speed limit thresholds, such as traffic speed limit thresholds for UDP, TCP-SYN, TCP-ACK, ICMP, etc. Here, the traffic threshold of the speed limit cluster TCP-SYN identifier is taken as speed _cluster (unit: bit/s) as an example. The forwarding processes running on N nodes in the cluster obtain the cluster TCP-SYN traffic speed limit speed _cluster , as shown in Figure 1.

In the above steps, this is achieved by adding the forwarding process of the node to the multicast address, such as the multicast address 224.0.1.100.

S2: The node controller counts the traffic of the specified data flow flowing into the network card per second and marks it, sends the node unique identifier and the network card traffic received per second through multicast to share with other nodes in the cluster, and calculates the node speed limit quota threshold of this node based on the multicast messages received from the cluster.

In the above steps, the node rate limit quota threshold of the node is calculated based on the multicast message received from the cluster, which includes:

When a node receives a multicast message sent by another node, it compares the node unique identifier in the multicast message with its own node unique identifier. If the comparison result is equal, the multicast message is ignored. If the comparison result is not equal, the multicast message is recorded and the node speed limit quota threshold of the node is calculated. The calculation formula is as follows:
speed _{node i} ＝speed _cluster *traffic _{node i} /∑ traffic _{node i}

Where speed _cluster is the cluster speed limit threshold of the specified data flow, traffic _{node i} is the traffic volume of the specified data flow flowing into the node network card of node i within one second, i is the node number, ranging from 1 to N, and N is the number of cluster nodes.

Continuing with the example in the previous steps, the node forwarding process counts the TCP-SYN traffic flowing into the network card every second (by parsing all the network card queue traffic packets, judging that the transport layer protocol number in the IP packet is 6, which is the TCP protocol traffic, judging that the SYN bit in the Flag field of the TCP layer header is 1 and the other fields of the Flag are all 0, then the TCP-SYN traffic that meets the statistics is marked as traffic _node1 , traffic _node2 , ... traffic _nodeN , and the node unique identifier ID _nodeN and the network card traffic received per second traffic _nodeN are sent out through multicast to share with other nodes in the cluster. When a node receives the multicast packets received by other nodes, it ignores the multicast packets based on the comparison between the node unique identifier ID _nodeN in the multicast packets and the node unique identifier ID _nodeN of the node. If they are equal, it excludes the ID sent by the node itself, and records the others. By calculating the traffic ratio of the node, it is the incremental TCP-SYN traffic traffic flowing into the node network card of the node within this second. Divide _nodei by the sum of the incremental TCP-SYN traffic flowing into the network cards of all nodes in the cluster within one second ∑traffic _nodei (i is the node number, ranging from 1 to N, N is the number of cluster nodes), and calculate the speed limit quota threshold of this node as speed _{node i} = speed _cluster * traffic _{node i} / ∑traffic _{node i} .

S3: The bandwidth allocation scheduler allocates the core speed limit quota threshold of each node according to the bandwidth ratio of each service core processing queue traffic in the node forwarding process to the total traffic of the node.

The calculation formula of the core speed limit quota threshold is as follows:
speed _{node(i)_lcore(n)} =speed _node(i) *traffic _{node(i)_lcore(n)} /∑traffic _{node(i)_lcore(n)}

Among them, speed _{node(i)_lcore(n)} is the traffic volume of the specified data flow flowing from node i to service core n within this second, i is the node number, ranging from 1 to N, N is the maximum number of nodes in the cluster, n is the service core number, ranging from 1 to M, M is the maximum number of service cores.

In the above steps, continue to use the examples in the previous steps for description. Get the speed limit quota of TCP-SYN traffic flowing into each queue of each node network card: each network card queue n (n is the network card queue number, taking Intel network card 82599 as an example, the network card has 16 queues, and n ranges from 0 to 15) of the node forwarding process is independently assigned a business core thread to process lcore(n) and bound to CPU(n) for exclusive processing. The lcore(n) business core thread counts the corresponding TCP-SYN traffic flowing into the bound network card queue n every second (by parsing the traffic message of queue n, judging that the transport layer protocol number in the IP message is 6 for TCP protocol traffic, judging that the SYN bit of the Flag field in the TCP layer header is 1 and the other fields of the Flag are all 0, then the TCP-SYN traffic that meets the statistics is identified as traffic _{node1_lcore1} , traffic _{node1_lcore2} , and traffic _{node1_lcoreM} . The speed limit quota threshold of a business core traffic of each node can be calculated as speed _{node(i)_lcore(n)} = speed _node(i) * traffic _{node(i)_lcore(n)} /∑ traffic _{node(i)_lcore(n)} .

S4: The token bucket manager uses the token bucket algorithm to periodically and dynamically adjust the real-time data of the proportion of traffic processed by each queue in the node to the total bandwidth of the node, obtain the proportion of each queue in the cluster's speed-limited bandwidth, and limit the speed of the traffic passing through the queue.

Among them, the bucket depth of the token bucket size _lcore(i) = speed _{node(i)_lcore(n)} , the token generation rate of the token bucket rate _lcore(i) = speed _{node(i)_lcore(n)} , and the number of tokens in the initialization token bucket token _lcore(i) = speed _{node(i)_lcore(n)} .

In the above steps, the examples in the above steps are continued to be used for explanation. The lcore(n) thread creates or adjusts a token bucket for the traffic marked with TCP-SYN every second, the depth of the bucket size _lcore(i) = speed _{node(i)_lcore(n)} , the bucket token generation rate rate _lcore(i) = speed _{node(i)_lcore(n)} , the token number of the token bucket is initialized token _lcore(i) = speed _{node(i)_lcore(n)} , and the length of each TCP-SYN message passing through the bucket length _pkt is generated according to the token bucket algorithm. The number of tokens rate _lcore(i) is generated per second, and the bucket depth does not exceed size _lcore(i) . If the length of the token in the bucket is greater than length _pkt , it passes and token = token-length _pkt . If length _pkt is greater than the token length, the message is discarded. The lcore(n) thread can limit the traffic speed of the queue according to the dynamically adjusted speed limit quota every second, so that the TCP-SYN traffic configured by the cluster reaches the cluster speed limit result, as shown in Figure 3. Each time a message length passes, the total length of the message that can pass is obtained by subtracting the number of tokens in the bucket from the message length. The above token indicates that the number of tokens in the current bucket is a dynamic variable, and token _lcore(i) indicates the initial value of the token bucket when it is initialized.

When a node service exits abnormally or a new node is added, the node controller in the forwarding process can still make dynamic adjustments based on the multicast messages received within the periodic time.

In summary, this application adopts the flow control method of the token bucket algorithm, combined with the communication and coordination mechanism within the cluster, through the data interaction between the cluster manager and the node controller, this technology can monitor the flow of each node and queue in real time, and dynamically generate and allocate tokens according to the shared flow rate limit threshold within the cluster. At the same time, this application can also adaptively adjust the flow rate limit strategy according to the status of the nodes and queues in the cluster to adapt to different load conditions and demand changes.

In a third aspect, the present application proposes, in one embodiment, an electronic device, comprising: a memory and one or more processors.

One or more application programs are stored in the memory, and the one or more application programs are suitable for being executed by the one or more processors to implement the method described in the first aspect.

The electronic device includes: a processor and a memory, wherein the processor and the memory are connected, for example, via a bus.

The structure of the electronic device does not constitute a limitation on the embodiments of the present application.

The processor may be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA or other programmable logic device, a transistor logic device, a hardware component or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of this application. The processor may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.

The bus may include a path to transmit information between the above components. The bus may be a PCI bus or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc.

The memory can be a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, or an EEPROM, CD-ROM or other optical disk storage, optical disk storage (including compressed optical disk, laser disk, optical disk, digital versatile disk, Blu-ray disk, etc.), magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to this.

In a fourth aspect, the present application proposes a computer-readable storage medium having a computer program stored thereon, wherein the computer program can be loaded by a processor and execute the method described in the first aspect.

Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims (10)

A rate limiting method for a cluster shared traffic threshold, comprising: S1, the cluster manager sends the cluster speed limit threshold of the specified data flow to each node in the network cluster, and adds the forwarding process of each node to the multicast to share traffic information between nodes, and the forwarding process in the node obtains the cluster speed limit threshold; S2: The node controller counts the traffic of the specified data flow flowing into the network card per second and marks it. It sends the node unique identifier and the network card traffic received per second through multicast to share with other nodes in the cluster. It calculates the node speed limit quota threshold of the node based on the multicast message received from the cluster. S3, the bandwidth allocation scheduler allocates the core speed limit quota threshold of each node according to the bandwidth ratio of each service core processing queue traffic in the node forwarding process to the total traffic of the node; S4: The token bucket manager uses the token bucket algorithm to periodically and dynamically adjust the real-time data of the proportion of traffic processed by each queue in the node to the total bandwidth of the node, obtain the proportion of each queue in the cluster's speed-limited bandwidth, and limit the speed of the traffic passing through the queue. The method according to claim 1, wherein in step S2, calculating the node rate limit quota threshold of the node according to the multicast message received from the cluster includes: When a node receives a multicast message sent by other nodes, it compares the node unique identifier in the multicast message with its own node unique identifier. If the comparison result is equal, the multicast message is ignored. If the comparison result is unequal, the multicast message is recorded and the node speed limit quota threshold of the node is calculated. According to the method of claim 2, the calculation formula of the node speed limit quota threshold is as follows:
speed _nodei ＝speed _cluster *traffic _nodei /∑ traffic _nodei Where speed _cluster is the cluster speed limit threshold of the specified data flow, traffic _nodei is the traffic volume of the specified data flow flowing into the node network card of node i within one second, i is the node number, ranging from 1 to N, and N is the number of cluster nodes. According to the method of claim 1, in step S3, the calculation formula of the core speed limit quota threshold is as follows:
speed _{node(i)_lcore(n)} =speed _node(i) *traffic _{node(i)_lcore(n)} /∑traffic _{node(i)_lcore(n)} Among them, speed _{node(i)_lcore(n)} is the traffic volume of the specified data flow flowing from node i to service core n within this second, i is the node number, ranging from 1 to N, N is the maximum number of nodes in the cluster, n is the service core number, ranging from 1 to M, M is the maximum number of service cores. A speed limiting device for cluster shared traffic threshold, the device comprising a cluster manager and multiple nodes located in the same network cluster; The cluster manager is connected to multiple nodes respectively, and is used to coordinate and manage the entire multi-node cluster, and set, adjust and issue cluster speed limit thresholds according to speed limit requirements; A node controller, which is arranged on each of the plurality of nodes and serves as a control main thread of the node forwarding process, and is used to monitor the local bandwidth utilization and flow; The bandwidth allocation scheduler is set in the business core of the service process of each node to ensure fair allocation and utilization efficiency of bandwidth resources; The token bucket manager is set on the thread bound to each queue served by each node. It is used to limit the speed of the traffic passing through the queue using the token bucket algorithm to achieve speed limit for the entire cluster. The device according to claim 5, wherein the node controller counts the local traffic data flowing into the network card every second, and shares the local traffic data through multicast. The device according to claim 6, wherein sharing the local traffic data includes: the node controller sends the local traffic data to each node in the cluster via multicast, the node receives traffic data from other nodes in the cluster, and after excluding its own traffic data, calculates the node's own node speed limit quota threshold based on the traffic ratio. According to the method of claim 5, ensuring fair allocation and utilization efficiency of bandwidth resources specifically includes: the bandwidth allocation scheduler determines the node speed limit quota threshold according to the traffic bandwidth ratio of the nodes in the cluster, and allocates the core speed limit quota threshold of each node according to the bandwidth ratio of each service core processing queue traffic in the node forwarding process to the total traffic of the node. According to the method of claim 5, the use of a token bucket algorithm to limit the speed of traffic passing through a queue specifically includes: the token bucket manager periodically and dynamically adjusts the proportion of each queue in the cluster speed-limited bandwidth based on real-time data of the proportion of traffic processed by each queue in the node to the total bandwidth of the node, so as to obtain the proportion of each queue in the cluster speed-limited bandwidth. An electronic device comprises: a memory unit and a processor unit, wherein the memory unit stores a computer program, and wherein the processor unit implements the method according to any one of claims 1 to 4 when executing the program.