Method and device for displaying congestion notification mark and electronic equipment

By allocating a separate queue to each active traffic flow and comparing cache usage with dynamic thresholds in real time, the problem of traditional explicit congestion notification algorithms failing to identify congested traffic is solved, thereby improving network throughput and reducing communication latency.

CN119052174AActive Publication Date: 2024-11-29TSINGHUA UNIVERSITY

Patent Information

Application Number
CN202410943105.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-15
Publication Date
2024-11-29
Estimated Expiration
2044-07-15

AI Technical Summary

Technical Problem

Traditional explicit congestion notification labeling algorithms cannot distinguish between traffic that is actually causing congestion and innocent, non-congested traffic, leading to incorrect labeling and restrictions, which affects network performance.

Method used

By acquiring multiple active traffic streams, counting their numbers, and assigning a separate queue to each stream, a dynamic marking threshold is determined based on the queue length and number. The cache usage is compared with the threshold in real time, and only traffic with cache usage exceeding the threshold is marked with a congestion notification.

Benefits of technology

It enables accurate identification and labeling of traffic that truly causes congestion, avoids mislabeling of innocent traffic, improves network throughput performance, and reduces communication latency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119052174A_ABST
    Figure CN119052174A_ABST
Patent Text Reader

Abstract

The invention discloses a congestion notification mark display method and device and electronic equipment, and relates to the technical field of network congestion control, and the method comprises the steps: obtaining a plurality of pieces of active traffic, and carrying out the statistics of the number of the plurality of pieces of active traffic, the active traffic being composed of one or more data packets which are queued in a switch at the current moment and have the same quintuple information; distributing a corresponding queue for each active flow according to the flow type; determining the cache occupation amount of each active flow according to the queue length of each active flow, and determining a dynamic mark threshold value of each active flow according to the number of the active flows; and comparing the cache occupation amount of each piece of active flow with a corresponding dynamic marking threshold value, and performing display congestion notification marking processing on the active flow of which the cache occupation amount is greater than the corresponding dynamic marking threshold value. According to the embodiment of the invention, the traffic which really causes congestion can be identified more accurately when congestion occurs, and innocent non-congestion traffic is effectively prevented from being wrongly marked and limited.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of network congestion control, and in particular to a method, device and electronic device for displaying congestion notification marking. Background Art

[0002] In a data center environment, network performance is critical for various applications and services. Different applications and services have different requirements for the network. For example, web content search and distributed cache services require lower communication latency, while cloud storage and parallel computing require higher network bandwidth. In order to meet these requirements at the same time, experts have proposed a variety of congestion control algorithms and congestion signal setting methods, among which Explicit Congestion Notification (ECN) is a congestion signal mechanism applied to the network layer.

[0003] The ECN mechanism is usually deployed on switches in the network. In the traditional explicit congestion notification marking algorithm, when congestion occurs, all data packets from the same port or the same queue will be marked to notify the sender to reduce the speed, avoid network overload and reduce transmission delay.

[0004] However, this traditional explicit congestion notification marking algorithm cannot distinguish the traffic that actually causes congestion from innocent non-congested traffic when congestion occurs, resulting in incorrect marking and restriction. For example, a low-priority or low-bandwidth demand flow may be incorrectly marked because it shares the same queue with other high-bandwidth flows, resulting in an unnecessary reduction in its sending rate, thus affecting network performance. Summary of the invention

[0005] In view of this, the present application provides a method, device and electronic device for displaying congestion notification marking, the main purpose of which is to solve the problem that the current traditional display congestion notification marking algorithm cannot identify the traffic that actually causes congestion and innocent non-congested traffic when congestion occurs, resulting in incorrect marking and restriction.

[0006] According to a first aspect of the present application, a method for displaying a congestion notification mark is provided, comprising:

[0007] Acquire multiple active flows and count the number of the multiple active flows, wherein the active flows are composed of one or more data packets with the same five-tuple information queued in the switch at the current moment;

[0008] Allocate a corresponding queue for each active flow according to the flow type;

[0009] Determining a cache occupancy of each of the active flows according to a queue length of each of the active flows, and determining a dynamic marking threshold of each of the active flows according to the number of the active flows;

[0010] The buffer occupancy of each active flow is compared with the corresponding dynamic marking threshold, and an explicit congestion notification marking process is performed on the active flow whose buffer occupancy is greater than the corresponding dynamic marking threshold.

[0011] According to a second aspect of the present application, a device for displaying a congestion notification mark is provided, comprising:

[0012] An acquisition module, used to acquire multiple active flows and count the number of the multiple active flows, wherein the active flow is composed of one or more data packets with the same five-tuple information queued in the switch at the current moment;

[0013] An allocation module, configured to allocate a corresponding queue to each active flow according to the flow type;

[0014] A first determination module is used to determine the cache occupancy of each active flow according to the queue length of each active flow, and determine the dynamic marking threshold of each active flow according to the number of active flows;

[0015] The comparison module is used to compare the cache occupancy of each active flow with the corresponding dynamic marking threshold, and perform a congestion notification marking process on the active flow whose cache occupancy is greater than the corresponding dynamic marking threshold.

[0016] According to a third aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method for displaying a congestion notification mark according to the first aspect is implemented.

[0017] According to a fourth aspect of the present application, an electronic device is provided, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the method for displaying a congestion notification mark as described in the first aspect when executing the computer program.

[0018] According to a fifth aspect of the present application, the present disclosure provides a computer program product, including: a computer program, which, when executed by a processor, implements the method for displaying a congestion notification mark according to the first aspect.

[0019] By means of the above technical scheme, a method, device and electronic device for displaying congestion notification marking provided by the present application can obtain multiple active flows and count the number of multiple active flows, wherein the active flow is composed of one or more data packets with the same five-tuple information queued in the switch at the current moment; a corresponding queue is allocated to each active flow according to the flow type; the cache occupancy of each active flow is determined according to the queue length of each active flow, and the dynamic marking threshold of each active flow is determined according to the number of active flows; the cache occupancy of each active flow is compared with the corresponding dynamic marking threshold, and the active flow whose cache occupancy is greater than the corresponding dynamic marking threshold is processed with a display congestion notification marking. For the embodiments of the present disclosure, by allocating a separate queue to each active flow, each active flow can be independently served and processed, avoiding the problem of mutual interference between different flows. At the same time, it also ensures that subsequent services can be processed differently according to the flow level, thereby improving service quality and network efficiency. By comparing the cache occupancy of each active flow with the corresponding dynamic marking threshold in real time, it is possible to more accurately identify the flow that actually causes congestion when congestion occurs, and to display congestion notification markings, effectively preventing innocent non-congested traffic from being incorrectly marked and restricted. This fine-grained congestion control helps improve network throughput performance and reduce communication latency.

[0020] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0022] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0023] Figure 1 A flowchart of a method for displaying a congestion notification mark provided by an embodiment of the present disclosure;

[0024] Figure 2 An overall working principle diagram of a method for displaying congestion notification marking provided by an embodiment of the present disclosure;

[0025] Figure 3A queue allocation diagram for enqueuing a data packet provided by an embodiment of the present disclosure;

[0026] Figure 4 A schematic diagram of the structure of a device for displaying a congestion notification mark provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0027] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to aid understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description. It should be noted that the embodiments of the present disclosure and the features therein may be combined with each other without conflict.

[0028] The following describes a method, device, and electronic device for displaying a congestion notification mark according to an embodiment of the present disclosure with reference to the accompanying drawings.

[0029] In order to more accurately identify the traffic that actually causes congestion when congestion occurs, and avoid innocent non-congested traffic from being wrongly marked and restricted. This embodiment provides a method for displaying congestion notification marking, such as Figure 1 As shown, the method includes:

[0030] Step 101: Acquire multiple active flows and count the number of the multiple active flows.

[0031] Among them, active traffic is composed of one or more packets with the same five-tuple information queued in the switch at the current moment. By tracing only active traffic, the traffic space that needs to be saved can be greatly reduced, because only those flows that are using network resources need to be tracked and processed, which helps to improve the efficiency and performance of the system.

[0032] The number of active flows may be counted, which may include:

[0033] Check that the traffic currently queued in the switch is active traffic;

[0034] Calculate the five-tuple hash value of each active flow according to the five-tuple information of each active flow;

[0035] According to the uniqueness of the five-tuple hash value, the number of active flows is counted.

[0036] For the embodiments of the present disclosure, specifically, real-time statistics of active traffic can be approximately achieved on the data plane.

[0037] First, a time period T may be initialized. In each time period T, a flow counter n may be maintained, wherein the flow counter may be used to count the number of active flows in the current period.

[0038] The start of a flow can be defined as the first appearance of a five-tuple packet in the switch in each time period. This means that every time a new five-tuple packet arrives at the switch, it is regarded as a new flow start.

[0039] When a data packet in each active flow arrives, the five-tuple hash value of the current data packet is calculated and stored in the register. The five-tuple hash value of the current data packet can be obtained by hashing the source IP address, source sending port, destination IP address, destination sending port and protocol type (i.e., five-tuple information).

[0040] If the current five-tuple hash value has never appeared, it means that this is the first appearance of the five-tuple data packet in the current time period, so the counter value n in the period will be increased by one, indicating that a new flow has started.

[0041] After each cycle ends, the registers and counter n responsible for storing the hash value can be reset and cleared to start a new cycle statistics.

[0042] For the disclosed embodiment, the present application designs two processing units to solve the coarse-grained problem of early display congestion notification marking, so as to adapt to different network traffic patterns, improve the network throughput performance, and reduce the network communication delay. The first unit is a queue mapping unit, which only traces the status of active flows and allocates a separate queue to each active flow as much as possible to ensure differentiated services at the subsequent flow level. The second unit is a selective marking unit, which reconstructs the marking criteria of the traditional display congestion notification based on the fairness of active flows, compares the cache occupancy of each active flow with the dynamic marking threshold separately, and accurately identifies the congested flow that needs to be marked. This method can accurately identify the traffic that actually causes congestion and display congestion notification marking when congestion occurs by combining the queue mapping unit and the selective marking unit, while protecting innocent non-congested traffic. Both of the above units can be implemented on the data plane, which is very suitable for deployment in ultra-high-speed data center networks, improving network throughput performance and reducing network communication delay. Among them, the specific implementation process can be referred to steps 102 to 104.

[0043] Step 102: Allocate a corresponding queue to each active flow according to the flow type.

[0044] The traffic type may include new active traffic and existing active traffic; the queue may be a first-in-first-out queue (FIFO), which may include an empty queue, an allocated queue, and a high-priority queue.

[0045] New active traffic may be traffic that is identified or detected by the system for the first time and meets the definition of active traffic (ie, consists of one or more data packets with the same quintuple information).

[0046] Existing active traffic can be traffic that has been identified and processed by the system and meets the definition of active traffic (i.e., composed of one or more data packets with the same five-tuple information). These traffic may already exist in a data structure of the system, such as a hash table or a queue.

[0047] An empty queue may be a queue to which no active traffic is currently allocated. An empty queue may be used to store newly arrived active traffic that has not yet been processed.

[0048] The allocated queue may be a queue to which active traffic has been allocated.

[0049] A high priority queue may be a queue with a higher processing priority in the queue system. A high priority queue may be used to store new active traffic that requires a faster response or higher quality of service. A high priority queue can ensure that important or urgent traffic is given priority when resources are limited.

[0050] In specific application scenarios, the emergence of new programmable switches has facilitated the implementation of differentiated congestion notification marking at the flow level. This application utilizes the advantages of multiple queues of programmable switches and combines the real-time traffic information of the data plane to have the ability to differentiate flows in complex traffic environments and differentially mark each flow.

[0051] Specifically, in order to achieve the flow-by-flow marking granularity, such as Figure 2 As shown, when a data packet of each active flow arrives at the switch, the queue mapping unit can check whether the data packet belongs to a new active flow. If it is a new active flow, an empty queue on the outbound port is allocated to it; if it is an existing active flow, it is placed in the allocated queue, so that each flow occupies a queue as much as possible, thereby ensuring that each flow can receive differentiated explicit congestion notification marking processing. The queue mapping unit can be implemented on the data plane of the programmable switch.

[0052] For the disclosed embodiment, different active flows can be distinguished by the source IP address, source sending port, destination IP address, destination sending port and protocol type (i.e., quintuple information) of the data packet. For data packets with the same quintuple information, it can be determined that they belong to the same active flow.

[0053] Specifically, when a data packet enters the switch, the switch can check whether the active flow of the data packet belongs to the new active flow according to the five-tuple information. If it is the new active flow, a new queue is allocated for the new data packet; if it is not the new active flow, the queue ID of the flow corresponding to the five-tuple information of the data packet is found and the data packet is allocated to the queue.

[0054] This application helps the system manage queue resources more effectively and avoid queue overload by tracking the buffer occupancy of each queue to determine which queues are empty. Once the last packet in a queue is dispatched, the state of the queue will be reset to empty. This management mechanism ensures the efficient use of queues and provides space for new traffic allocation.

[0055] For the embodiments of the present disclosure, correspondingly, the traffic type of the active traffic can be determined based on the five-tuple information of each active traffic, wherein the five-tuple information includes the source IP address, the source sending port, the destination IP address, the destination sending port and the protocol type.

[0056] Specifically, determining the traffic type of the active traffic according to the five-tuple information of each active traffic may include:

[0057] Check whether there is a five-tuple information record corresponding to the five-tuple information in the preset flow table, wherein the preset flow table contains the five-tuple information of the existing active flow;

[0058] If it does not exist, the traffic type of the active traffic is determined to be new active traffic;

[0059] If it exists, the flow type of the active flow is determined to be the existing active flow.

[0060] For the disclosed embodiments, although a separate queue has been assigned to each active flow, in rare cases, the number of burst active flows may exceed the number of available queues provided by the switch. To solve this problem, the present application reserves a queue that is assigned a high priority (i.e., a high priority queue). When new active traffic arrives and all other queues are occupied, it will enter the high priority queue. Packets in the high priority queue will be dequeued before packets in other queues. This means that even when all other queues are full, packets in the high priority queue will be processed first. When a packet of a flow belonging to a high priority queue enters the switch, the system checks in real time whether there is a new empty queue. Once a new empty queue appears, the system can transfer the packet to the new empty queue to ensure that the packets in the high priority queue can be processed as soon as possible (i.e., an empty queue that is found in real time through the principle of "queuing while detecting").

[0061] In order to prevent large flows from excessively occupying high-priority queue resources and affecting delay-sensitive small flows, the present application limits the number of bytes of each new active flow entering the high-priority queue (that is, the number of bytes of new active flows passing through the high-priority queue is less than or equal to a preset fixed threshold). This limitation helps to balance the processing of different flows and ensure that small flows can be served.

[0062] By reserving a high-priority queue, the present application provides an opportunity for newly arriving active traffic to enter an empty queue, thereby avoiding the problem of long-term use of the high-priority queue causing starvation of other queues, and also preventing large traffic from occupying the high-priority queue for a long time.

[0063] This application builds a complete logic for packet enqueue by allocating a separate queue for each active flow and reserving a high-priority queue, ensuring that even when queue resources are tight, the system can effectively manage and allocate queues to meet the needs of different flows, such as Figure 3 shown.

[0064] Accordingly, a corresponding queue is allocated to each active flow according to the traffic type, which may include:

[0065] If the traffic type of the active traffic is new active traffic and there is an empty queue in the queue, the empty queue is allocated to the new active traffic;

[0066] If the traffic type of the active traffic is new active traffic, there is no empty queue in the queue, and the number of bytes passed by the new active traffic in the high priority queue is less than or equal to the preset fixed threshold, the high priority queue is allocated to the new active traffic;

[0067] If the traffic type of the active traffic is existing active traffic, the corresponding allocated queue is allocated to the existing active traffic.

[0068] Step 103: Determine the buffer occupancy of each active flow according to the queue length of each active flow, and determine the dynamic marking threshold of each active flow according to the number of active flows.

[0069] The cache occupancy may be the amount of space occupied in a cache (such as a cache of a network device, a buffer of an operating system, etc.) for processing active traffic. The cache occupancy may be determined according to the traffic size and transmission speed of the active traffic.

[0070] The dynamic marking threshold can be a limit value used in network management and traffic control to decide whether to perform certain operations (such as marking, packet loss, etc.) on active traffic. This threshold changes dynamically and can be adjusted according to factors such as the current network status, the amount of active traffic, and system resources.

[0071] The dynamic marking threshold can be used to distinguish congested flows from non-congested flows, so as to decide whether to perform explicit congestion notification marking on packets according to the dynamic marking threshold.

[0072] For the disclosed embodiments, modern programmable switches, such as Tofino 1 and Tofino 2, are equipped with multi-queue capabilities. For example, Tofino 1 has 32 queues per egress port, and Tofino 2 has 128 queues per egress port. This means that in general, each egress port has enough queues to allocate a separate queue for each active flow.

[0073] Since each active flow can be assigned to a separate queue, the present application can use the queue length as the cache occupancy of the active flow (i.e., determine the cache occupancy of each active flow according to the queue length of each active flow), so that the selective marking unit can identify the rationality of congested flows and non-congested flows by analyzing the queue length, so as to determine whether it is necessary to display congestion notification marking for the data packet based on the queue length.

[0074] In the present application, a congested flow may be a flow whose bandwidth exceeds the fair share, which is higher than the bandwidth of a non-congested flow. In the present application, each active flow enters a separate FIFO queue, and a high-bandwidth flow will naturally generate more packets in the same time, resulting in a longer queue. Therefore, under flow-level management, the more cache occupancy of an active flow, the more likely it is that the active flow will become a congested flow, thereby ensuring the rationality of identifying congested flows by queue length.

[0075] The queue length can be obtained directly from the interface provided by the programmable switch, without the need to manually update the cache occupancy of each active flow when each data packet enters and leaves the switch. In this way, the present application avoids performing a large number of calculations and update operations during each data packet processing, thereby reducing computing overhead and improving processing efficiency.

[0076] For the disclosed embodiment, the dynamic marking threshold can be determined based on a conventional setting method to facilitate implementation in a switch. In addition, the dynamic marking threshold can be determined by the number of active flows to ensure that each active flow can obtain fair bandwidth allocation.

[0077] To simplify deployment, the present application may set the same dynamic marking threshold for each queue of active traffic.

[0078] With the idea of ​​flow-by-flow marking, network management is no longer based on ports and queues, but processes flow rates individually from the perspective of flows to ensure that each active flow receives fair bandwidth allocation.

[0079] In the scenario where N active flows share one egress port, each active flow should obtain a fair share of bandwidth C / N. In order to ensure that each active flow can obtain fair bandwidth, this application reconstructs the calculation method of the marking threshold. The new marking threshold K p It can be calculated according to the following formula:

[0080] K p =λ*RTT*C / N flow

[0081] In the formula, K p It can represent the dynamic marking threshold of each active flow, λ can represent the inherent coefficient of the traditional explicit congestion notification marking calculation method, which is usually 1, RTT can represent the communication delay from the sender to the receiver of each active flow, C can represent the sending rate of each active flow on the outbound port, N flow Can represent the amount of active flow.

[0082] As flows start and end, the number of active flows N flow It will also change dynamically in real time, indicating that the marking threshold K p It can be adjusted dynamically according to the current traffic situation to ensure that each flow can obtain fair outbound port bandwidth.

[0083] New marking threshold K p Indicates the congestion boundary when the bandwidth of each flow reaches its fair share. This means that only when the bandwidth of an active flow reaches the marking threshold will it be marked with congestion notification to avoid excessive bandwidth consumption.

[0084] Accordingly, the dynamic marking threshold of each active flow is determined according to the number of active flows, including:

[0085] Obtain the intrinsic coefficient, the communication delay of each active flow from the sending end to the receiving end, and the sending rate of each active flow at the egress port;

[0086] Substitute the inherent coefficient, communication delay, sending rate and the number of active flows into the preset dynamic marking threshold calculation formula to obtain the dynamic marking threshold of each active flow;

[0087] The preset dynamic marking threshold calculation formula is:

[0088] K p =λ*RTT*C / N flow

[0089] In the formula, K p represents the dynamic marking threshold of each active flow, λ represents the inherent coefficient, RTT represents the communication delay from the sender to the receiver of each active flow, C represents the sending rate of each active flow on the outbound port, and Nflow Indicates the number of active flows.

[0090] Step 104: Compare the buffer occupancy of each active flow with the corresponding dynamic marking threshold, and perform explicit congestion notification marking processing on the active flow whose buffer occupancy is greater than the corresponding dynamic marking threshold.

[0091] Among them, Explicit Congestion Notification (ECN) can be a congestion signal mechanism applied to the network layer. This mechanism is usually deployed on switches in the network. When congestion occurs, it actively marks data packets to inform the sender to reduce the speed, avoid network overload and reduce transmission delay. The ECN algorithm mainly determines whether to mark the data packet based on the queue length. When the queue length exceeds the preset threshold, the corresponding ECN bit in the TOS (Type of Service) field of the subsequent data packet will be set to 1. After the sender receives the congestion signal, it will reduce the sending speed until the queue length drops below the threshold. The ECN algorithm is widely used in network devices, such as routers, switches, etc., to manage the marking and congestion control of data packets. It plays a key role in the congestion control mechanism in TCP / IP networks, helping network devices to dynamically adjust the sending speed according to the degree of congestion to improve network performance and fairness.

[0092] For the disclosed embodiments, the application mainly utilizes the queue mapping unit to ensure that each active flow enters a separate queue as much as possible, thereby isolating the cache occupancy of each active flow from other flows. Then, when the data packet of the active flow is dequeued, the selective marking unit can decide whether to differentially display congestion notification markings for the data packets based on the buffer occupancy of each active flow. This method ensures that the display congestion notification marking of each active flow is independent, only related to the queue length of the flow, and is not affected by other flows.

[0093] Specifically, the cache occupancy of each active flow can be compared with the corresponding dynamic marking threshold. If the cache occupancy exceeds the dynamic marking threshold, the data packet is marked; otherwise, it is not marked, so as to realize the difference display of congestion notification marking at the flow level, ensuring that only the flow that actually causes congestion will be marked, thereby protecting other flows from being affected, such as Figure 2 As shown, the selective marking unit can be implemented on the data plane of the programmable switch; the congested flow can be the flow whose bandwidth exceeds the fair share.

[0094] Although the application reacts quickly to congested traffic, it is necessary to avoid over-suppressing it. When the rate of the congested flow successfully drops to the fair share or other flows reach the required bandwidth, the application no longer displays the congestion notification mark for it to avoid excessive restriction of high-demand traffic. Therefore, it is crucial to set the threshold of the occupied buffer as the boundary to distinguish congested flows from non-congested flows. The setting of the marking threshold helps to accurately control high-demand traffic and protect low-rate traffic.

[0095] For the disclosed embodiments, the fine-grained marking of the flow level of the ECN algorithm in the switch is implemented through the queue mapping unit and the selective marking unit. Compared with the previous traditional methods, the current network traffic situation is obtained in the ultra-high-speed network. When congestion occurs, the flow-level difference display congestion notification marking algorithm is implemented by utilizing the multi-queue and data plane real-time traffic information of the programmable switch, and the traffic that actually causes congestion and needs to be marked is accurately identified. At the same time, the remaining innocent traffic is protected from interference, the communication delay of delay-sensitive applications is reduced, the network performance is better improved, and the fairness of bandwidth is ensured. This technology can effectively enhance the automatic adaptability of the ultra-high-speed network to different traffic patterns, thereby improving the communication capability of the network, and is very suitable for deployment and application in ultra-high-speed network scenarios with complex traffic processing requirements.

[0096] In summary, according to a method for displaying congestion notification marking disclosed in the present invention, multiple active flows can be obtained and the number of multiple active flows can be counted, wherein the active flow is composed of one or more data packets with the same five-tuple information queued in the switch at the current moment; a corresponding queue is allocated to each active flow according to the flow type; the cache occupancy of each active flow is determined according to the queue length of each active flow, and the dynamic marking threshold of each active flow is determined according to the number of active flows; the cache occupancy of each active flow is compared with the corresponding dynamic marking threshold, and the active flow whose cache occupancy is greater than the corresponding dynamic marking threshold is subjected to display congestion notification marking processing. For the embodiments of the present disclosure, by allocating a separate queue to each active flow, each active flow can be independently served and processed, avoiding the problem of mutual interference between different flows. At the same time, it also ensures that subsequent services can be processed differentially according to the flow level, thereby improving service quality and network efficiency. By comparing the cache occupancy of each active flow with the corresponding dynamic marking threshold in real time, it is possible to more accurately identify the flow that actually causes congestion when congestion occurs, and to display congestion notification markings, effectively preventing innocent non-congested traffic from being incorrectly marked and restricted. This fine-grained congestion control helps improve network throughput performance and reduce communication latency.

[0097] Based on the above Figure 1 The specific implementation of the method shown in the embodiment provides a device for displaying a congestion notification mark, such as Figure 4As shown, the device includes: an acquisition module 31, an allocation module 32, a first determination module 33, and a comparison module 34;

[0098] An acquisition module 31 is used to acquire multiple active flows and count the number of the multiple active flows, wherein the active flow is composed of one or more data packets with the same five-tuple information queued in the switch at the current moment;

[0099] An allocation module 32, configured to allocate a corresponding queue to each active flow according to the flow type;

[0100] A first determination module 33 is used to determine the cache occupancy of each active flow according to the queue length of each active flow, and to determine the dynamic marking threshold of each active flow according to the number of active flows;

[0101] The comparison module 34 is used to compare the buffer occupancy of each active flow with the corresponding dynamic marking threshold, and perform a congestion notification marking process on the active flow whose buffer occupancy is greater than the corresponding dynamic marking threshold.

[0102] In a specific application scenario, the traffic type includes new active traffic and existing active traffic; the queue includes an empty queue, an allocated queue, and a high priority queue, and the allocation module 32 can be used to allocate the empty queue to the new active traffic if the traffic type of the active traffic is the new active traffic and there is an empty queue in the queue;

[0103] If the traffic type of the active traffic is the new active traffic, there is no empty queue in the queue, and the number of bytes passed by the new active traffic in the high priority queue is less than or equal to a preset fixed threshold, the high priority queue is allocated to the new active traffic;

[0104] If the traffic type of the active traffic is the existing active traffic, the corresponding allocated queue is allocated to the existing active traffic.

[0105] In a specific application scenario, the device further includes: a second determination module 35;

[0106] The second determination module 35 is used to determine the traffic type of the active traffic according to the five-tuple information of each active traffic, wherein the five-tuple information includes a source IP address, a source sending port, a destination IP address, a destination sending port and a protocol type.

[0107] In a specific application scenario, the second determination module 35 may be used to query whether there is a five-tuple information record corresponding to the five-tuple information in a preset flow table, wherein the preset flow table contains five-tuple information of existing active traffic;

[0108] If not, determining the flow type of the active flow to be new active flow;

[0109] If it exists, it is determined that the flow type of the active flow is the existing active flow.

[0110] In a specific application scenario, the acquisition module 31 can be used to query whether the traffic queued in the switch at the current moment is active traffic;

[0111] Calculate a five-tuple hash value of each active flow according to the five-tuple information of each active flow;

[0112] According to the uniqueness of the quintuple hash value, the amount of the active traffic is counted.

[0113] In a specific application scenario, the first determination module 33 may be used to obtain an inherent coefficient, a communication delay of each active flow from a sending end to a receiving end, and a sending rate of each active flow at an egress port;

[0114] Substituting the inherent coefficient, the communication delay, the sending rate and the number of active flows into a preset dynamic marking threshold calculation formula to obtain a dynamic marking threshold for each active flow;

[0115] The preset dynamic marking threshold calculation formula is:

[0116] K p =λ*RTT*C / N flow

[0117] In the formula, K p represents the dynamic marking threshold of each active flow, λ represents the inherent coefficient, RTT represents the communication delay from the sending end to the receiving end of each active flow, C represents the sending rate of each active flow on the outbound port, N flow Indicates the number of active flows.

[0118] It should be noted that for other corresponding descriptions of the functional units involved in the device for displaying a congestion notification mark provided in this embodiment, reference can be made to Figure 1 The corresponding description of the method in will not be repeated here.

[0119] Based on the above Figure 1 The method shown in the present disclosure, accordingly, also provides a computer-readable storage medium having a computer program stored thereon, which implements the above-mentioned Figure 1 The method shown.

[0120] Based on this understanding, the technical solution of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of various implementation scenarios of the present disclosure.

[0121] Based on the above Figure 1 The method shown, and Figure 4 In order to achieve the above-mentioned purpose, the embodiment of the present disclosure also provides an electronic device that can be configured on the vehicle (such as an electric vehicle) side, and the device includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to achieve the above-mentioned Figure 1 The method shown.

[0122] Optionally, the above-mentioned physical device may also include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, etc. The user interface may include a display, an input unit such as a keyboard, etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface), etc.

[0123] Those skilled in the art will appreciate that the above-mentioned physical device structure provided by the present disclosure does not constitute a limitation on the physical device, and may include more or fewer components, or a combination of certain components, or different arrangements of components.

[0124] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the above-mentioned physical device, and supports the operation of the information processing program and other software and / or programs. The network communication module is used to realize the communication between the components inside the storage medium, and the communication with other hardware and software in the information processing physical device.

[0125] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present disclosure can be implemented by means of software plus necessary general hardware platforms, or by hardware. Compared with the prior art, the display congestion notification marking method, device and electronic device provided by the present disclosure obtain multiple active flows and count the number of multiple active flows, wherein the active flow is composed of one or more data packets with the same five-tuple information queued in the switch at the current moment; assign a corresponding queue to each active flow according to the flow type; determine the cache occupancy of each active flow according to the queue length of each active flow, and determine the dynamic marking threshold of each active flow according to the number of active flows; compare the cache occupancy of each active flow with the corresponding dynamic marking threshold, and perform display congestion notification marking processing on the active flow whose cache occupancy is greater than the corresponding dynamic marking threshold. For the embodiment of the present disclosure, by assigning a separate queue to each active flow, each active flow can be independently served and processed, avoiding the problem of mutual interference between different flows. At the same time, it also ensures that subsequent services can be processed differently according to the flow level, improving service quality and network efficiency. By comparing the cache occupancy of each active flow with the corresponding dynamic marking threshold in real time, it is possible to more accurately identify the flow that actually causes congestion when congestion occurs, and to display congestion notification markings, effectively preventing innocent non-congested traffic from being incorrectly marked and restricted. This fine-grained congestion control helps improve network throughput performance and reduce communication latency.

[0126] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprising a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0127] The above is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments herein, but will conform to the widest scope consistent with the principles and novel features applied for herein.

Claims

1. A method for displaying a congestion notification mark, characterized in that: The method comprises: Acquire multiple active flows and count the number of the multiple active flows, wherein the active flows are composed of one or more data packets with the same five-tuple information queued in the switch at the current moment; Allocate a corresponding queue for each active flow according to the flow type; Determining a cache occupancy of each of the active flows according to a queue length of each of the active flows, and determining a dynamic marking threshold of each of the active flows according to the number of the active flows; The buffer occupancy of each active flow is compared with the corresponding dynamic marking threshold, and an explicit congestion notification marking process is performed on the active flow whose buffer occupancy is greater than the corresponding dynamic marking threshold.

2. The method according to claim 1, characterized in that The traffic type includes new active traffic and existing active traffic; the queue includes an empty queue, an allocated queue and a high priority queue; The allocating a corresponding queue to each active flow according to the flow type includes: If the traffic type of the active traffic is the new active traffic and there is an empty queue in the queue, allocating the empty queue to the new active traffic; If the traffic type of the active traffic is the new active traffic, there is no empty queue in the queue, and the number of bytes passed by the new active traffic in the high priority queue is less than or equal to a preset fixed threshold, the high priority queue is allocated to the new active traffic; If the traffic type of the active traffic is the existing active traffic, the corresponding allocated queue is allocated to the existing active traffic.

3. The method according to claim 2, characterized in that The method further comprises: The traffic type of the active traffic is determined according to the five-tuple information of each active traffic, wherein the five-tuple information includes a source IP address, a source sending port, a destination IP address, a destination sending port, and a protocol type.

4. The method according to claim 3, characterized in that The determining the traffic type of the active traffic according to the five-tuple information of each active traffic includes: Querying whether there is a five-tuple information record corresponding to the five-tuple information in a preset flow table, wherein the preset flow table contains five-tuple information of existing active traffic; If not, determining the flow type of the active flow to be new active flow; If it exists, it is determined that the flow type of the active flow is the existing active flow.

5. The method according to claim 1, characterized in that The counting of the number of the plurality of active flows includes: Check that the traffic currently queued in the switch is active traffic; Calculate a five-tuple hash value of each active flow according to the five-tuple information of each active flow; According to the uniqueness of the quintuple hash value, the amount of the active traffic is counted.

6. The method according to claim 1, characterized in that The step of determining the dynamic marking threshold of each active flow according to the number of active flows includes: Obtaining an inherent coefficient, a communication delay of each active flow from a sending end to a receiving end, and a sending rate of each active flow at an egress port; Substituting the inherent coefficient, the communication delay, the sending rate and the number of active flows into a preset dynamic marking threshold calculation formula to obtain a dynamic marking threshold for each active flow; The preset dynamic marking threshold calculation formula is: K p =λ*RTT*C / N flow In the formula, K p represents the dynamic marking threshold of each active flow, λ represents the inherent coefficient, RTT represents the communication delay from the sending end to the receiving end of each active flow, C represents the sending rate of each active flow on the outbound port, N flow Indicates the number of active flows.

7. A device for displaying a congestion notification mark, characterized in that: The device comprises: An acquisition module, used to acquire multiple active flows and count the number of the multiple active flows, wherein the active flow is composed of one or more data packets with the same five-tuple information queued in the switch at the current moment; An allocation module, configured to allocate a corresponding queue to each active flow according to the flow type; A first determination module is used to determine the cache occupancy of each active flow according to the queue length of each active flow, and determine the dynamic marking threshold of each active flow according to the number of active flows; The comparison module is used to compare the buffer occupancy of each active flow with the corresponding dynamic marking threshold, and perform a congestion notification marking process on the active flow whose buffer occupancy is greater than the corresponding dynamic marking threshold.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method described in claims 1 to 6 is implemented.

9. An electronic device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, characterized in that: When the processor executes the computer program, the method described in claims 1-6 is implemented.

10. A computer program product, characterized in that The invention comprises a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 6.

Citation Information

Patent Citations

  • Display congestion marking method and device combining enqueue marking with dequeue marking

    CN107070804A

  • Micro burst flow oriented congestion sensing and marking method in multi-queue data center environment

    CN110061927A

  • Congestion control method and device

    CN113507420A

  • Self-learning explicit mark congestion control method and system based on token bucket

    CN118055072A

  • Network active congestion control method and device, electronic equipment and storage medium

    CN118138530A

Cited By

  • Data management method and device, storage medium and computer equipment

    CN119945986A

  • Port-by-port flow control method suitable for BCube topology, medium and equipment

    CN119966899A

  • A per-port flow control method, medium and device suitable for BCube topology

    CN119966899B