A dynamic bandwidth-aware active queue management method and system

Abstract

The application provides a kind of active queue management method based on dynamic bandwidth perception, comprising: S1, real-time response data packet arrival, obtain the available bandwidth value of current network link and the length of data packet queue;S2, according to the available bandwidth value of current network link and the preset target forwarding delay, calculate the current congestion threshold;S3, compare the length of current data packet queue and the size of current congestion threshold, when the length of current data packet queue is greater than or equal to current congestion threshold, it is congestion state;When in congestion state, further judge whether the data stream to which the current arrived data packet belongs is aggressive flow, if it is aggressive flow, then according to the preset packet loss method, the current arrived data packet is randomly discarded.The application maps the changing link available bandwidth into dynamic threshold to manage queue, so that network device can efficiently utilize network bandwidth;Through dynamic identification of aggressive flow and targeted setting of packet loss strategy, network device can provide forwarding service more fairly.

Description

Technical Field

[0001] This invention relates to the field of data communication, specifically to the field of network equipment, and more specifically to a proactive queue management method based on dynamic bandwidth awareness. Background Technology

[0002] Queue management is a key technology for addressing network congestion and reducing transmission latency by managing packet queues. Traditional queue management schemes often employ tail-drop, discarding newly arriving packets only when the queue is full, which can easily lead to TCP global synchronization problems and thus degrade network performance. Active Queue Management (AQM) schemes, such as RED, CoDel, and PIE, randomly drop or mark packets when the queue is not full and send congestion signals in advance to notify the sender to pause transmission, thereby reducing average queue length and end-to-end latency. For example, RED calculates an exponentially weighted average of the queue length and actively drops packets with a certain probability when the length exceeds a preset threshold to prevent congestion. CoDel, on the other hand, directly uses packet queuing latency as a control indicator, only dropping packets when a sustained high latency is detected, thus achieving precise latency control. In addition, there are some improved AQM schemes, such as ARED, which optimizes forwarding performance by dynamically adjusting the packet drop probability, and WRED, which implements differentiated packet drop based on IP priority. All of the above AQM schemes have two characteristics: they determine whether congestion will occur based on a fixed threshold (queue length or queue forwarding delay); and they do not identify data streams, or they identify data streams and then apply a static packet loss strategy to those data streams.

[0003] However, existing AQM schemes exhibit significant limitations in 5G network equipment due to their fixed threshold settings and static packet loss strategies. Experiments show that in complex environments like 5G networks, characterized by drastic fluctuations in available bandwidth, multi-stream contention, and deep buffers, existing AQM schemes, while partially alleviating congestion, cannot effectively utilize bandwidth or guarantee data flow fairness. For example, while AQM schemes like RED and CoDel can reduce average queue length, their fixed threshold settings mean that when available network bandwidth drops sharply, congestion can quickly occur even if the queue length has not yet reached the set threshold. Conversely, when available network bandwidth increases, excessively low thresholds can lead to excessive packet loss, impacting network equipment throughput. Furthermore, when aggressive flows (service flows with extremely high traffic within a certain timeframe) exist, the lack of packet loss strategies for these flows allows queues to quickly fill up, resulting in higher tail latency for other service flows.

[0004] In summary, existing AQM solutions have two major problems when facing complex environments like 5G networks, such as drastic fluctuations in available bandwidth, multi-stream competition, and deep buffers: fixed threshold settings prevent network devices from efficiently utilizing the currently available network bandwidth; and static packet loss policies prevent network devices from providing fair forwarding services for different data streams.

[0005] It should be noted that the background information presented here is only for illustrating relevant information about the present invention to aid in understanding the technical solution of the present invention, and does not imply that the relevant information is necessarily prior art. The relevant information was submitted and disclosed together with the present invention, and should not be considered prior art unless there is evidence that the relevant information was disclosed before the filing date of the present invention. Summary of the Invention

[0006] Therefore, the purpose of this invention is to overcome the shortcomings of the prior art and provide an active queue management method based on dynamic bandwidth awareness.

[0007] The objective of this invention is achieved through the following technical solution:

[0008] According to a first aspect of the present invention, a dynamic bandwidth-aware active queue management method is provided for managing whether data packets arriving in real time in a network device are placed into a data packet queue, wherein the data packet queue is used to store data packets to be forwarded, and the method includes the following steps: S1. Real-time response to data packet arrival, obtaining the available bandwidth value of the current network link and the length of the current data packet queue. S2. Calculating the current congestion threshold according to the available bandwidth value of the current network link and a preset target forwarding delay according to a preset rule. S3. Comparing the length of the current data packet queue with the current congestion threshold to obtain the current state; a congestion state is defined as the length of the current data packet queue being greater than or equal to the current congestion threshold. In the congestion state, it is determined whether the data flow to which the currently arriving data packet belongs is an aggressive flow according to a preset method; if it is an aggressive flow, the currently arriving data packet is randomly discarded according to a preset packet dropping method.

[0009] This scheme achieves at least the following beneficial technical effects: 1. It can detect changes in link bandwidth in real time and correlate the dynamically changing link bandwidth with congestion thresholds, enabling network devices to utilize the fluctuating link bandwidth more efficiently and improving their actual throughput. 2. It defines aggressive flows and only causes random packet loss to aggressive flows, improving the fairness of network forwarding services.

[0010] Optionally, the available bandwidth value of the current network link can be obtained in step S1 by measuring the amount of transmitted data on the downlink path of the network device at a fixed sampling period, and then obtaining the available bandwidth value of the current network link as follows:

[0011]

[0012] in, This represents the available bandwidth value of the current network link. Represents a fixed sampling period. This represents the amount of data transmitted as measured.

[0013] Optionally, the preset rule in step S2 is:

[0014]

[0015] in, Represents the congestion threshold. This represents the preset target forwarding latency. This represents the available bandwidth value of the current network link. It represents the maximum transmission unit that is fixed in a network link.

[0016] Optionally, the preset method in step S3 includes: S31, obtaining the length of the data stream to which the currently arriving data packet belongs in the data packet queue. and the number of data streams currently in the packet queue S32, through Calculate the length percentage of the data stream to which the currently arriving data packet belongs in the data packet queue. ,in Represents the length of the current packet queue. S33, via Calculate the current flow percentage threshold ,in This represents the set tolerance error. S34, Comparison and The size is used to determine whether the data stream to which the currently arriving data packet belongs is an aggressive stream, where, if Then the data stream is an aggressive stream.

[0017] This solution can achieve at least the following beneficial technical effects: it provides an optional aggressive flow judgment method, in which... and The behavior changes constantly with the arrival of new data packets, so whether the currently forwarded data stream is considered an aggressive stream is also constantly changing, which provides better fairness.

[0018] Optionally, the preset packet loss method includes: T1, calculating the packet loss probability of the currently arriving data packet according to a preset packet loss probability rule; T2, generating a random number for the currently arriving data packet; T3, comparing the packet loss probability of the currently arriving data packet with the random number generated in step T2, wherein if the packet loss probability is greater than or equal to the random number, the arriving data packet is discarded.

[0019] This solution can achieve at least the following beneficial technical effects: it provides an optional packet loss method, adopts a probabilistic packet loss strategy to smooth queue changes, rather than a one-time, large-scale packet loss, thus avoiding "flow starvation" or a sharp drop in throughput caused by excessive packet loss.

[0020] Optionally, the preset packet loss probability rule in T1 is:

[0021]

[0022] in, This represents the probability of packet loss for the currently arriving data packets. and Represents the weighting coefficient. Represents the length of the current packet queue. Represents the current congestion threshold. This represents the proportion of the data stream to which the currently arriving data packet belongs in the data packet queue.

[0023] This scheme can achieve at least the following beneficial technical effects: it provides an optional rule for calculating packet loss probability, and the rule considers both the global congestion level through the first term and the weighting of aggressive flows through the second term. Therefore, the packet loss probability calculated using this rule can adjust the queue length more quickly and accurately, thereby making better use of bandwidth resources.

[0024] According to a second aspect of the present invention, a dynamic bandwidth-aware active queue management system based on the method described in the first aspect of the present invention is provided, comprising a congestion indicator and a congestion controller, wherein the congestion indicator is used to: respond in real time to the arrival of data packets, obtain the available bandwidth value of the current network link and the length of the current data packet queue; obtain a congestion threshold based on the available bandwidth value of the current network link and a preset target forwarding delay; compare the length of the current data packet queue with the current congestion threshold to obtain the current state and send it to the congestion controller. The congestion controller is used to receive the current state, and if the received current state is a congestion state, determine whether the data stream to which the currently arriving data packet belongs is an aggressive stream according to a preset method; if it is an aggressive stream, randomly discard the currently arriving data packet according to a preset packet dropping method.

[0025] Compared with the prior art, the advantages of the present invention are: by managing queues by mapping the changing available bandwidth of links to dynamic thresholds, network devices can efficiently utilize the currently available network bandwidth; by dynamically identifying aggressive flows and setting targeted packet loss policies, network devices can provide forwarding services more fairly. Attached Figure Description

[0026] The embodiments of the present invention will be further described below with reference to the accompanying drawings, wherein:

[0027] Figure 1 This is a schematic diagram illustrating the processing flow of arriving data packets using the dynamic bandwidth-aware active queue management method according to an embodiment of the present invention.

[0028] Figure 2 This is a schematic diagram of the architecture of a dynamic bandwidth-aware active queue management system according to an embodiment of the present invention;

[0029] Figure 3 This is a schematic diagram illustrating the queue length control effect of the active queue management method based on dynamic bandwidth awareness according to an embodiment of the present invention.

[0030] Figure 4 This is a schematic diagram comparing the queue length control effect of the dynamic bandwidth-aware active queue management method according to an embodiment of the present invention with other methods;

[0031] Figure 5 This is a schematic diagram comparing the throughput and packet loss rate of the dynamic bandwidth-aware active queue management method according to an embodiment of the present invention with other methods.

[0032] Figure 6 This is a schematic diagram comparing the tail latency of the dynamic bandwidth-aware active queue management method according to an embodiment of the present invention with other methods.

[0033] Figure 7 This is a schematic diagram comparing the data packet round-trip latency of the dynamic bandwidth-aware active queue management method according to an embodiment of the present invention with other methods;

[0034] Figure 8 This diagram illustrates a comparison of fairness between the dynamic bandwidth-aware active queue management method according to an embodiment of the present invention and other methods. Detailed Implementation

[0035] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative and are not intended to limit the invention.

[0036] As mentioned in the background section, existing AQM solutions have two major problems when facing complex scenarios such as 5G networks with drastic fluctuations in available bandwidth, multi-stream competition, and deep buffers: fixed threshold settings prevent network devices from efficiently utilizing the currently available network bandwidth; and static packet loss policies prevent network devices from providing fair forwarding services for different data streams.

[0037] The inventors believe that the key to solving the above problems lies in real-time sensing of network link bandwidth changes and mapping them to queue management strategies. This enables network devices to issue timely congestion warnings and appropriately drop packets when bandwidth suddenly drops, and relax packet loss restrictions when bandwidth recovers, thereby improving the network devices' ability to utilize link bandwidth. Furthermore, fairness is enhanced by identifying different data flows and implementing dynamic packet loss management for different data flows. Based on this design concept, this invention proposes a Dynamic Bandwidth Aware Active Queue Management (DBAQM) method to address the problems existing in the prior art.

[0038] In summary, the method provided by this invention includes the following steps: S1. Real-time response to data packet arrival, obtaining the available bandwidth value of the current network link and the length of the current data packet queue. S2. Calculating the current congestion threshold according to the available bandwidth value of the current network link and the preset target forwarding delay according to preset rules. S3. Comparing the length of the current data packet queue with the current congestion threshold to obtain the current state; a congestion state is defined as the length of the current data packet queue being greater than or equal to the current congestion threshold. In the congestion state, it is determined whether the data stream to which the currently arriving data packet belongs is an aggressive flow according to a preset method; if it is an aggressive flow, the currently arriving data packet is randomly discarded according to a preset packet dropping method.

[0039] As shown in steps S1 and S2, in DBAQM, the available bandwidth of the current network link is directly linked to the congestion threshold. Therefore, when the link bandwidth fluctuates, the congestion threshold changes dynamically. Thus, in the early stages of a rapid decrease in link bandwidth and impending congestion, packet loss is quickly triggered, effectively suppressing the growth of the packet queue and controlling network congestion. After the link bandwidth recovers, packet loss is immediately stopped or mitigated, thus fully utilizing the link bandwidth. It is evident that DBAQM overcomes the limitations of the fixed configuration of traditional AQM, achieving fine-grained queue control in highly dynamic network scenarios by real-time sensing of network link status and adaptive adjustment. Furthermore, as shown in step S3, DBAQM also defines aggressive flows and applies targeted packet loss strategies to them. That is, aggressive flows "pay" for congestion costs, protecting non-aggressive flows and thus achieving fair data forwarding.

[0040] To better understand the present invention, each step will be described in detail below with reference to specific embodiments.

[0041] According to one embodiment of the present invention, the method for obtaining the available bandwidth value of the current network link in step S1 is as follows: the amount of transmitted data is measured at a fixed sampling period on the downlink path of the network device, and then the available bandwidth value of the current network link is obtained as follows:

[0042]

[0043] in, This represents the available bandwidth value of the current network link. Represents a fixed sampling period. This represents the amount of data transmitted as measured.

[0044] According to an embodiment of the present invention, the preset rule in step S2 is:

[0045]

[0046] in, Represents the congestion threshold. This represents the preset target forwarding latency. This represents the available bandwidth value of the current network link. This represents the maximum transmission unit (MTU) of a fixed network link. This rule shows that as the available link bandwidth decreases, Automatically decreases, thus triggering congestion before the packet queue has grown significantly; as available link bandwidth increases, Automatically increases the queue size, allowing for moderate accumulation of data packets to fully utilize the increased link bandwidth and improve the actual throughput of network devices.

[0047] According to an embodiment of the present invention, the preset method in step S3 includes: S31, obtaining the length of the data stream to which the currently arriving data packet belongs in the data packet queue. and the number of data streams currently in the packet queue S32, through Calculate the length percentage of the data stream to which the currently arriving data packet belongs in the data packet queue. ,in Represents the length of the current packet queue. S33, via Calculate the current flow percentage threshold ,in This represents the set tolerance error. S34, Comparison and The size is used to determine whether the data stream to which the currently arriving data packet belongs is an aggressive stream, where, if This data stream is then considered an aggressive stream. It should be noted that as data packets continuously arrive, are enqueued, and are forwarded, and The values ​​of these values ​​are constantly changing, so a data stream that is currently considered an aggressive stream may become a non-aggressive stream in the next moment, and vice versa. This allows network devices to provide fairer forwarding services for different data streams and avoid non-aggressive streams bearing higher latency.

[0048] According to an embodiment of the present invention, the preset packet loss method includes: T1, calculating the packet loss probability of the currently arriving data packet using a preset packet loss probability rule; T2, generating a random number for the currently arriving data packet; T3, comparing the packet loss probability of the currently arriving data packet with the random number generated in step T2, wherein if the packet loss probability is greater than or equal to the random number, the arriving data packet is discarded. This embodiment uses a probabilistic packet loss strategy to smooth queue changes, rather than a one-time, large-scale packet loss, thus avoiding "flow starvation" or a sharp drop in network throughput caused by excessive packet loss.

[0049] According to an embodiment of the present invention, the preset packet loss probability rule in T1 is:

[0050]

[0051] in, This represents the probability of packet loss for the currently arriving data packets. and Represents the weighting coefficient. Represents the length of the current packet queue. Represents the current congestion threshold. This represents the proportion of the data stream to which the currently arriving data packet belongs in the packet queue. This packet loss probability rule considers both the global congestion level through the first term and the weighting of aggressive flows through the second. This composite packet loss strategy based on dynamic bandwidth thresholds and flow proportions ensures that the queue length fluctuates within a small range around the threshold. Therefore, the packet loss probability calculated using this strategy can adjust the queue length more quickly and accurately, thereby better utilizing the actual available link bandwidth resources.

[0052] To better understand the above embodiments, please refer to... Figure 1The processing flow for arriving data packets is explained as follows: When a new data packet arrives, the available bandwidth of the current network link is monitored and obtained in real time. Simultaneously, the current queue length information is retrieved from the data packet queue, and the current congestion threshold is calculated. The network is then assessed for congestion based on the current queue length and the congestion threshold. If the network is not currently congested, the arriving data packet is placed in the queue, and the processing flow for that data packet ends. If the network is congested, the length percentage of the data stream to which the data packet belongs in the queue is further obtained. Based on this percentage and the number of data streams in the queue, it is determined whether the data stream belongs to an aggressive flow. If the data stream is not an aggressive flow, the arriving data packet is placed in the queue, and the processing flow for that data packet ends. If the data stream belongs to an aggressive flow, the packet loss probability is calculated, and a random number is generated for comparison with the packet loss probability. If the probability value is not lower than the random number, the arriving data packet is placed in the queue; otherwise, the data packet is discarded, and the processing flow for that data packet ends. Overall, the DBAQM solution effectively avoids excessive backlog in network device buffers (i.e., packet queues) by sensing dynamic changes in link bandwidth and queue length in real time and taking differentiated congestion control measures for different data streams, thus ensuring the stability of overall network forwarding performance and the fairness of network resource allocation.

[0053] According to one embodiment of the present invention, the inventors also provide a dynamic bandwidth-aware active queue management system based on the above embodiments. (See also...) Figure 2 Network devices contain packet queues and queue entry buffers. When a new packet arrives at the network device, it typically waits in the queue entry buffer before entering the queue. The system then determines whether to discard the new packet or add it to the packet queue, and executes the decision. This system includes a congestion indicator and a congestion controller. The congestion indicator is used to: respond to packet arrivals in real time; obtain the available bandwidth of the current network link and the length of the current packet queue; obtain a congestion threshold based on the available bandwidth of the current network link and a preset target forwarding delay; compare the length of the current packet queue with the current congestion threshold to obtain the current status and send it to the congestion controller. The congestion controller receives the current status. If the received current status is congested, it determines whether the data stream to which the currently arriving packet belongs is an aggressive flow according to a preset method. If it is an aggressive flow, it randomly discards the currently arriving packet according to a preset packet dropping method.

[0054] To verify the technical effectiveness of this invention, the inventors conducted separate experiments on the DBAQM scheme and comparative experiments on several schemes, including FIFO, RED, CoDel, and DBAQM, in the NS-3 simulation environment. The experimental evaluation metrics included queue length control effectiveness, throughput, packet loss rate, round-trip time (especially the 99th percentile round-trip time, i.e., tail latency), and network fairness. The specific experimental environment configuration was as follows: the network device's packet queue capacity was set to a deep buffer scenario; the available bandwidth of the link simulated the frequent fluctuations of a 5G network; multiple data streams were started, and the sending end used congestion control protocols such as Cubic, BBR, and Copa to send the data streams. The specific experimental results are explained below:

[0055] First, let's explain separately the effect of the DBAQM scheme on controlling queue length (number of data packets).

[0056] Reference Figure 3 The dynamic relationship between the number of packets, the threshold, and the probability of packet loss in the displayed packet queue over time is as follows: In the initial stage, the available bandwidth of the link is high, and the queue length has not exceeded the threshold, so no packet loss is triggered (corresponding to...). Figure 3 During approximately the first second of the transmission cycle); when bandwidth suddenly drops, the queue length rapidly exceeds the threshold, triggering packet loss and a rapid increase in the probability of packet loss. This forces the sender to slow down data transmission, and the queue length begins to decrease (corresponding to...). Figure 3 During the initial 1-5 seconds, DBAQM continuously and dynamically adjusts the threshold and packet loss probability. When the queue length exceeds the threshold, the packet loss probability increases; when it falls below the threshold, packet loss stops, thus causing the queue length to oscillate slightly around the threshold, achieving stable control of the queue length (corresponding to...). Figure 3 (Approximately 5-25 seconds).

[0057] Next, we will use a comparison to illustrate the advantages and disadvantages of the DBAQM solution compared to other solutions.

[0058] Reference first Figure 4 The demonstration shows the changes in queue length (number of packets) over time in the FIFO, RED, CoDel, and DBAQM schemes: FIFO queues have almost no control mechanism during congestion, resulting in a surge in queue length, while RED, CoDel, and DBAQM effectively control queue length. Based on this, refer to... Figure 5 By simultaneously transmitting a 98.58MB file, the throughput and packet loss rates of RED, CoDel, and DBAQM, which have better queue control, were further compared. Figure 5As shown in the red and yellow bars, DBAQM has significant advantages over RED and CoDel in terms of throughput and packet loss rate. Specifically, compared to RED, it increases throughput by 16.2% and reduces packet loss rate by 57.1%; compared to CoDel, it increases throughput by 30.3% and reduces packet loss rate by 66.7%.

[0059] Then refer to the appendix. Figure 6 The comparison of tail latency (99% round-trip latency) under different queue management strategies is shown: When running Cubic, Copa, and BBR congestion control protocols at the sending end, DBAQM significantly reduces tail latency compared to FIFO, and slightly increases it compared to RED and CoDel, but it remains within an acceptable level. For a more intuitive understanding of the latency and tail latency of the different schemes, refer to... Figure 7 The cumulative latency distribution shown is as follows: When running FIFO, less than 20% of the data packets have a round-trip latency of less than 50ms, and less than 40% have a round-trip latency of less than 100ms (tail latency is not ideal); when running RED or CoDel, approximately 95% of the data packets have a round-trip latency of less than 50ms, and 100% have a round-trip latency of less than 100ms; when running DBAQM, approximately 85% of the data packets have a round-trip latency of less than 50ms, and more than 99.5% have a round-trip latency of less than 100ms, further demonstrating that the latency and tail latency metrics of DBAQM are not significantly different from those of RED and CoDel.

[0060] Finally refer to Figure 8 Fairness comparison: Thanks to DBAQM's definition of aggressive flow and its differentiated packet loss strategy for aggressive flow, DBAQM's Jain fairness coefficient reaches 0.910 under multi-flow competition, which is better than all other queue management solutions.

[0061] Therefore, the experimental results show that, under the experimental conditions, network devices using the DBAQM scheme have significant advantages in throughput and packet loss rate compared to other AQM schemes, and also have certain advantages in terms of network fairness; although they have some disadvantages in terms of latency compared to some RED and CoDel schemes, it is not significant.

[0062] In summary, this invention manages queues by mapping changing link available bandwidth to dynamic thresholds, enabling network devices to efficiently utilize currently available network bandwidth. By dynamically identifying aggressive flows and setting targeted packet loss policies, it enables network devices to provide forwarding services more fairly. Based on these advantages, the solution of this invention is particularly suitable for network environments such as 5G networks where available bandwidth changes frequently and service fairness is required.

[0063] It should be noted that although the steps are described in a specific order above, it does not mean that the steps must be executed in the above specific order. In fact, some of these steps can be executed concurrently, or even in a different order, as long as the required function can be achieved.

[0064] This invention can be a network device, a computer program product, or a computer-readable storage medium. Network devices primarily refer to devices used for data forwarding, such as routers, switches, firewalls, VPNs, wireless controllers, wireless access points, OLTs, and ONUs, that manage whether incoming data packets are placed into their data packet queues using this solution. Computer program products primarily refer to software products that implement this solution through computer programs. Computer-readable storage media can be tangible devices that hold and store instructions used by instruction execution devices. Computer-readable storage media can include, but are not limited to, electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusion structures storing instructions thereon, and any suitable combination thereof.

[0065] The various embodiments of the present invention have been described above. These descriptions are exemplary and not exhaustive, nor are they limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is chosen to best explain the principles, practical application, or technical improvements to the embodiments in the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A dynamic bandwidth-aware active queue management method for managing whether real-time arriving data packets in a network device are placed into a data packet queue, wherein, A data packet queue is used to store data packets to be forwarded, characterized in that the method includes the following steps: S1. Real-time response to incoming data packets, obtaining the available bandwidth of the current network link and the length of the current data packet queue; S2. Calculate the current congestion threshold according to the available bandwidth of the current network link and the preset target forwarding delay according to the preset rules. S3. Compare the length of the current data packet queue with the current congestion threshold to obtain the current state. The current state is congested when the length of the current data packet queue is greater than or equal to the current congestion threshold. In the congested state, determine whether the data stream to which the currently arriving data packet belongs is an aggressive stream according to a preset method. If it is an aggressive stream, randomly discard the currently arriving data packet according to a preset packet dropping method.

2. The active queue management method based on dynamic bandwidth awareness according to claim 1, characterized in that, The method for obtaining the available bandwidth value of the current network link in step S1 is as follows: the amount of transmitted data is measured at a fixed sampling period on the downlink path of the network device, and then the available bandwidth value of the current network link is obtained as follows: in, This represents the available bandwidth value of the current network link. Represents a fixed sampling period. This represents the amount of data transmitted as measured.

3. The active queue management method based on dynamic bandwidth awareness according to claim 1, characterized in that, The preset rule mentioned in step S2 is: in, Represents the congestion threshold. This represents the preset target forwarding latency. This represents the available bandwidth value of the current network link. It represents the maximum transmission unit that is fixed in a network link.

4. The active queue management method based on dynamic bandwidth awareness according to claim 1, characterized in that, The preset methods mentioned in step S3 include: S31. Obtain the length of the data stream to which the currently arriving data packet belongs in the data packet queue. and the number of data streams currently in the packet queue ; S32, Through Calculate the length percentage of the data stream to which the currently arriving data packet belongs in the data packet queue. ,in Represents the length of the current packet queue; S33, Through Calculate the current flow percentage threshold ,in This represents the set tolerance error; S34, Comparison and The size is used to determine whether the data stream to which the currently arriving data packet belongs is an aggressive stream, where, if Then the data stream is an aggressive stream.

5. The active queue management method based on dynamic bandwidth awareness according to claim 1, characterized in that, The preset packet loss methods include: T1. Calculate the packet loss probability of the currently arriving data packet using the preset packet loss probability rules; T2. Generate a random number for the currently arriving data packet; T3. Compare the packet loss probability of the currently arriving data packet with the size of the random number generated in step T2. If the packet loss probability is greater than or equal to the random number, then discard the arriving data packet.

6. The active queue management method based on dynamic bandwidth awareness according to claim 5, characterized in that, The preset packet loss probability rule mentioned in T1 is as follows: in, This represents the probability of packet loss for the currently arriving data packets. and Represents the weighting coefficient. Represents the length of the current packet queue. Represents the current congestion threshold. This represents the proportion of the data stream to which the currently arriving data packet belongs in the data packet queue.

7. A dynamic bandwidth-aware active queue management system based on the method of any one of claims 1-6, characterized in that, The system includes a congestion indicator and a congestion controller, wherein: Congestion indicator is used for: Real-time response to incoming data packets, obtaining the available bandwidth of the current network link and the length of the current data packet queue; The congestion threshold is obtained based on the available bandwidth of the current network link and the preset target forwarding delay; Compare the current packet queue length with the current congestion threshold to obtain the current state and send it to the congestion controller; The congestion controller receives the current status. If the received current status is a congestion status, it determines whether the data stream to which the currently arriving data packet belongs is an aggressive stream according to a preset method. If it is an aggressive stream, it randomly discards the currently arriving data packet according to a preset packet dropping method.

8. A network device, characterized in that, The network device manages whether real-time arriving data packets are placed into its data packet queue using the method described in any one of claims 1-6.

9. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the method as described in any one of claims 1-6.

10. A computer-readable storage medium, characterized in that, It stores a computer program, which is executed by a processor to implement the steps of the method according to any one of claims 1-6.

Images