Dynamic buffer management method and system for Wi-Fi network
By using a dynamic buffer management method to identify latency-sensitive flows and adjust buffer thresholds, the buffering problem caused by the time-varying characteristics of wireless links in Wi-Fi networks is solved, thereby improving network performance and QoS.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-01-06
- Publication Date
- 2026-04-07
AI Technical Summary
Traditional static queue management strategies are difficult to effectively cope with the time-varying characteristics of wireless links in Wi-Fi networks, leading to sudden accumulation of buffer packets, packet loss, and a sharp increase in RTT, which affects network performance.
By identifying latency-sensitive flows through access control lists, calculating the total latency and available bandwidth of data packets in the wireless link, dynamically adjusting the three-level buffer thresholds, and combining a weighted random early detection strategy, the data packet transmission rate and dropping strategy are adjusted to achieve dynamic buffer management.
It effectively suppresses sudden accumulation in the buffer, improves network throughput, reduces packet loss rate, ensures QoS for latency-sensitive traffic, and adapts to the high volatility of wireless links.
Smart Images

Figure CN121815329A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the field of computer network congestion control algorithm, and particularly relates to a dynamic buffer management method for Wi-Fi network. BACKGROUND
[0002] With the rapid development of distributed systems, cloud computing and edge computing, end-to-end network performance optimization is facing new challenges. In the modern hybrid network architecture, wired networks (including backbone networks and access networks) and Wi-Fi networks usually coexist, forming a complex end-to-end path. However, due to the inherent signal attenuation, multipath interference and dynamic channel competition of wireless links, its transmission stability is significantly lower than that of wired links, which has a multi-faceted impact on network performance. TCP protocol, as the mainstream transmission protocol, its congestion control mechanism relies on the interpretation of packet loss and RTT changes. However, in a wireless environment, the high volatility of wireless links can easily cause burst accumulation of router buffer packets, resulting in a large number of packet losses and sharp growth of RTT.
[0003] Traditional static queue management strategies are difficult to effectively cope with the time-varying characteristics of wireless links. Although the virtual queue technology achieves differentiated scheduling at the router level by allocating independent queues for different priority traffic, the high volatility of Wi-Fi links makes it difficult for static virtual queue configuration to adapt to the rapid changes in channel state. The active queue management technology optimized for wireless environments also has adaptation problems. Although the weighted random early detection (WRED) prevents buffer overflow by probabilistic packet loss, the rate adaptation mechanism (MCS adjustment) unique to Wi-Fi links can cause step changes in bandwidth, and the static threshold of traditional WRED cannot track this non-linear change in real time, resulting in premature packet loss (false congestion) when the link speed is reduced or a delayed response when the link speed is increased.
[0004] In view of this, it is necessary to propose a dynamic buffer management algorithm for Wi-Fi networks to solve the above problems. SUMMARY
[0005] The present application aims to at least solve one of the technical problems existing in the related art to some extent.
[0006] The present application aims to provide a dynamic buffer management method and system for Wi-Fi networks, which can effectively suppress the burst accumulation of buffers caused by the high volatility of wireless links by dynamically adjusting the multi-level threshold based on real-time sensing of wireless link state and estimating the buffer queue length, thereby achieving high throughput and low latency of the network.
[0007] To achieve the above purpose, the present application provides a dynamic buffer management method for Wi-Fi network, comprising the following steps:
[0008] S1, identifying the flow to which the data packet entering the router belongs through an access control list, and marking the data packet meeting the delay-sensitive flow feature;
[0009] S2, calculating the total delay of the marked data packet in wireless link transmission, including long-term queuing delay, short-term queuing delay and transmission delay;
[0010] S3, reasoning and estimating the available bandwidth of the current wireless link by analyzing the frame interaction behavior of the Wi-Fi MAC layer on the router side for the Wi-Fi network;
[0011] S4, dynamically calculating the bandwidth delay product according to the transmission delay and the available bandwidth of the current wireless link, and setting a three-level buffer dynamic threshold that is updated in real time with the change of the link state based on the bandwidth delay product;
[0012] S5, setting the buffer control mechanism according to the obtained three-level buffer dynamic threshold and the target delay of the marked data packet, adjusting the sending rate of the data packet entering the router, or probabilistically discarding the unmarked data packet based on the weighted random early detection strategy.
[0013] The further preferred technical solution of the application is that in step S1, the flow to which the data packet entering the router belongs is identified through an access control list, specifically:
[0014] According to the service characteristics of the data flow and the sensitivity of the data flow to delay and packet loss, the determination criteria of various flow types are determined, so as to establish the flow classification rule for distinguishing the key flow and the non-key flow;
[0015] Based on the flow classification rule, the access control list rule is constructed combined with the identification information of the data packet, and the key flow data packet is identified.
[0016] As a preferred, in step S1, the data packet meeting the delay-sensitive flow feature is marked; specifically:
[0017] When the data packet entering the router matches the constructed access control list rule successfully, the key flow data packet is identified, and the router adds metadata information to the data packet in the data plane as a mark; at the same time, the data packet is explicitly marked in the least significant bit of the service type field of the IP packet header to indicate the priority of the data packet.
[0018] As a preferred, in step S2, the total delay of the marked data packet in wireless link transmission is calculated, specifically:
[0019] In the wireless link, the total delay of the data packet from entering the router to reaching the receiver is divided into queuing delay and transmission delay;
[0020] Wherein, the queuing delay is the time consumption of the data packet from arriving at the access point to leaving the queue and being scheduled to enter the underlying driver; the queuing delay includes long-term queuing delay and short-term queuing delay ;
[0021] The long-term queuing delay is the time from the data packet entering the queue to becoming the head-of-line packet, and is expressed as:
[0022] ;
[0023] Wherein, is the current queue length, is the average dequeue rate measured in a sliding window;
[0024] The short-term queuing delay is the time consumption of the data packet from becoming the head-of-line to finally being dequeued, which is obtained by measuring the residence time of the head-of-line packet of the current queue;
[0025] The transmission delay is the time experienced by the data packet from being dequeued to being successfully transmitted to the receiving end, which is constrained by the transmission capability of the wireless channel; the estimation of the transmission delay adopts a sliding window method to measure the average interval between the network layer queue packets, and the sliding window length needs to cover at least two sending end burst periods;
[0026] The total delay of the data packet in the wireless link is expressed as:
[0027] .
[0028] As preferred, the step S3 is described as estimating and calculating the available bandwidth of the current wireless link by analyzing the frame interaction behavior of the Wi-Fi MAC layer on the router side for the Wi-Fi network; specifically:
[0029] Defining that the router sends data frames in the A-MPDU manner, and the maximum number of frames allowed in a batch is M which is determined by negotiation with the receiving end, the actual number of aggregated frames sent by the router is recorded as b, if there is no data backlog in the buffer of the router, the router sends aggregated frames in the actual batch size b, the receiving end confirms the whole batch of data through the Block ACK, and the router calculates the instantaneous dequeue rate according to the ACK interval :
[0030] ;
[0031] Wherein, is the number of bits of a single frame;
[0032] When there is data backlog in the buffer, that is, the instantaneous dequeue rate i.e. the link bandwidth;
[0033] When the router buffer has no data backlog, i.e. the theoretical ACK interval at full batch Since the mean of ACK intervals of different batch sizes is linear with slope S / R, the theoretical ACK interval is expressed as:
[0034] ;
[0035] where R is the physical layer transmission bit rate;
[0036] Further, the available bandwidth of the current wireless link is estimated according to the theoretical ACK interval , and the calculation formula is:
[0037] .
[0038] As a preferred embodiment, step S4 comprises dynamically calculating the bandwidth-delay product according to the transmission delay and the available bandwidth of the current wireless link, and setting the three-level buffer dynamic threshold which is updated in real time according to the link state based on the bandwidth-delay product; specifically:
[0039] According to the transmission delay and the available bandwidth of the current wireless link , the bandwidth-delay product of the wireless link at the current time is dynamically calculated , and the calculation formula is:
[0040] ;
[0041] Based on the bandwidth-delay product of the wireless link at the current time , the three-level buffer dynamic threshold is set to distinguish different degrees of buffer congestion state, and the three levels are respectively:
[0042] ;
[0043] ;
[0044] ;
[0045] wherein, is a preset proportion coefficient, , .
[0046] As a preferred embodiment, the buffer regulation mechanism in step S5 comprises:
[0047] the buffer queue length compared with the three-level buffer dynamic threshold, when the buffer queue length exceeds the first-level threshold and does not reach the second-level threshold, i.e. , an acceleration mark is added to the data packet to prompt the sending end to increase the sending rate; when the buffer queue length exceeds the second-level threshold and does not reach the third-level threshold, i.e. , an explicit congestion notification mark is added to the data packet to inform the sending end to reduce the sending rate; when the buffer queue length exceeds the third-level threshold, i.e. , the router probabilistically discards the unmarked data packet based on a weighted random early detection strategy.
[0048] As a preferred, the buffer regulation mechanism in step S5 further comprises:
[0049] judging whether the total delay of the marked data packet in the wireless link transmission is within a preset maximum end-to-end allowed delay upper limit , if , the router probabilistically discards the unmarked data packet based on a weighted random early detection strategy; is a preset proportion coefficient, satisfying .
[0050] Another aspect of the present application provides a dynamic buffer management system for a Wi-Fi network, comprising:
[0051] a data packet classification module, configured to identify the flow to which the data packet entering the router belongs through an access control list, and mark the data packet satisfying the delay-sensitive flow characteristics;
[0052] a data packet delay prediction module, configured to calculate the total delay of the marked data packet in the wireless link transmission, including long-term queuing delay, short-term queuing delay and transmission delay;
[0053] a wireless link bandwidth estimation module, configured to infer and estimate the available bandwidth of the current wireless link through analyzing the frame interaction behavior of the Wi-Fi MAC layer on the router side for the Wi-Fi network;
[0054] a dynamic threshold setting module, configured to dynamically calculate the bandwidth delay product according to the transmission delay and the available bandwidth of the current wireless link, and set the three-level buffer dynamic threshold which is updated in real time with the change of the link state based on the bandwidth delay product;
[0055] a buffer dynamic regulation module, configured to set the regulation mechanism of the buffer according to the obtained three-level buffer dynamic threshold and the target delay of the marked data packet, adjust the sending rate of the data packet entering the router, or probabilistically discard the unmarked data packet based on a weighted random early detection strategy.
[0056] Another aspect of the present application provides a non-transitory computer readable storage medium having stored thereon computer instructions that cause a computer to perform the dynamic buffer management method for Wi-Fi network.
[0057] Still another aspect of the present application provides an electronic device comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus, and the processor invokes logical instructions in the memory to perform the dynamic buffer management method for Wi-Fi network.
[0058] Still another aspect of the present application provides a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, which, when executed by a processor, causes a computer to perform the dynamic buffer management method for Wi-Fi network.
[0059] Beneficial effects: The present application formalizes the dynamic buffer management problem in Wi-Fi network, solves the problem that the traditional static queue management strategy is difficult to effectively cope with the time-varying characteristics of wireless link, and maximally reduces the router buffer overflow packet loss problem. First, when the data packet enters the router, the ACL module marks the delay-sensitive traffic as a critical flow according to the characteristics of the traffic, providing a basis for subsequent differentiated control and effectively improving the QoS of delay-sensitive traffic. Secondly, estimate the data packet delay and buffer queue length at the router, and estimate the wireless link bandwidth, dynamically set the buffer three-level threshold, and improve the sensitivity and adaptability of congestion control. Finally, compare the queue length and data packet delay with the buffer threshold, respectively trigger the accelerated marking, ECN marking or WRED-based probabilistic packet loss behavior, effectively reducing the buffer overflow packet loss. BRIEF DESCRIPTION OF DRAWINGS
[0060] Figure 1 The flow chart of the dynamic buffer management method for Wi-Fi network of the present application.
[0061] Figure 2 The system architecture diagram of the dynamic buffer management system for Wi-Fi network of the present application.
[0062] Figure 3 The composition structure diagram of the total delay of the data packet transmission in the wireless link of the present application.
[0063] Figure 4 The simulation verification result diagram based on the dynamic buffer management method for Wi-Fi network of the present application. DETAILED DESCRIPTION
[0064] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in conjunction with the drawings in the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. They should not be understood as limitations to the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work belong to the scope of protection of the present application. In the description of the present application, it should be understood that the terms used are only for the purpose of description and should not be understood as indicating or implying relative importance.
[0065] The technical solutions of the present application will be described below in conjunction with Figures 1-4 The present application provides a dynamic buffer management method and system for Wi-Fi network.
[0066] Embodiment 1: The present embodiment provides a dynamic buffer management method (W-DBM) for Wi-Fi network, as shown in the following steps: Figure 1
[0067] S1, identifying the flow to which the data packet entering the router belongs through an access control list (ACL), and marking the data packet meeting the delay-sensitive flow characteristics.
[0068] When the data packet enters the data plane of the router, the router classifies and identifies the flow to which the data packet belongs through the ACL, and identifies the flow meeting the delay-sensitive characteristics (including but not limited to small-scale flow, service flow sensitive to packet loss and with low delay requirement) according to the preset matching rule, and adds identification information to the data packet for subsequent processing. Specifically:
[0069] Establishing a flow classification rule, formulating a flow type judgment standard according to the service attribute of the data flow and its sensitivity to delay and packet loss, for identifying and distinguishing key flow and non-key flow;
[0070] Based on the flow classification rule, combining the identification information of the data packet (including but not limited to five-tuple, flow identifier, sequence number range, etc.), configuring corresponding matching rules in the ACL, respectively corresponding to key flow and non-key flow, and constructing a distinguishable rule entry set;
[0071] When the data packet matches the ACL rule successfully, the router adds metadata information for identification to the data packet in the data plane as a mark; at the same time, the least significant bit of the service type field (DSCP) in the IP packet header is used to mark the data packet explicitly to identify its priority level, for subsequent differentiated processing.
[0072] S2, calculating the total delay of the marked data packet in the wireless link.
[0073] For tagged data packets, the system breaks down the delay in their transmission path, such as... Figure 3 As shown, this includes queuing delay and transmission delay.
[0074] Queuing delay refers to the time it takes for a data packet to travel from the arrival of the access point queue to its scheduling into the underlying driver, and can be further subdivided into long-term queuing delay. and short-term queuing delay Long-term queuing delay and short-term queuing delay are estimated separately to achieve accurate, real-time prediction of the overall packet delay.
[0075] Long queue delay The time it takes for a data packet to go from entering the queue to becoming the first packet in the queue is represented as:
[0076] ;
[0077] in, This is the current queue length. The average dequeue rate is measured within a sliding window; this metric reflects the queuing backlog trend caused by contention or RTC burst traffic in the wireless link.
[0078] Short-term queuing delay The queuing delay, from the moment a data packet becomes the head of the queue to its final departure, is primarily influenced by link-layer scheduling patterns (such as MAC data unit aggregation) and is susceptible to fluctuations in current link contention trends and scheduling policies. This delay is estimated by real-time measurement of the dwell time of the current head-of-the-queue data packet, and can sensitively reflect the decreasing trend of link bandwidth. Therefore, the total queuing delay can be expressed as… and The sum. To avoid overestimating latency due to sudden batch dequeueing caused by link-layer aggregation, a correction mechanism is introduced to calibrate the queue length. The correction formula is:
[0079] ;
[0080] in This represents the maximum amount of aggregated outgoing data that may occur at the link layer with a 1ms time granularity.
[0081] Transmission delay This refers to the link time between a data packet being dequeued from the queue and successfully transmitted to the receiver, which is constrained by the transmission capacity of the wireless channel. Since a wireless channel can only transmit one data unit at any given time, this transmission delay is estimated by measuring the average of the network layer data packet dequeue intervals using a sliding window method. The sliding window length must cover at least two burst cycles from the transmitter to ensure the stability and continuity of the estimation results.
[0082] Total delay of the final data packet transmitted in the wireless link , expressed as:
[0083] .
[0084] S3. Infer and estimate the available bandwidth of the current wireless link.
[0085] On the router side for Wi-Fi networks, by analyzing the frame interaction behavior of the Wi-Fi MAC layer, including but not limited to parameters such as data frame transmission timing and acknowledgment frame (ACK) intervals, infer and estimate the available bandwidth of the current wireless link.
[0086] The router sends data frames (MPDUs) in batches through the Aggregate MAC Protocol Data Unit (A-MPDU) mechanism in the 802.11n standard. After negotiation with the receiving end, the router determines that the maximum number of frames allowed for each aggregation batch is M. When there is no obvious data backlog in the router buffer, the actual number of aggregated frames sent by the router is denoted as b (b < M). The receiving end confirms the entire aggregated frame batch through a single Block ACK. Based on this ACK feedback, the instantaneous dequeue rate tr(t) at the corresponding moment can be calculated, and its calculation formula is:
[0087] ;<00S4. Based on the transmission delay and the available bandwidth of the current wireless link, dynamically calculate the bandwidth-delay product, and set a three-level buffer dynamic threshold that is updated in real time with changes in the link status based on the bandwidth-delay product.
[0095] Based on transmission delay and the available bandwidth of the current wireless link Dynamically calculate the bandwidth-delay product of the wireless link at the current moment. The calculation formula is:
[0096] ;
[0097] The bandwidth-delay product of the wireless link at the current moment Three levels of dynamic thresholds are set to distinguish different degrees of buffer congestion:
[0098] ;
[0099] ;
[0100] ;
[0101] in, This is a preset proportional coefficient. , The T1 threshold indicates that the current data packet transmission rate is not fully utilizing the wireless link bandwidth; the T2 threshold indicates that the transmission rate is slightly higher than the link's carrying capacity; the T3 threshold reflects that the buffer is about to experience severe congestion, requiring timely and stronger control measures.
[0102] S5. Based on the obtained three-level buffer dynamic threshold, set the buffer control mechanism and the target delay of the labeled data packets, adjust the sending rate of the data packets entering the router, or based on the weighted random early detection strategy, probabilistically discard unlabeled data packets.
[0103] To ensure QoS requirements, the buffer queue length qSize is compared with a three-level dynamic threshold, and it is determined whether the total delay of critical data packets (totalDelay) is within the preset maximum allowable end-to-end delay limit. The transmission is completed within the specified timeframe. When the marked critical data packet meets the specified condition... This indicates that the data packet is not expected to be transmitted within the required latency limit, making it difficult to meet the needs of latency-sensitive services. The preset proportional coefficient satisfies At this point, the router, based on the WRED policy, probabilistically discards unlabeled, non-critical data packets, thereby freeing up buffer resources for critical flows and ensuring their transmission timeliness.
[0104] Length of the buffer queue Compared with the dynamic threshold of the three-level buffer, when the buffer queue length exceeds the first-level threshold but does not reach the second-level threshold, i.e. This indicates that the transmission rate is not fully utilizing the wireless link bandwidth. To improve link utilization, an acceleration marker is added to the data packet to prompt the sender to increase the transmission rate.
[0105] When the buffer queue length exceeds the second-level threshold but does not reach the third-level threshold, i.e. If there is a risk of congestion in the buffer, an explicit congestion notification flag (ECN) is added to the data packet to notify the sender to reduce the sending rate.
[0106] When the length of the buffer queue exceeds the third-level threshold, i.e. To prevent the buffer from continuously expanding, the router uses a weighted random early detection strategy to probabilistically discard untagged packets.
[0107] This invention formalizes the dynamic buffer management problem in Wi-Fi networks, addressing the difficulty of traditional static queue management strategies in effectively handling the time-varying characteristics of wireless links and minimizing router buffer overflow packet loss. First, when packets enter the router, the ACL module marks latency-sensitive traffic as critical flows based on traffic characteristics, providing a basis for subsequent differentiated control and effectively improving the QoS of latency-sensitive traffic. Second, at the router, packet latency and buffer queue length are estimated, and wireless link bandwidth is predicted, dynamically setting three-level buffer thresholds to improve the sensitivity and adaptability of congestion control. Finally, based on a comparison between queue length, packet latency, and buffer thresholds, acceleration marking, ECN marking, or probabilistic packet loss behavior based on WRED is triggered, effectively reducing buffer overflow packet loss.
[0108] To further verify the effectiveness of this invention, the FCT performance of this invention and the BBR algorithm in a wired-wireless hybrid network was compared using a WebSearch traffic model on the NS-3 simulation platform. The results are shown in Table 1 and... Figure 4 As shown.
[0109] Table 1 Comparison of throughput and packet loss rate
[0110]
[0111] Experimental results show that the Wi-Fi network performance optimization algorithm based on dynamic buffer management can effectively cope with the high volatility of wireless links, improve network throughput, reduce packet loss rate and significantly reduce FCT, and its overall performance is better than BBR.
[0112] Example 2: This example provides a dynamic buffer management system for Wi-Fi networks, such as... Figure 2 As shown, it includes:
[0113] The packet classification module is used to identify the flow to which the packets entering the router belong using access control lists, and to mark the packets that meet the characteristics of delay-sensitive flows.
[0114] The packet delay prediction module is used to calculate the total delay of the transmission of tagged packets in the wireless link, including long queuing delay, short queuing delay and transmission delay;
[0115] The wireless link bandwidth estimation module is used on the router side to infer and estimate the available bandwidth of the current wireless link by analyzing the frame interaction behavior of the Wi-Fi MAC layer.
[0116] The dynamic threshold setting module is used to dynamically calculate the bandwidth-delay product based on the transmission delay and the available bandwidth of the current wireless link, and set a three-level buffer dynamic threshold that is updated in real time with changes in the link status based on the bandwidth-delay product.
[0117] The buffer dynamic control module is used to set the buffer control mechanism based on the obtained three-level buffer dynamic threshold and the target delay of the labeled data packets, adjust the sending rate of data packets entering the router, or, based on a weighted random early detection strategy, probabilistically discard unlabeled data packets.
[0118] Example 3: This example provides a non-transitory computer-readable storage medium storing computer instructions that cause a computer to execute a dynamic buffer management method for a Wi-Fi network. The method includes the following steps:
[0119] S1. Identify the flow to which the data packets entering the router belong using the access control list, and mark the data packets that meet the characteristics of a delay-sensitive flow.
[0120] S2. Calculate the total delay of the tagged data packet in the wireless link, including long queuing delay, short queuing delay and transmission delay;
[0121] S3. On the router side, for the Wi-Fi network, by analyzing the frame interaction behavior of the Wi-Fi MAC layer, infer and estimate the available bandwidth of the current wireless link;
[0122] S4. Based on the transmission delay and the available bandwidth of the current wireless link, dynamically calculate the bandwidth-delay product, and set a three-level buffer dynamic threshold that is updated in real time with changes in the link status based on the bandwidth-delay product.
[0123] S5. Based on the obtained dynamic threshold of the three-level buffer and the target delay of the labeled data packets, set the buffer control mechanism to adjust the sending rate of data packets entering the router, or based on the weighted random early detection strategy, probabilistically discard unlabeled data packets.
[0124] Example 4: This example provides an electronic device that may include a processor, a communication interface, memory, and a communication bus. The processor, communication interface, and memory communicate with each other via the communication bus. The processor can invoke logical instructions from the memory to execute a dynamic buffer management method for Wi-Fi networks. This method includes the following steps:
[0125] S1. Identify the flow to which the data packets entering the router belong using the access control list, and mark the data packets that meet the characteristics of a delay-sensitive flow.
[0126] S2. Calculate the total delay of the tagged data packet in the wireless link, including long queuing delay, short queuing delay and transmission delay;
[0127] S3. On the router side, for the Wi-Fi network, by analyzing the frame interaction behavior of the Wi-Fi MAC layer, infer and estimate the available bandwidth of the current wireless link;
[0128] S4. Based on the transmission delay and the available bandwidth of the current wireless link, dynamically calculate the bandwidth-delay product, and set a three-level buffer dynamic threshold that is updated in real time with changes in the link status based on the bandwidth-delay product.
[0129] S5. Based on the obtained dynamic threshold of the three-level buffer and the target delay of the labeled data packets, set the buffer control mechanism to adjust the sending rate of data packets entering the router, or based on the weighted random early detection strategy, probabilistically discard unlabeled data packets.
[0130] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, and can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0131] Example 5: This example provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute a dynamic buffer management method for Wi-Fi networks, which includes the following steps:
[0132] S1. Identify the flow to which the data packets entering the router belong using the access control list, and mark the data packets that meet the characteristics of a delay-sensitive flow.
[0133] S2. Calculate the total delay of the tagged data packet in the wireless link, including long queuing delay, short queuing delay and transmission delay;
[0134] S3. On the router side, for the Wi-Fi network, by analyzing the frame interaction behavior of the Wi-Fi MAC layer, infer and estimate the available bandwidth of the current wireless link;
[0135] S4. Based on the transmission delay and the available bandwidth of the current wireless link, dynamically calculate the bandwidth-delay product, and set a three-level buffer dynamic threshold that is updated in real time with changes in the link status based on the bandwidth-delay product.
[0136] S5. Based on the obtained dynamic threshold of the three-level buffer and the target delay of the labeled data packets, set the buffer control mechanism to adjust the sending rate of data packets entering the router, or based on the weighted random early detection strategy, probabilistically discard unlabeled data packets.
[0137] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0138] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0139] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A dynamic buffer management method for Wi-Fi networks, characterized in that, Includes the following steps: S1. Identify the flow to which the data packets entering the router belong using the access control list, and mark the data packets that meet the characteristics of a delay-sensitive flow. S2. Calculate the total delay of the tagged data packet in the wireless link, including long queuing delay, short queuing delay and transmission delay; S3. On the router side, for the Wi-Fi network, by analyzing the frame interaction behavior of the Wi-Fi MAC layer, infer and estimate the available bandwidth of the current wireless link; S4. Based on the transmission delay and the available bandwidth of the current wireless link, dynamically calculate the bandwidth-delay product, and set a three-level buffer dynamic threshold that is updated in real time with changes in the link status based on the bandwidth-delay product. S5. Based on the obtained dynamic threshold of the three-level buffer and the target delay of the labeled data packets, set the buffer control mechanism to adjust the sending rate of data packets entering the router, or based on the weighted random early detection strategy, probabilistically discard unlabeled data packets.
2. The dynamic buffer management method for Wi-Fi networks according to claim 1, characterized in that, The step S1, which involves identifying the flow to which the data packets entering the router belong through the access control list, specifically includes: Based on the business characteristics of data flow and the sensitivity of data flow to latency and packet loss, the criteria for judging various types of flow are determined, thereby establishing flow classification rules to distinguish between critical and non-critical flows; Based on the flow classification rules and combined with the packet identification information, access control list rules are constructed to identify critical flow packets.
3. The dynamic buffer management method for Wi-Fi networks according to claim 2, characterized in that, In step S1, data packets that meet the characteristics of delay-sensitive flows are marked; Specifically: When a data packet entering the router successfully matches a rule in the constructed access control list and a critical flow data packet is identified, the router appends metadata information to the data packet in the data plane as a marker; at the same time, it explicitly marks the data packet in the least significant bit of the type of service field in the IP header to indicate the priority of the data packet.
4. The dynamic buffer management method for Wi-Fi networks according to claim 1, characterized in that, Step S2 calculates the total transmission delay of the tagged data packets in the wireless link, specifically as follows: In a wireless link, the total delay from when a data packet enters the router to when it arrives at the receiver is divided into queuing delay and transmission delay. Queuing latency is the time it takes for a data packet to travel from the arrival of the access point to leaving the queue and being scheduled into the underlying driver; queuing latency includes long-term queuing latency. and short-term queuing delay ; Long queue delay The time it takes for a data packet to go from entering the queue to becoming the first packet in the queue is represented as: ; in, This is the current queue length. The average dequeue rate is measured within a sliding window. Short-term queuing delay The time it takes for a data packet to go from becoming the head of the queue to being dequeued is obtained by measuring the dwell time of the head packet in the current queue. Transmission delay Transmission delay is the time it takes for a data packet to travel from dequeue to successful transmission to the receiver, constrained by the transmission capacity of the wireless channel. The estimation uses a sliding window method to measure the average interval between packets output from the network layer queue. The length of the sliding window needs to cover at least two burst cycles from the sending end. Total delay of data packets transmitted in the wireless link , is represented as: 。 5. The dynamic buffer management method for Wi-Fi networks according to claim 4, characterized in that, Step S3 describes how, on the router side, for the Wi-Fi network, the available bandwidth of the current wireless link is inferred and estimated by analyzing the frame interaction behavior of the Wi-Fi MAC layer. Specifically: The router is defined to send data frames using A-MPDU method, and negotiates with the receiver to determine the maximum allowed number of frames in a batch as M. The actual number of aggregated frames sent by the router is denoted as b. If there is no data backlog in the router's buffer, the router sends aggregated frames with the actual batch size b. The receiver acknowledges the entire batch of data through Block ACK, and the router determines the batch size based on the ACK interval. Calculate instantaneous dequeue rate : ; in, This refers to the number of bits per frame. When there is a backlog of data in the buffer, i.e. At that time, the instantaneous outgoing rate That is, the link bandwidth; When there is no data backlog in the router buffer, i.e. Calculate the theoretical ACK interval when the batch is full. Since the mean ACK interval for different batch sizes has a linear relationship with a slope of S / R, the theoretical ACK interval... Represented as: ; Where R is the physical layer transmission bit rate; Furthermore, based on the theoretical ACK interval Estimate the available bandwidth of the current wireless link. The calculation formula is: 。 6. The dynamic buffer management method for Wi-Fi networks according to claim 5, characterized in that, Step S4 involves dynamically calculating the bandwidth-delay product based on the transmission delay and the available bandwidth of the current wireless link, and setting a dynamic threshold for a three-level buffer that updates in real time with changes in the link state based on this bandwidth-delay product; specifically: Based on transmission delay and the available bandwidth of the current wireless link Dynamically calculate the bandwidth-delay product of the wireless link at the current moment. The calculation formula is: ; The bandwidth-delay product of the wireless link at the current moment Three levels of dynamic thresholds are set to distinguish different degrees of buffer congestion: ; ; ; in, This is a preset proportional coefficient. , .
7. The dynamic buffer management method for Wi-Fi networks according to claim 6, characterized in that, The buffer control mechanism in step S5 includes: Length of the buffer queue Compared with the dynamic threshold of the three-level buffer, when the buffer queue length exceeds the first-level threshold but does not reach the second-level threshold, i.e. An acceleration flag is added to the data packet to prompt the sender to increase the sending rate; when the buffer queue length exceeds the second-level threshold but does not reach the third-level threshold, i.e. Explicit congestion notification flags are added to data packets to notify the sender to reduce the transmission rate; when the buffer queue length exceeds the third-level threshold, i.e. The router uses a weighted random early detection strategy to probabilistically discard unlabeled data packets.
8. The dynamic buffer management method for Wi-Fi networks according to claim 6, characterized in that, The buffer control mechanism in step S5 also includes: Determine the total delay of the transmission of the tagged data packets in the wireless link. Is it within the preset maximum end-to-end allowed latency limit? inside, if The router uses a weighted random early detection strategy to probabilistically discard unlabeled data packets; The preset proportional coefficient satisfies .
9. A dynamic buffer management system for Wi-Fi networks, characterized in that, include: The packet classification module is used to identify the flow to which the packets entering the router belong using access control lists, and to mark the packets that meet the characteristics of delay-sensitive flows. The packet delay prediction module is used to calculate the total delay of the transmission of tagged packets in the wireless link, including long queuing delay, short queuing delay and transmission delay; The wireless link bandwidth estimation module is used on the router side to infer and estimate the available bandwidth of the current wireless link by analyzing the frame interaction behavior of the Wi-Fi MAC layer. The dynamic threshold setting module is used to dynamically calculate the bandwidth-delay product based on the transmission delay and the available bandwidth of the current wireless link, and set a three-level buffer dynamic threshold that is updated in real time with changes in the link status based on the bandwidth-delay product. The buffer dynamic control module is used to set the buffer control mechanism based on the obtained three-level buffer dynamic threshold and the target delay of the labeled data packets, adjust the sending rate of data packets entering the router, or, based on a weighted random early detection strategy, probabilistically discard unlabeled data packets.
10. A non-transitory computer-readable storage medium, characterized in that, It stores computer instructions that cause the computer to execute the dynamic buffer management method for Wi-Fi networks as described in any one of claims 1-8.