Buffer management method and system based on priority feedback and medium
By setting the queue threshold and feedback threshold in the switch buffer, the problem of low-priority queue preemption of high-priority queues is solved, ensuring the stable transmission of high-priority traffic and the basic requirement of low-priority traffic, achieving the balance of priority isolation and traffic absorption.
Patent Information
- Application Number
- CN202510800886.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-16
- Publication Date
- 2025-08-15
AI Technical Summary
In the prior art, in multi-priority queue transmission, low-priority queues are prone to preempt buffer space of high-priority queues, resulting in an increase in packet loss rate of high-priority traffic and a lack of fine threshold adjustment strategies.
By setting queue thresholds, feedback thresholds and minimum guarantee thresholds for low-priority queues in the switch buffer, the buffer management strategy is optimized in combination with preset algorithms, including initial state, feedback state and minimum guarantee state, ensuring priority transmission of high-priority traffic and limiting the occupation of low-priority traffic.
It realizes the stable transmission of high-priority traffic under high load conditions and the basic demand guarantee of low-priority traffic, reduces the packet loss rate and buffer occupancy rate of high-priority traffic, and improves the priority isolation and traffic absorption capacity.
Smart Images
Figure CN120499137A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of computer network systems, and in particular to a buffer management method, system and medium based on priority feedback. Background Art
[0002] In modern data center networks, the diversity of traffic types and their varying Quality of Service (QoS) requirements are becoming increasingly prominent. To optimize transmission performance and ensure QoS, data center networks widely employ various strategies, such as traffic scheduling. A core approach is to prioritize traffic based on its importance. To effectively support this mechanism, network switches typically configure multiple priority queues on each output port, enabling different types of traffic to be allocated resources based on their priority, ensuring that high-priority traffic receives preferential processing. However, while the introduction of multiple priority queues improves scheduling flexibility, it also places higher demands on buffer management. On the one hand, buffer management strategies must fully utilize their fundamental function—temporarily absorbing bursts to reduce packet loss and maintain network stability. On the other hand, due to the inherent characteristics of the priority mechanism, the transmission requirements of high-priority traffic must be strictly guaranteed, while low-priority traffic should not encroach on the buffer resources of high-priority traffic. This means that isolation between different priority queues must be ensured, but also that low-priority traffic should not fall into a "starvation" state, where its transmission requirements remain unmet for extended periods.
[0003] Among them, the ABM solution is the latest technical solution to solve the priority isolation problem in recent years. The ABM solution first considers the phenomenon that low-priority traffic will preempt high-priority traffic when the traditional DT solution is expanded to the scenario of multi-priority traffic transmission, which is the priority isolation problem; then the ABM solution proposes that the reason why priority isolation is affected is that the queue threshold of each priority queue depends not only on the configuration parameters of the current queue, but also on the number of queues of other priorities in a congested state. Therefore, ABM attempts to adjust the threshold by introducing the number of congested queues into the queue threshold calculation formula. It takes into account the priority isolation problem during the execution of the buffer management solution and introduces the number of congested queues to try to control the threshold.
[0004] However, this technology only relies on the number of congested queues after reaching a steady state, and cannot adjust the threshold according to the transmission of packets of different priorities in a transient state. Therefore, although it can alleviate the priority inversion phenomenon in a steady state, packets in low-priority queues will continue to occupy the buffer space of high-priority queues during transmission. In other words, this threshold control is lagging and its threshold adjustment method is not precise enough. Therefore, a buffer management strategy is needed that can feedback the threshold when the low-priority queue is about to reach congestion to prevent the excessive number of low-priority congested queues from affecting the buffer occupancy of high-priority queues. Summary of the Invention
[0005] The present application provides a buffer management method, system, and medium based on priority feedback, to at least solve the problem in related technologies of providing feedback on a threshold when a low-priority queue is about to reach congestion, thereby avoiding the problem that an excessive number of low-priority congested queues affects the buffer occupancy of a high-priority queue.
[0006] In a first aspect, the present application provides a buffer management method based on priority feedback, the method comprising:
[0007] Assigning queue thresholds for each priority queue in the switch buffer, feedback thresholds for low-priority queues, and minimum guaranteed thresholds for low-priority queues based on a preset algorithm, wherein the priority queues include low-priority queues and high-priority queues;
[0008] A buffer state strategy of the buffer is executed based on the real-time data capacity of each priority queue in the buffer, wherein the buffer state strategy includes an initial state strategy, a feedback state strategy, and a minimum guarantee state strategy, wherein:
[0009] The initial state strategy is a state in which the length of the low priority queue does not reach the feedback threshold;
[0010] The feedback status policy is triggered when the length of the low priority queue reaches the feedback threshold;
[0011] The minimum guarantee state policy is triggered to enter when the system is in the feedback state and the low priority queue reaches the minimum guarantee threshold.
[0012] Optionally, the initial state strategy includes:
[0013] When each priority queue arrives at the buffer, the length of the priority queue is judged: when the length of the priority queue exceeds the queue threshold, it is directly discarded; when the length of the priority queue does not exceed the queue threshold, the priority queue is accepted.
[0014] Optionally, the feedback status strategy includes:
[0015] If the low priority queue reaches the buffer, it is directly discarded;
[0016] If the high priority queue arrives at the buffer zone, the length of the high priority queue is judged: when the length of the high priority queue exceeds the queue threshold, it is directly discarded; when the length of the high priority queue does not exceed the queue threshold, the high priority queue is received and the feedback threshold is lowered.
[0017] Optionally, the high priority queue is received and the feedback threshold is lowered, and the reduction range of the feedback threshold is the length of the high priority queue.
[0018] Optionally, under the minimum guarantee status strategy, the feedback threshold is the same as the minimum guarantee threshold, including:
[0019] If the low priority queue arrives at the buffer zone, the length of the low priority queue is judged: when the length of the low priority queue exceeds the minimum guarantee threshold, the low priority queue is directly discarded; when the length of the low priority queue does not exceed the minimum guarantee threshold, the low priority queue is received;
[0020] If the high priority queue reaches the buffer zone, the length of the high priority queue is judged: when the length of the high priority queue exceeds the queue threshold, it is directly discarded; when the length of the high priority queue does not exceed the queue threshold, the high priority queue is received and the feedback threshold is not lowered.
[0021] Optionally, assigning a queue threshold to each priority queue in the switch buffer based on a preset algorithm includes:
[0022] Based on a first preset algorithm, queue thresholds of the low-priority queue and the high-priority queue are obtained, where the first preset algorithm is:
[0023] T p (t) = α p ·(BQ(t))
[0024] Where p is the priority of the corresponding queue, α p The threshold configuration parameters for this priority queue are: B is the total buffer size of the switch, and Q(t) is the size of the buffer occupied at time t;
[0025] The queue threshold is assigned as a packet admission control threshold of the high-priority queue; and the queue threshold is assigned as a packet reference value of the low-priority queue.
[0026] Optionally, assigning a feedback threshold to a low-priority queue in a switch buffer based on a preset algorithm includes:
[0027] Based on a second preset algorithm, a feedback threshold of the low-priority queue is calculated and obtained, where the feedback threshold is a proportionally reduced value of the queue threshold. The second preset algorithm is:
[0028] T feedback (t) = β·T p (t)-△
[0029] Wherein, β is a feedback proportional factor, and its value range is (0, 1); Δ is a downward adjustment value of the feedback threshold of the low priority queue in the feedback state.
[0030] Optionally, assigning a minimum guaranteed threshold to a low-priority queue in a switch buffer based on a preset algorithm includes:
[0031] Based on a third preset algorithm, the minimum guaranteed threshold of the low priority queue is calculated and obtained. The third preset algorithm is:
[0032]
[0033] Where p is the priority of the corresponding queue, α p is the threshold configuration parameter of the priority queue, B is the total buffer size of the switch, ∑ p α p (p∈P, P is the set of all priorities) is the sum of the threshold configuration parameters corresponding to all priority queues.
[0034] In a second aspect, the present application provides a buffer management system based on priority feedback, the system comprising:
[0035] A threshold assignment module is used to assign queue thresholds of each priority queue in the switch buffer, feedback thresholds of low-priority queues, and minimum guarantee thresholds of low-priority queues based on a preset algorithm. The priority queues include low-priority queues and high-priority queues.
[0036] A policy switching module is configured to execute a buffer state policy of the buffer based on the real-time data capacity of each priority queue in the buffer, wherein the buffer state policy includes an initial state policy, a feedback state policy, and a minimum guarantee state policy, wherein:
[0037] The initial state strategy is a state in which the length of the low priority queue does not reach the feedback threshold;
[0038] The feedback status policy is triggered when the length of the low priority queue reaches the feedback threshold;
[0039] The minimum guarantee state policy is triggered to enter when the system is in the feedback state and the low priority queue reaches the minimum guarantee threshold.
[0040] In a third aspect, the present application provides a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the buffer management method provided in the first aspect above.
[0041] Compared with related technologies, the buffer management method, system, and medium based on priority feedback provided by this application have at least the following technical effects:
[0042] By configuring and optimizing the queue thresholds of each priority queue in the switch buffer, the feedback thresholds of the low-priority queue, and the minimum guaranteed thresholds of the low-priority queue, buffer resources are rationally allocated according to the transmission requirements of queues of different priority levels. While maximizing the traffic absorption capacity, a buffer management strategy that isolates different priorities is ensured. High-priority traffic can use the initial threshold to control the admission of its packets to ensure priority transmission, while the admission of packets to low-priority queues is controlled using the feedback threshold to ensure that high-priority traffic is not preempted. At the same time, the minimum guaranteed threshold can also ensure the minimum performance requirements of low-priority traffic.
[0043] The details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0045] Figure 1 The flowchart of the buffer management method based on priority feedback is shown according to an exemplary embodiment.
[0046] Figure 2 The figure is a schematic diagram showing a switch queue during data transmission according to an exemplary embodiment.
[0047] Figure 3 The flowchart of a buffer management algorithm based on priority feedback is shown according to an exemplary embodiment.
[0048] Figure 4 The figure is a state transition diagram of a switch queue during data transmission according to an exemplary embodiment.
[0049] Figure 5 FIG. 4 is a diagram showing experimental results of packet loss rates at different priorities according to another exemplary embodiment.
[0050] Figure 6 is a diagram showing experimental results of buffer zone occupancy according to an exemplary embodiment. DETAILED DESCRIPTION
[0051] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is described and illustrated below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely used to explain this application and are not intended to limit this application. Based on the embodiments provided in this application, all other embodiments obtained by those of ordinary skill in the art without making any creative efforts are within the scope of protection of this application.
[0052] Obviously, the drawings described below are merely examples or embodiments of the present application. Those skilled in the art can, without inventive effort, apply the present application to other similar scenarios based on these drawings. Furthermore, it is also understood that, although the effort involved in such a development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, changes in design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as an insufficiency of the content disclosed in this application.
[0053] References to "embodiments" in this application mean that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it refer to independent or alternative embodiments that are mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described in this application may be combined with other embodiments unless there is a conflict.
[0054] Unless otherwise defined, the technical or scientific terms used in this application should have the ordinary meaning understood by a person of ordinary skill in the technical field to which this application belongs. The words "one", "a", "the" and the like used in this application do not indicate a limit on quantity and may indicate the singular or plural. The terms "include", "comprise", "have" and any variations thereof used in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units that are inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The word "multiple" used in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific order for the objects.
[0055] In related technologies, the ABM solution technology only relies on the number of congested queues after reaching a steady state, and cannot adjust the threshold according to the transmission of data packets of different priorities in a transient state. Therefore, although it can alleviate the priority inversion phenomenon in a steady state, the data packets in the low-priority queue will continue to occupy the buffer space of the high-priority queue during the transmission process. That is, this threshold control is lagging, and its threshold adjustment method is not precise enough. Therefore, a buffer management strategy is needed to be able to feedback the threshold when the low-priority queue is about to reach congestion, so as to avoid the excessive number of low-priority congested queues affecting the buffer occupancy of the high-priority queue.
[0056] Based on the above situation, an embodiment of the present invention provides a buffer management method, system and medium based on priority feedback, which are described in detail below with reference to specific embodiments and drawings.
[0057] Example 1
[0058] An embodiment of the present invention provides a buffer management method based on priority feedback. Figure 1 The flowchart of the buffer management method based on priority feedback is shown according to an exemplary embodiment. Figure 2 FIG. 1 is a schematic diagram of a switch queue during data transmission according to an exemplary embodiment. Figure 1-2 As shown, the method includes:
[0059] Step S101: assigning queue thresholds for each priority queue in a switch buffer, feedback thresholds for low-priority queues, and minimum guaranteed thresholds for low-priority queues based on a preset algorithm. The priority queues include low-priority queues and high-priority queues.
[0060] Among them, based on the first preset algorithm, the queue thresholds of the low-priority queue and the high-priority queue are obtained. For the high-priority queue, the initial threshold is directly used as the control threshold for data packet admission; for the low-priority queue, the initial threshold is used as a reference value for calculating the feedback threshold. The initial value of the feedback threshold is usually set to a certain ratio of the initial threshold. The first preset algorithm is:
[0061] T p (t) = α p ·(BQ(t))
[0062] Where p is the priority of the corresponding queue, α p The threshold configuration parameters for this priority queue are: B is the total buffer size of the switch, and Q(t) is the size of the buffer occupied at time t;
[0063] The queue threshold is assigned as a packet admission control threshold of the high-priority queue; and the queue threshold is assigned as a packet reference value of the low-priority queue.
[0064] Based on the second preset algorithm, the feedback threshold of the low-priority queue is calculated and obtained. The feedback threshold is mainly used to regulate the threshold of the low-priority queue. The feedback threshold is a proportional reduction value of the queue threshold. When the length of the low-priority queue reaches the feedback threshold, it indicates that the queue is gradually entering a congested state. At this time, the switch switches to the feedback state. In this state, data packets from the low-priority queue will be directly discarded when they arrive; at the same time, each time a high-priority queue enters the queue, the threshold of the low-priority queue will be lowered to prevent it from occupying the buffer zone that the high-priority queue deserves. The second preset algorithm is:
[0065] T feedback (t) = β·T p (t)-△
[0066] Wherein, β is a feedback proportional factor, and its value range is (0, 1); Δ is a downward adjustment value of the feedback threshold of the low priority queue in the feedback state.
[0067] Based on the third preset algorithm, the minimum guaranteed threshold of the low-priority queue is calculated. Although according to the nature of priority traffic, high-priority traffic should obtain more buffer resources to ensure its priority transmission, if high-priority traffic continues to arrive at the switch, the low-priority queue will be unable to obtain buffer resources for a long time and enter a starvation state. Therefore, the present invention sets a lower limit for the feedback interval to ensure minimum fairness. The third preset algorithm is:
[0068]
[0069] Where p is the priority of the corresponding queue, α p is the threshold configuration parameter of the priority queue, B is the total buffer size of the switch, ∑ p α p (p∈P, P is the set of all priorities) is the sum of the threshold configuration parameters corresponding to all priority queues;
[0070] Based on the three thresholds mentioned above, high-priority traffic can use the initial threshold to control the admission of its packets, while the threshold of the low-priority queue will slide between the feedback threshold and the minimum guarantee threshold, ensuring that high-priority traffic is not preempted while also ensuring the minimum performance requirements of low-priority traffic.
[0071] Step S102: executing a buffer state strategy of the buffer based on the real-time data capacity of each priority queue in the buffer, wherein the buffer state strategy includes an initial state strategy, a feedback state strategy, and a minimum guarantee state strategy, wherein:
[0072] The initial state strategy is a state in which the length of the low priority queue does not reach the feedback threshold;
[0073] The feedback status policy is triggered when the length of the low priority queue reaches the feedback threshold;
[0074] The minimum guarantee state policy is triggered to enter when the system is in the feedback state and the low priority queue reaches the minimum guarantee threshold.
[0075] Figure 3 The flowchart of a buffer management algorithm based on priority feedback is shown according to an exemplary embodiment. Figure 4 FIG. 1 is a state transition diagram of a switch queue during data transmission according to an exemplary embodiment. Figure 2-Figure 4 ,Based on the above three thresholds of the queue, the switch is in different states and performs state transition.
[0076] Specifically, in the embodiment of the present application, the initial state strategy includes:
[0077] When each priority queue arrives at the buffer, the length of the priority queue is judged: when the length of the priority queue exceeds the queue threshold, it is directly discarded; when the length of the priority queue does not exceed the queue threshold, the priority queue is accepted.
[0078] Specifically, in the embodiment of the present application, the feedback status strategy includes:
[0079] If the low priority queue reaches the buffer, it is directly discarded;
[0080] If the high priority queue reaches the buffer zone, the length of the high priority queue is judged: when the length of the high priority queue exceeds the queue threshold, it is directly discarded; when the length of the high priority queue does not exceed the queue threshold, the high priority queue is received and the feedback threshold is lowered, wherein the reduction range of the feedback threshold is consistent with the length of the high priority queue.
[0081] Specifically, in the embodiment of the present application, under the minimum guarantee state strategy, the feedback threshold is the same as the minimum guarantee threshold, including:
[0082] If the low priority queue arrives at the buffer zone, the length of the low priority queue is judged: when the length of the low priority queue exceeds the minimum guarantee threshold, the low priority queue is directly discarded; when the length of the low priority queue does not exceed the minimum guarantee threshold, the low priority queue is received;
[0083] If the high priority queue reaches the buffer zone, the length of the high priority queue is judged: when the length of the high priority queue exceeds the queue threshold, it is directly discarded; when the length of the high priority queue does not exceed the queue threshold, the high priority queue is received and the feedback threshold is not lowered.
[0084] Experimental Verification: To comprehensively evaluate the performance of the PFBM solution, we conducted simulation experiments comparing it with two state-of-the-art buffer management strategies for multi-priority queues: DT (delayed queue management) with multiple priorities and ABM (automated buffer management). DT (delayed queue management) with multiple priorities configures different parameters for each priority level, with the core concept being to dynamically adjust thresholds based on global buffer status. ABM, on the other hand, is a newer, more optimized buffer management strategy for multi-priority queues, taking into account factors such as queue draining time and isolation.
[0085] Network Configuration: In our simulation experiments, we used a leaf-spine network topology as the experimental infrastructure. To simulate the real-world scenario of a large-scale data center network, we configured eight spine switches and eight leaf switches, each connected to 32 servers. All links had a bandwidth of 100 Gbps and a link propagation delay of 1 μs. Switch buffers were sized to 9.6 KB per port per Gbps.
[0086] Workload: This workload simulates the query-response behavior of a distributed file system in a data center. Specifically, a server requests file data from different locations and aggregates the data to the requesting server. This request generates incast traffic, and the traffic size is controlled between 10% and 80% of the total buffer size.
[0087] The experimental results are as follows:
[0088] The experiments aim to evaluate the usage of buffers under the Incast workload.
[0089] Figure 5 FIG is an experimental result diagram of packet loss rate under different priorities according to another exemplary embodiment. Figure 5 As shown, first of all, in the experimental results of packet loss rate under different priorities, Figure 5 (a) and Figure 5 (b) illustrates the packet loss rate in the high priority queue and the low priority queue respectively. Figure 5 The results in (a) show that as the link load increases, the packet loss rate of high-priority traffic under the DT and ABM strategies shows a nearly linear growth trend. This phenomenon indicates that under high load conditions, these two strategies are unable to effectively protect high-priority traffic, causing its buffer resources to be preempted by low-priority traffic, resulting in significant packet loss. In contrast, the PFBM strategy fully demonstrates its ability to protect high-priority traffic, maintaining a low and stable packet loss rate as the load increases. Figure 5(b) reveals the impact of PFBM on low-priority traffic. The results show that the packet loss rate of low-priority traffic under PFBM also shows a linear growth trend with the increase of load. However, although PFBM sacrifices some performance of low-priority traffic in order to prioritize high-priority traffic, its overall packet loss rate of low-priority traffic is still lower than or close to the levels of DT and ABM strategies. This phenomenon can be explained from the following aspects: First, high-priority traffic usually has higher processing efficiency, can complete transmission quickly and release buffer resources, thereby indirectly alleviating the congestion pressure of low-priority queues; second, PFBM ensures that low-priority queues will not be "starved" due to excessive resource deprivation by setting a minimum fairness threshold. Therefore, although the packet loss rate of low-priority traffic has increased, the overall performance degradation is still within an acceptable range.
[0090] Figure 6 FIG is an experimental result diagram of the buffer occupancy rate according to an exemplary embodiment. Figure 6 As shown, in terms of buffer occupancy, Figure 6 (a) Comparison of the overall buffer occupancy of the three buffer management strategies under different load factors. The results show that both PFBM and ABM are able to maintain a relatively stable buffer occupancy, avoiding excessive buffer congestion caused by increased load. However, PFBM performs better overall, with a significantly lower buffer occupancy than ABM. Figure 6 (b) further demonstrates PFBM's performance in priority isolation. The results show that as traffic increases its buffer demand, PFBM effectively prevents low-priority traffic from preempting high-priority traffic. Specifically, under high load conditions, high-priority traffic consistently receives more buffer resources, while low-priority traffic's buffer usage is strictly limited. This feature ensures quality of service for high-priority traffic while maintaining PFBM's superior performance in priority isolation.
[0091] In summary, the priority feedback-based buffer management method provided by the embodiment of the present invention reasonably allocates buffer resources according to the transmission requirements of different priority queues through the configuration and optimization of the queue thresholds of each priority queue in the switch buffer, the feedback thresholds of the low-priority queues, and the minimum guarantee thresholds of the low-priority queues. While maximizing the traffic absorption capacity, it ensures the isolation of different priority levels in the buffer management strategy. High-priority traffic can use the initial threshold to control the access of its data packets to ensure priority transmission of high priority, while the access of data packets of low-priority queues is controlled by the feedback threshold, thereby ensuring that high-priority traffic is not preempted. At the same time, the minimum guarantee threshold can also ensure the minimum performance requirements of low-priority traffic.
[0092] Example 2
[0093] Embodiment 2 of the present invention provides a buffer management system based on priority feedback, the system comprising:
[0094] A threshold assignment module is used to assign queue thresholds of each priority queue in the switch buffer, feedback thresholds of low-priority queues, and minimum guarantee thresholds of low-priority queues based on a preset algorithm. The priority queues include low-priority queues and high-priority queues.
[0095] A policy switching module is configured to execute a buffer state policy of the buffer based on the real-time data capacity of each priority queue in the buffer, wherein the buffer state policy includes an initial state policy, a feedback state policy, and a minimum guarantee state policy, wherein:
[0096] The initial state strategy is a state in which the length of the low priority queue does not reach the feedback threshold;
[0097] The feedback status policy is triggered when the length of the low priority queue reaches the feedback threshold;
[0098] The minimum guarantee state policy is triggered to enter when the system is in the feedback state and the low priority queue reaches the minimum guarantee threshold.
[0099] Example 3
[0100] Embodiment 3 of the present invention provides a computer-readable storage medium having a program stored thereon. When the program is executed by a processor, the buffer management method provided in the first aspect described above is implemented.
[0101] The readable storage medium may include, but is not limited to, a portable disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical storage device, a magnetic storage device, or any suitable combination thereof.
[0102] In a possible implementation manner, the present invention may also be implemented in the form of a program product, which includes program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps of implementing the buffer management method in Example 1.
[0103] The program code for executing the present invention may be written in any combination of one or more programming languages, and may be executed entirely on the user device, partially on the user device, as an independent software package, partially on the user device and partially on a remote device, or entirely on the remote device.
[0104] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0105] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.
Claims
1. A buffer management method based on priority feedback, characterized in that: The method comprises: Assigning queue thresholds for each priority queue in the switch buffer, feedback thresholds for low-priority queues, and minimum guaranteed thresholds for low-priority queues based on a preset algorithm, wherein the priority queues include low-priority queues and high-priority queues; A buffer state strategy of the buffer is executed based on the real-time data capacity of each priority queue in the buffer, wherein the buffer state strategy includes an initial state strategy, a feedback state strategy, and a minimum guarantee state strategy, wherein: The initial state strategy is a state in which the length of the low priority queue does not reach the feedback threshold; The feedback status policy is triggered when the length of the low priority queue reaches the feedback threshold; The minimum guarantee state policy is triggered to enter when the system is in the feedback state and the low priority queue reaches the minimum guarantee threshold.
2. The buffer management method according to claim 1, wherein: The initial state strategy includes: When each priority queue arrives at the buffer, the length of the priority queue is judged: when the length of the priority queue exceeds the queue threshold, it is directly discarded; when the length of the priority queue does not exceed the queue threshold, the priority queue is accepted.
3. The management method according to claim 1, characterized in that: The feedback status strategy includes: If the low priority queue reaches the buffer, it is directly discarded; If the high priority queue arrives at the buffer zone, the length of the high priority queue is judged: when the length of the high priority queue exceeds the queue threshold, it is directly discarded; when the length of the high priority queue does not exceed the queue threshold, the high priority queue is received and the feedback threshold is lowered.
4. The buffer management method according to claim 3, characterized in that: The high priority queue is received and the feedback threshold is lowered, and the reduction range of the feedback threshold is the length of the high priority queue.
5. The buffer management method according to claim 1, wherein: Under the minimum guarantee status strategy, the feedback threshold is the same as the minimum guarantee threshold, including: If the low priority queue arrives at the buffer zone, the length of the low priority queue is judged: when the length of the low priority queue exceeds the minimum guarantee threshold, the low priority queue is directly discarded; when the length of the low priority queue does not exceed the minimum guarantee threshold, the low priority queue is received; If the high priority queue reaches the buffer zone, the length of the high priority queue is judged: when the length of the high priority queue exceeds the queue threshold, it is directly discarded; when the length of the high priority queue does not exceed the queue threshold, the high priority queue is received and the feedback threshold is not lowered.
6. The buffer zone management method according to claim 1, wherein: The queue threshold assigned to each priority queue in the switch buffer based on a preset algorithm includes: Based on a first preset algorithm, queue thresholds of the low-priority queue and the high-priority queue are obtained, where the first preset algorithm is: T p (t)=α p ·(B-Q(t)) Where p is the priority of the corresponding queue, α p The threshold configuration parameters for this priority queue are: B is the total buffer size of the switch, and Q(t) is the size of the buffer occupied at time t; The queue threshold is assigned as a packet admission control threshold of the high-priority queue; and the queue threshold is assigned as a packet reference value of the low-priority queue.
7. The buffer management method according to claim 1, wherein: The feedback threshold assigned to the low-priority queue in the switch buffer based on a preset algorithm includes: Based on a second preset algorithm, a feedback threshold of the low-priority queue is calculated and obtained, where the feedback threshold is a proportionally reduced value of the queue threshold. The second preset algorithm is: T feedback (t)=β·T p (t)-△ Wherein, β is a feedback proportional factor, and its value range is (0, 1); Δ is a downward adjustment value of the feedback threshold of the low priority queue in the feedback state.
8. The buffer management method according to claim 1, wherein: The minimum guaranteed threshold value assigned to the low priority queue in the switch buffer based on the preset algorithm includes: Based on a third preset algorithm, the minimum guaranteed threshold of the low priority queue is calculated and obtained. The third preset algorithm is: Where p is the priority of the corresponding queue, α p is the threshold configuration parameter of the priority queue, B is the total buffer size of the switch, ∑ p α p The sum of the threshold configuration parameters corresponding to all priority queues.
9. A buffer management system based on priority feedback, characterized in that: The system comprises: A threshold assignment module is used to assign queue thresholds of each priority queue in the switch buffer, feedback thresholds of low-priority queues, and minimum guarantee thresholds of low-priority queues based on a preset algorithm. The priority queues include low-priority queues and high-priority queues. A policy switching module is configured to execute a buffer state policy of the buffer based on the real-time data capacity of each priority queue in the buffer, wherein the buffer state policy includes an initial state policy, a feedback state policy, and a minimum guarantee state policy, wherein: The initial state strategy is a state in which the length of the low priority queue does not reach the feedback threshold; The feedback status policy is triggered when the length of the low priority queue reaches the feedback threshold; The minimum guarantee state policy is triggered to enter when the system is in the feedback state and the low priority queue reaches the minimum guarantee threshold.
10. A computer-readable storage medium, characterized in that A program is stored thereon, and when the program is executed by a processor, the buffer management method according to any one of claims 1 to 8 is implemented.