CN115412497B - Performance optimization method of BBR congestion control algorithm - Google Patents
Performance optimization method of BBR congestion control algorithm Download PDFInfo
- Publication number
- CN115412497B CN115412497B CN202211352738.5A CN202211352738A CN115412497B CN 115412497 B CN115412497 B CN 115412497B CN 202211352738 A CN202211352738 A CN 202211352738A CN 115412497 B CN115412497 B CN 115412497B
- Authority
- CN
- China
- Prior art keywords
- bandwidth
- point
- bottleneck link
- flow
- state
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0803—Configuration setting
- H04L41/0823—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/25—Flow control; Congestion control with rate being modified by the source upon detecting a change of network conditions
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/30—Flow control; Congestion control in combination with information about buffer occupancy at either end or at transit nodes
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention relates to the technical field of computer networks, and discloses a performance optimization method of a BBR congestion control algorithm.
Description
Technical Field
The invention relates to the technical field of computer networks, in particular to a performance optimization method of a BBR congestion control algorithm.
Background
Network congestion can cause network performance degradation and even collapse, and congestion control is an effective way to solve this problem. With the development of network infrastructure, the shortcomings of the conventional congestion control algorithm based on packet loss are continuously revealed, so that a new congestion control algorithm more suitable for the current network development is urgently needed. The BBR congestion control algorithm aims to maintain a high sending rate while keeping low latency, and has been deployed in a number of practical application scenarios, achieving significant performance improvements.
The BBR congestion control algorithm does not rely on packet loss, but adjusts the sending rate by periodically estimating the bottleneck link bandwidth and round-trip propagation time. Due to its excellent performance, it is gradually replacing the traditional congestion control algorithm and is widely used globally. Therefore, optimization of the performance of the BBR congestion control algorithm becomes crucial. The existing optimization methods are mainly divided into the following methods:
(1) Throughput: in order to meet the requirement of high-speed mass data transmission of the current network, the throughput is an important index for measuring the performance of a congestion control algorithm. In a bandwidth detection stage (ProbeBW stage) of the BBR congestion control algorithm, the BBR congestion control algorithm first increases a sending rate according to a gain of 1.25 to detect whether there is a surplus available bandwidth resource, then rapidly decreases the sending rate according to a gain of 0.75 to empty a queue generated in a buffer in the previous stage, and then smoothly transmits 6 Round-Trip times (RTT) with 8 Round-Trip times as a cycle, and the cycle is repeated. However, a cycle period of 8 round trip times may result in the BBR congestion control algorithm not being able to respond quickly to the increase in bandwidth, resulting in a waste of bandwidth resources and a decrease in average throughput. Reducing the time for smooth transmission from 6 round trip times to a random value less than 6 round trip times can reduce the length of the cycle period, more sensitively capture the increase of bandwidth resources, and is beneficial to the improvement of average throughput and bandwidth utilization rate. On the other hand, in the delay detection stage (ProbeRTT stage) of the BBR Congestion control algorithm, the BBR Congestion control algorithm sets a Congestion Window (cwnd) to 4 Maximum packet lengths (MSS) to detect the minimum round-trip time, and this emptying detection method may cause a large reduction and fluctuation in throughput. To alleviate this problem, congestion control algorithms, represented by BBR v2, reduce cwnd to 50% instead of draining directly.
(2) Fairness: fairness issues are mainly divided into two categories, namely inter-algorithm fairness issues and intra-algorithm fairness issues. The fairness problem among algorithms mainly refers to the problem of bandwidth allocation imbalance generated when the BBR congestion control algorithm and the congestion control algorithms such as Cubic coexist. In an environment with a high packet loss rate, when the BBR congestion control algorithm and the Cubic congestion control algorithm compete, the BBR congestion control algorithm often occupies more than 90% of bandwidth resources, so that the Cubic congestion control algorithm is hardly available. In an environment with a low packet loss rate, the Cubic congestion control algorithm can suppress the BBR congestion control algorithm to occupy most of the bandwidth because the Cubic congestion control algorithm can send excessive data to occupy a bottleneck link buffer until packet loss occurs. The fairness problem among algorithms needs to be solved by considering the competition situation of the BBR congestion control algorithm and the conventional congestion control algorithms such as Cubic, reno and the like in each common scene, so that the bandwidth allocation is approximately balanced. The fairness problem in the algorithm is mainly expressed as an imbalance problem of bandwidth allocation among a plurality of BBR streams with different round trip times. The long round-trip time BBR streams occupy most of the bandwidth, and even if the round-trip time difference between two BBR streams is small, the bandwidth imbalance is quite severe. The existing solutions improve round-trip time fairness by reducing the detection time of a BBR stream with a long round-trip time, and some solutions adjust the amount of data to be transmitted according to the round-trip time by multiplying a Bandwidth-Delay Product (BDP) by an adjustment factor related to the round-trip time when calculating the BDP. The adjustment factor of the BBR flow with long round-trip time is smaller than that of the BBR flow with short round-trip time, so that the reduction of the bandwidth-delay product of the BBR flow with long round-trip time is larger than that of the BBR flow with short round-trip time, and the round-trip time fairness is improved.
(3) And (4) retransmission: in the start phase (start phase), the BBR congestion control algorithm exponentially increases the sending rate to quickly detect the ceiling of the current link until it is detected that the rate increases by no more than 25% within 3 round-trip times, but in most cases, the 3 round-trip times are too long, which not only generates a large number of queues in the buffer but also causes packet loss, so the start phase should be ended earlier, the generation of queues in the buffer is reduced, and retransmission is reduced.
(4) Convergence rate: in a practical application environment, frequent joining and exiting of flows often exist in network links. Ideally, when a flow is added or withdrawn, the coexisting flows can be converged to the state of the average bandwidth relatively quickly, so that network fluctuation and throughput variation generated when the flow is added or withdrawn can be reduced to the maximum extent, the overall delay is reduced, and the overall bandwidth utilization rate is improved.
Due to the rapid development of the internet, the traditional congestion control algorithm cannot adapt to the current needs of network infrastructure and data transmission, and although the BBR congestion control algorithm proposed in recent years achieves more remarkable performance improvement, in some specific scenarios, many problems still exist. In a network environment with relatively limited cache resources, the BBR congestion control algorithm periodically injects excessive data into a buffer area to form more persistent queues, so that the packet loss rate is higher and more network resources are required for retransmission. In an environment with relatively abundant cache resources, a small dynamic gain factor (paging gain) results in a low bandwidth change rate, and a plurality of BBR streams sharing a bottleneck link bandwidth take a long time to converge to an ideal state of averaging the bottleneck link bandwidth. Meanwhile, in an actual network environment, a flow is often added or withdrawn, the faster convergence speed can reduce the oscillation of the network, the overall delay of the network is reduced, and the link utilization rate is improved. The existing solutions mainly adjust the congestion window through a heuristic algorithm to optimize the performance, reduce retransmission to a certain extent, and improve the fairness among multiple BBR streams, but these solutions cannot effectively improve the convergence speed, and cannot perform adaptive adjustment according to the cache resource condition in the network and the network congestion degree.
Disclosure of Invention
In order to solve the above technical problems, the present invention provides a performance optimization method for BBR congestion control algorithm, which can adaptively adjust the value of gain according to the size of the network bottleneck link buffer and the network congestion condition on the basis of modeling the convergence behavior of BBR flow in the ProbeBW stage, thereby effectively increasing the convergence speed and reducing retransmission and average delay.
In order to solve the technical problems, the invention adopts the following technical scheme:
a performance optimization method of a BBR congestion control algorithm comprises the following steps:
s1, behavior modeling, specifically comprising:
step S11: in the bandwidth detection stage of the BBR congestion control algorithm, the bandwidths obtained from the bottleneck link bandwidth by the a and B streams sharing the bottleneck link bandwidth are respectively represented in the convergence domain: in the convergence domain, the horizontal axis represents the bandwidth of the a stream divided from the bottleneck link bandwidth, the vertical axis represents the bandwidth of the B stream divided from the bottleneck link bandwidth, and the bottleneck link bandwidth is marked as
The initial bandwidths of the A and B streams are recorded as
And
obtaining a state one, which is represented as a point in the convergence domain
Point of contact
Has the coordinates of
(ii) a In the convergence domain, the values of the,
the line represents a constraint that all flows share the bottleneck link bandwidth,
for the size of the bottleneck link buffer,
the line represents the bottleneck link bandwidth plus the bottleneck link bufferTotal available area boundary of the conflict area;
wire and
the area between the lines represents the buffer of the bottleneck link,
the line represents the ideal state of multiple flows equally dividing the bottleneck link bandwidth;
step S12: the A flow carries out bandwidth detection to obtain a state two, and the state two is represented as a point in a convergence domain
Point of contact
Has the coordinates of
Wherein g is a dynamic gain coefficient when each stream is subjected to bandwidth detection;
step S13: and after the bandwidth detection of the A flow is finished, queue emptying is required to be carried out, and a state III is obtained and is represented as a point in a convergence domain
I.e. A-flow queue emptying line and
intersection of lines, points
Has the coordinates of
The emptying line of the A flow queue is the origin and the point
The connecting line of (1);
step S14: b flow carries out bandwidth detection to obtain a state four, and the state four is represented as a point in a convergence domain
Point of contact
Having coordinates of
；
Step S15: and after the bandwidth detection of the B flow is finished, queue emptying is required to be carried out, and a state five is obtained and is represented as a point in a convergence domain
I.e. B-flow queue emptying line and
intersection of lines, points
Has the coordinates of
(ii) a The emptying line of the B flow queue is an origin and a point
The connecting line of (1);
s2, solving an optimization problem, which specifically comprises the following steps: in order to converge the a and B flows to the ideal state of sharing the bottleneck link bandwidth, a minimization point is needed
And the balance point
Distance between, equilibrium point
Is B = A line and
the intersection point of the lines and the constraint condition are that the A flow and the B flow cannot exceed the buffer area of the bottleneck link in the bandwidth detection process, so that the optimization problem is obtained
Solving by adopting a Lagrange multiplier method to obtain a dynamic gain coefficient
；
Step S3, self-adaptive control, which specifically comprises the following steps: when network congestion control is actually performed, the bottleneck link is controlled by
Sharing of each stream, dynamic gain factor
；
Is a first
The transmission rate of the individual streams is,
。
further, in step S3, if it is detected that packet loss occurs, the dynamic gain factor is adjusted
Multiplying by an adjustment factor
：
Compared with the prior art, the invention has the beneficial technical effects that:
the behavior of the BBR congestion control algorithm in the Probe BW stage is modeled, the value problem of the dynamic gain coefficient is converted into an optimization problem, and the optimization problem is solved; on the basis, the invention defines a regulating factor
The self-adaptive adjustment method according to the network cache resource and the network congestion condition is designed and realized. Compared with the original BBR, the invention can improve the convergence rate by 21%, reduce the retransmission by 73% and reduce the average delay by 46%.
Drawings
FIG. 1 is an optimization process of the BBR congestion control algorithm of the present invention;
FIG. 2 is a schematic diagram of the convergence behavior of the BBR congestion control algorithm of the present invention;
fig. 3 illustrates an adaptive control method according to the present invention.
Detailed Description
A preferred embodiment of the present invention will be described in detail with reference to the accompanying drawings.
The state machine adopted in the BBR congestion control algorithm is divided into four stages, namely a start stage (start stage), an empty stage (Drain stage), a bandwidth detection stage (ProbeBW stage), and a delay detection stage (ProbeRTT stage).
The invention provides a performance optimization method of a BBR congestion control algorithm by modeling the behavior of a bandwidth detection stage (Probe BW stage) of the BBR congestion control algorithm, and the performance optimization method is mainly divided into three stages as shown in figure 1.
The first phase is a behavior modeling phase. Modeling the behavior of a BBR congestion control algorithm in a Probe BW stage, respectively representing bandwidths obtained by an A flow and a B flow sharing the bottleneck link bandwidth from the bottleneck link bandwidth in a convergence domain, defining constraint conditions which need to be met when the flows share the bottleneck link bandwidth, and representing an ideal state when the flows converge to the equalized bottleneck link bandwidth. On the basis, the process of respectively detecting the bandwidth and emptying the queue in the buffer by the flow A and the flow B is described.
The second phase is the optimization problem solving phase. In the invention, it is expected that the stream A and the stream B can converge to an ideal state of sharing bandwidth as soon as possible, but in the detection process, the overflowed excessive data can form a queue in the buffer area, and it is required to ensure that the excessive data does not exceed the buffer area in the detection process to cause packet loss. In the convergence domain, as shown in fig. 2, it can be described that it is desirable that the points representing the bandwidth allocation states of the a-stream and the B-stream be as close as possible to
Line, not exceeding during probing at the same time
And dotted line
An enclosed area. The problem can be converted into a typical optimization problem and solved by using a Lagrange multiplier method.
The third stage is an adaptive control stage. Detecting the congestion state of the network, if detecting that packet loss is generatedMultiplying by an adjustment factor on the basis of the solution obtained in the second stage
，
The value of (d) is related to the current sending rate, and if no packet loss is detected, the value of gain is the solution obtained at the second stage.
The flow of the performance optimization method is specifically as follows.
(1) Behavior modeling phase
As shown in fig. 2, in the convergence domain, the horizontal axis represents the bandwidth of the a-stream divided from the bottleneck link bandwidth, and the vertical axis represents the bandwidth of the B-stream divided from the bottleneck link bandwidth. Total bottleneck Link Bandwidth as
The initial bandwidth of the A and B streams are denoted as
And
then this state (denoted as state one) can be represented as a point in the convergence domain
Point of contact
Has the coordinates of
。
In the figure
The line represents the constraint that all flows share the bottleneck link bandwidth, which is shown with the upper right dotted line
The area in between represents the buffer of the bottleneck link,
the line represents the ideal situation where multiple flows share the bottleneck link bandwidth. In the ProbeBW stage of the BBR congestion control algorithm, a dynamic gain factor (paging gain) is multiplied to detect whether available bandwidth exists when calculating the sending rate, and for convenience, the dynamic gain factor is abbreviated as
。
It is not assumed that the A-flow is detected first and then changes to state two, i.e. the point in the figure
The coordinates in the convergence domain are
。
At this time, since a queue is generated due to the excessive data injected into the buffer, the queue needs to be emptied after the probe is finished, state three is obtained, and then the process returns to
Points on the line
Point of contact
The coordinates of (a) are:
。
then, the B stream enters the detection stage and is multiplied by the dynamic gain factor
In the form ofState four, the slave point in the convergence domain
Become a point
Point of contact
The coordinates are expressed as:
。
it is worth mentioning that B-stream detection is not at the point
On the basis of the ordinate by
. Since the estimated bottleneck link bandwidth (BtlBw) for a flow is a maximum over 10 round-trip times, the point
Has a ordinate smaller than the point
Thus does not trigger an update of the estimated bottleneck link bandwidth (BtlBw) of the B-stream, and so the point
On the ordinate is at the point
Multiplication on the basis of the ordinate
The result of (1). At this point a queue is created in the buffer, and the queue still needs to be drained, getting state five, and returning to
Points on the line
Point of contact
Has the coordinates of
。
(2) Optimization problem solving phase
In the step (1), the behavior of the BBR congestion control algorithm in the ProbeBW stage is analyzed and modeled, and in order to converge to an ideal state as soon as possible without exceeding a buffer zone, the BBR congestion control algorithm is converted into an optimization problem and solved in the stage. Considering the points obtained after the A flow and the B flow are respectively subjected to bandwidth detection once
The goal is to minimize this point and the balance point
The constraint condition is to ensure that the stream A and the stream B do not exceed the buffer area during the detection process. For the above optimization objectives and constraint conditions, the corresponding optimization problems can be obtained, and after necessary simplification, the optimization problems are as follows:
in order to solve the optimization problem, a lagrangian multiplier method is adopted, and then the lagrangian function of the optimization problem is as follows:
the KKT condition is:
solving the KKT condition to obtain:
coefficient of dynamic gain
；
Lagrange multiplier
；
Dynamic gain coefficient of the event
Should take
This value is related to the bottleneck link buffer size and the currently allocated bandwidth,
is the bottleneck link buffer size.
(3) Adaptive control phase
Assume bottleneck link
Has a bandwidth of
From
The streams are shared. Order to
Is shown as
The number of streams is such that,
to represent
The round-trip time of (a) is,
to represent
The transmission rate of (1) is, in an ideal state, that
As shown. Determining a dynamic gain coefficient according to the solution obtained in (2) in combination with the size of the buffer
Is taken as the value of (2), then the dynamic gain coefficient
As shown in fig. 3. When there are n streams, the n streams may be divided into an a stream and a stream other than a, so the bandwidth detection behaviors of the two streams and the n streams are identical, so the solution obtained in (2) may be applied to the case where there are n streams.
Furthermore, dynamic gain factor
The value is also influenced by the congestion state of the network, and if the network is detected to have packet loss, the dynamic gain coefficient is used for not increasing the congestion
Should be further reduced. In the invention, an adjusting factor less than 1 is multiplied on the basis of the value obtained by original calculation
，
The value should take into account the transmission rate at that time
Relative to the maximum bandwidth that has not expired
The magnitude of the drop. Transmission rate at this time
The smaller, the more heavily the degree of network congestion is indicated,
the smaller the value of (c). Therefore, will
The values of (A) are set as:
wherein
And
is the maximum and minimum transmission rates observed during an estimated bottleneck link bandwidth (BtlBw) validity period, which is typically 10 round-trip times (RTTs).
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein, and any reference signs in the claims are not intended to be construed as limiting the claim concerned.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
Claims (2)
1. A performance optimization method of a BBR congestion control algorithm comprises the following steps:
s1, behavior modeling, specifically comprising:
step S11: in the bandwidth detection stage of the BBR congestion control algorithm, the bandwidths obtained from the bottleneck link bandwidth by the a and B streams sharing the bottleneck link bandwidth are respectively represented in the convergence domain: in the convergence domain, the horizontal axis represents the bandwidth of the a stream divided from the bottleneck link bandwidth, the vertical axis represents the bandwidth of the B stream divided from the bottleneck link bandwidth, and the bottleneck link bandwidth is marked as
The initial bandwidths of the A and B streams are recorded as
And
to obtainTo State one, the State one is represented as a point in the converged Domain
Point of contact
Has the coordinates of
(ii) a In the converged domain, the sum of the coefficients,
the line represents the constraint that all flows share the bottleneck link bandwidth,
for the bottleneck link buffer size,
the line represents the bottleneck link bandwidth plus the total available area boundary of the bottleneck link buffer;
wire and
the area between the lines represents the buffer of the bottleneck link,
the line represents the ideal state of multiple flows equally dividing the bottleneck link bandwidth;
step S12: the A flow carries out bandwidth detection to obtain a state two, and the state two is expressed as a point in a convergence domain
Point of contact
Has the coordinates of
Wherein g is a dynamic gain coefficient when each stream is subjected to bandwidth detection;
step S13: and after the bandwidth detection of the A flow is finished, queue emptying is required to be carried out, and a state III is obtained and is represented as a point in a convergence domain
I.e. A-flow queue drain and
intersection of lines, points
Has the coordinates of
The emptying line of the A flow queue is the origin and the point
The connecting line of (1);
step S14: b flow carries out bandwidth detection to obtain a state four, and the state four is represented as a point in a convergence domain
Point of contact
The coordinates are
；
Step S15: and after the bandwidth detection of the B flow is finished, queue emptying is required to be carried out, and a state five is obtained and is represented as a point in a convergence domain
I.e. B-flow queue emptying line and
intersection of lines, points
Has the coordinates of
(ii) a The emptying line of the B flow queue is an origin and a point
The connecting line of (1);
s2, solving an optimization problem, which specifically comprises the following steps: in order to converge the A and B streams to the ideal state of sharing the bottleneck link bandwidth, a minimum point is needed
And the balance point
Distance between, equilibrium point
Is B = A line and
the intersection point of the lines is constrained by the condition that the A flow and the B flow cannot exceed the buffer area of the bottleneck link in the bandwidth detection process, so that the optimization problem is obtained
Solving by adopting Lagrange multiplier method to obtain dynamic gain coefficient
；
Step S3, self-adaptive control, which specifically comprises the following steps: when network congestion control is actually carried out, the bottleneck link is controlled by
Sharing of individual streams, dynamic gain factor
；
Is as follows
The transmission rate of the individual streams is,
。
2. the method of performance optimization of a BBR congestion control algorithm of claim 1, wherein: in step S3, if the packet loss is detected, the dynamic gain coefficient is adjusted
Multiplying by an adjustment factor
：
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211352738.5A CN115412497B (en) | 2022-11-01 | 2022-11-01 | Performance optimization method of BBR congestion control algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211352738.5A CN115412497B (en) | 2022-11-01 | 2022-11-01 | Performance optimization method of BBR congestion control algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115412497A CN115412497A (en) | 2022-11-29 |
| CN115412497B true CN115412497B (en) | 2022-12-30 |
Family
ID=84168413
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211352738.5A Active CN115412497B (en) | 2022-11-01 | 2022-11-01 | Performance optimization method of BBR congestion control algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115412497B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115865827B (en) * | 2023-02-09 | 2023-05-23 | 合肥综合性国家科学中心人工智能研究院(安徽省人工智能实验室) | Event-driven cross-layer end network cooperative congestion control method |
| CN116155825B (en) * | 2023-04-04 | 2023-06-30 | 山东大学 | Optimization method for BBR congestion control algorithm data retransmission |
| CN116915706B (en) * | 2023-09-13 | 2023-12-26 | 合肥综合性国家科学中心人工智能研究院(安徽省人工智能实验室) | Data center network congestion control method, device, equipment and storage medium |
| CN117499317B (en) * | 2023-12-28 | 2024-04-12 | 苏州元脑智能科技有限公司 | Link congestion control method and device, storage medium and electronic equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110572333A (en) * | 2019-10-25 | 2019-12-13 | 中南大学 | transmission rate adjusting method and congestion control algorithm in bandwidth detection stage |
| CN111683393A (en) * | 2020-05-25 | 2020-09-18 | 华中科技大学 | Adaptive congestion control method for dynamically adjusting gain coefficient |
| CN113518040A (en) * | 2021-04-30 | 2021-10-19 | 东北大学 | Multipath coupling congestion control method for delay sensitive service |
| US11398984B1 (en) * | 2018-09-24 | 2022-07-26 | Reservoir Labs, Inc. | Systems and methods for quality of service (QoS) based management of bottlenecks and flows in networks |
| CN114945004A (en) * | 2022-05-23 | 2022-08-26 | 国网信息通信产业集团有限公司 | Congestion control method and device, electronic equipment and storage medium |
-
2022
- 2022-11-01 CN CN202211352738.5A patent/CN115412497B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11398984B1 (en) * | 2018-09-24 | 2022-07-26 | Reservoir Labs, Inc. | Systems and methods for quality of service (QoS) based management of bottlenecks and flows in networks |
| CN110572333A (en) * | 2019-10-25 | 2019-12-13 | 中南大学 | transmission rate adjusting method and congestion control algorithm in bandwidth detection stage |
| CN111683393A (en) * | 2020-05-25 | 2020-09-18 | 华中科技大学 | Adaptive congestion control method for dynamically adjusting gain coefficient |
| CN113518040A (en) * | 2021-04-30 | 2021-10-19 | 东北大学 | Multipath coupling congestion control method for delay sensitive service |
| CN114945004A (en) * | 2022-05-23 | 2022-08-26 | 国网信息通信产业集团有限公司 | Congestion control method and device, electronic equipment and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| BBR拥塞控制算法延迟及带宽探测优化;黄宏平;《计算机与现代化》;20221031;全文 * |
| Improving Inter-Protocol Fairness Based on Estimated Behavior of Competing Flows;Satoshi Utsumi;《IEEE》;20220722;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115412497A (en) | 2022-11-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115412497B (en) | Performance optimization method of BBR congestion control algorithm | |
| CN114145001B (en) | Rate optimized congestion management | |
| CN105610820B (en) | A kind of jamming control method and device based on MPTCP | |
| CN113746748B (en) | Explicit congestion control method in named data network | |
| JP5512699B2 (en) | Method for controlling flow in a packet switching network | |
| MXPA05006550A (en) | Flow control in network devices. | |
| WO2020253488A1 (en) | Method and device for congestion control, communication network, and computer storage medium | |
| CN110730469B (en) | Method for predicting bandwidth based on extended Kalman wireless network and congestion control thereof | |
| CN104994031A (en) | Active queue self-adaptive management method ASRED | |
| CN109698925B (en) | Real-time video congestion control method and device based on data driving | |
| Baklizi et al. | Dynamic stochastic early discovery: a new congestion control technique to improve networks performance | |
| CN116963182A (en) | Time delay optimal task unloading method and device, electronic equipment and storage medium | |
| CN109787861B (en) | Network data delay control method | |
| Abrantes et al. | Explicit congestion control algorithms for time varying capacity media | |
| Shan et al. | Micro-burst in data centers: Observations, implications, and applications | |
| CN116319569A (en) | Network parameter updating method, network parameter updating device, medium and electronic equipment | |
| Sun et al. | RFBBR: a RTT faireness awared algorithm based on BBR | |
| CN114915600B (en) | BBRv2 congestion control method under deep buffer | |
| CN116389375A (en) | Network queue management method, device and router for live video stream | |
| JP2016213581A (en) | Information processing system, control method of information processing system and receiver | |
| TWI423623B (en) | Network buffer management methods and computer program products | |
| Fridovich-Keil et al. | A model predictive control approach to flow pacing for TCP | |
| Vyakaranal et al. | Performance evaluation of TCP using AQM schemes for congestion control | |
| Hu et al. | An active queue management mechanism for minimizing queueing delay | |
| JP2005184494A (en) | Data communication management method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |