CN113518040A - Multipath coupling congestion control method for delay sensitive service - Google Patents
Multipath coupling congestion control method for delay sensitive service Download PDFInfo
- Publication number
- CN113518040A CN113518040A CN202110480669.5A CN202110480669A CN113518040A CN 113518040 A CN113518040 A CN 113518040A CN 202110480669 A CN202110480669 A CN 202110480669A CN 113518040 A CN113518040 A CN 113518040A
- Authority
- CN
- China
- Prior art keywords
- rtt
- sub
- max
- cwnd
- flow
- 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.)
- Granted
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
- H04L47/125—Avoiding congestion; Recovering from congestion by balancing the load, e.g. traffic engineering
-
- 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
- H04L47/28—Flow control; Congestion control in relation to timing considerations
- H04L47/283—Flow control; Congestion control in relation to timing considerations in response to processing delays, e.g. caused by jitter or round trip time [RTT]
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention provides a time delay sensitive service-oriented multipath coupling congestion control method, and relates to the technical field of transmission optimization application. In the state transfer process of the method, a BBR algorithm mechanism is used, and in a Probe BW stage, in order to meet fairness and balance congestion, data packets cached in a sending buffer area are distributed in proportion according to Max BW multiplied by Min RTT of each sub-flow measured currently to regulate and control competition of each sub-flow on bandwidth. When multiple sub-streams compete with each other on the bottleneck link, the bandwidth obtained by each sub-stream is proportional to the size of the buffer it occupies in the link queue. Therefore, the competition of each sub-flow for the bandwidth is regulated by adjusting the coefficient of cwnd. Compared with the traditional heuristic multipath coupling congestion control algorithm, the method of the invention reduces the round-trip delay and the packet loss rate, and is a session service transmission control mechanism which is more suitable for delay sensitive services.
Description
Technical Field
The invention relates to the technical field of transmission optimization application, in particular to a multi-path coupling congestion control method for delay sensitive services.
Background
Today, the demand for network services is no longer limited to traditional services such as voice and short messages, and the demand for real-time communication is provided in many service scenes, so that the time-delay sensitive conversation services become a new network service growth point.
Due to the end-to-end network bandwidth bottleneck effect and the delay of the transmission optimization technology, the delay-sensitive real-time session service is generally limited to the enterprise private network, and although the WeChat telecommunication is widely applied, the technical scheme is not open and mature enough, and the delay-sensitive session service is difficult to support effectively. The end-to-end service transmission quality not only depends on an access network, but also depends on an end-to-end transmission path selection and transmission control method, and a mature and open transmission optimization solution for the real-time session service does not exist at present. Therefore, a video transmission optimization technology meeting the transmission requirement of the time-delay sensitive conversational service is urgently needed to be designed. The delay sensitive conversational services put the following requirements on the transmission: firstly, the transmission delay of the delay sensitive service is reduced as much as possible, and strict delay constraint is met. The end-to-end transmission delay needs to be controlled below 250ms, and can not exceed 400ms at most. The QoE is considered only in the transmission link, and the smaller the transmission delay without packet loss is, the better the QoE is; and secondly, the requirement of high reliability is met, different from the traditional expression, the current real-time conversation service is a service sensitive to both time delay and packet loss, the jitter and packet loss caused by network random congestion can be effectively avoided, and if the random congestion of a bearer network is inevitable, a corresponding mechanism is needed to overcome the influence of the network abnormality on the service QoE.
Aiming at the requirement of delay sensitive service on transmission, multipath transmission is an effective solution, and can transmit data by using the transmission capability of a plurality of paths simultaneously, so that the resources of a plurality of network cards and a plurality of links are fully utilized, and better experience quality is provided for terminal users. The use of multipath to achieve parallel transmission of data can improve end-to-end communication performance in several respects, for example: aggregate bandwidth, improve reliability of transmission, etc.
One important research point of a multipath transmission control mechanism is the problem of congestion control, network congestion refers to the fact that transmitted data exceeds the transmission capability which can be carried by a network, the reasons for network congestion are the following points, and the problems of network congestion are caused by WiFi/4G signal attenuation, core transmission link attenuation, network export competition, retransmission storms, equipment performance attenuation and the like. The congestion control is that the sending end controls the rhythm of sending packets to the network according to the network condition. Without congestion control mechanisms, it is doubtful whether the internet can evolve to serve half of the world's population today.
In summary, for the transmission optimization problem of the delay sensitive service, congestion control is a crucial part, and a reasonable congestion control mechanism can adaptively and dynamically adjust the rhythm of sending data packets, and reasonably utilize bandwidth, thereby achieving the purpose of reducing delay and improving network throughput.
The conventional congestion control algorithm is based on a packet loss feedback protocol. The protocol based on the packet loss feedback is a passive congestion control mechanism, which determines the network congestion according to the packet loss event in the network. Even if the load in the network is high, as long as congestion packet loss is not generated, the protocol does not actively reduce the sending speed of the protocol. Multipath TCP maintains a congestion window cwnd at the sender, which controls the amount of transmission. And adjusting cwnd by adopting AIMD (advanced information technology), namely, an increasing and multiplicative decreasing mode, and increasing the window in an increasing mode in a congestion avoiding stage and decreasing the window in a multiplicative mode when packet loss occurs. The assumption of this congestion control algorithm is that packet losses are all due to congestion. Network packet loss may be caused by various factors, such as: packet loss caused by router strategies, error packets caused by WIFI signal interference and the like. These packet losses are not caused by network congestion, but can cause a large window drop of multipath TCP, even if the network bandwidth is good, the bandwidth is still not well utilized, and due to the increasing update of network services, the current delay-sensitive conversational services do not simply pursue high throughput, but rather want to achieve strict delay constraints under the premise of high reliability.
Disclosure of Invention
The technical problem to be solved by the present invention is to provide a delay sensitive service oriented multipath coupling congestion control method to ensure the requirements of low delay and high reliability of service transmission, aiming at the deficiencies of the prior art.
In order to solve the technical problems, the technical scheme adopted by the invention is as follows:
a multipath coupling congestion control method facing to delay sensitive service includes the following steps:
startup is an acceleration stage of a multipath coupling BBR congestion control algorithm, and in the initial state, each substream increases the sending rate pacingRate and cwnd by a gain factor 2/ln2 larger than 1, and the bottleneck bandwidth is expected to be rapidly detected; the pacingRate and cwnd are calculated at this stage as follows:
step 1-1, detecting minimum delay Min RTT and maximum bandwidth Max BW of each sub-flow;
step 1-2, calculating the time delay bandwidth product BDP according to Min RTT and Max BW of each sub-flow;
step 1-3, if the detected Max BW is not increased to more than 25% of the original Max BW for more than 3 times, indicating that the bottleneck bandwidth has arrived, turning to step 2, if an RTT value smaller than or equal to the minimum RTT of the last period is not measured in a time window, jumping to step 4, otherwise, continuing to execute the step 1-4;
1-4, calculating pacingRate and cwnd for each sub-flow according to Min RTT, Max BW, BDP and a gain coefficient 2/ln 2;
the multipath coupling BBR congestion control reduces the sending rate in the stage, each sub-flow reduces the sending rate pacinRate and cwnd to ln 2/2 with the gain factor less than 1, which is to drain the network buffer caused by the StartUP stage; the pacingRate and cwnd are calculated at this stage of Drain as follows:
step 2-1, detecting the minimum delay Min RTT and the maximum bandwidth Max BW of each sub-flow;
step 2-2, calculating the time delay bandwidth product BDP according to Min RTT and Max BW of each sub-flow;
step 2-3, if the flight data packet inflight in the link is BDP, going to step 3, if no RTT value smaller than or equal to the minimum RTT in the last period is measured in a time window, jumping to step 4, otherwise, continuing to execute step 2-4;
step 2-4, calculating pacingRate and cwnd according to Min RTT, Max BW, BDP and gain coefficient ln 2/2 for each sub-flow;
this state is a stable state of the multipath coupled BBR congestion control algorithm, where the majority of the multipath coupled BBR congestion control runs, with a gain factor of 8 sequential cycles: 1.25, 0.75, 1, calculating pacingRate, calculating cwnd by using a coupled gain coefficient cwnd _ gain, and regulating and controlling the competition of each substream for bandwidth by adjusting the coefficient of cwnd; a complete cycle comprises 8 phases, each phase having a duration of Rtprop; the pacingRate and cwnd are calculated at this stage of ProbeBW as follows:
step 3-1, detecting the minimum delay Min RTT and the maximum bandwidth Max BW of each sub-flow;
step 3-2, calculating the time delay bandwidth product BDP according to Min RTT and Max BW of each sub-flow;
3-3, if no RTT value smaller than or equal to the minimum RTT of the last period is measured in a time window, jumping to the step 4, otherwise, continuing to execute the step 3-4;
3-4, circularly calculating pacingRate according to Min RTT, Max BW, BDP and 8 gain coefficients by each sub-flow, and calculating cwnd by using a coupling gain coefficient cwnd _ gain;
step 4, entering a ProbeRTT state;
the previously reserved Min RTT represents the minimum RTT when no queue is cached, so the increase of the RTT measurement value represents the increase of the cache queue, and at the moment, the multipath coupling BBR considers that the network is congested and needs to enter a Probe RTT state to empty the network and measure the RTT again; in the ProbeRTT state, cwnd is set to 4 MSS, and RTT is measured again and lasts for 200 ms; the procedure at this stage of ProbeRTT is as follows:
step 4-1, detecting the minimum delay Min RTT and the maximum bandwidth Max BW of each sub-flow;
after the steps 4-2 and 200ms, if the detected Max BW is not increased to more than 25% of the original Max BW for more than 3 times, turning to the step 3, otherwise, turning to the step 1.
Adopt the produced beneficial effect of above-mentioned technical scheme to lie in: the time delay sensitive service-oriented multipath coupling congestion control method provided by the invention can meet strict time delay constraint and high reliability while achieving high bandwidth utilization rate, and simultaneously ensures fairness to single-path TCP. In the procedure of multipath coupling BBR state transition, a BBR algorithm mechanism is used, and in a Probe BW stage, in order to meet fairness and balance congestion, data packets cached in a sending buffer area are allocated in proportion according to Max BW multiplied by Min RTT of each sub-flow measured currently to regulate and control competition of each sub-flow on bandwidth. When multiple sub-streams compete with each other on the bottleneck link, the bandwidth obtained by each sub-stream is proportional to the size of the buffer it occupies in the link queue. Therefore, the competition of each sub-flow for the bandwidth is regulated by adjusting the coefficient of cwnd. Compared with the traditional heuristic multipath coupling congestion control algorithm, the method of the invention reduces the round-trip delay and the packet loss rate, and is a session service transmission control mechanism which is more suitable for delay sensitive services.
Drawings
FIG. 1 is a diagram of a multi-path coupling BBR operation model provided by an embodiment of the present invention;
FIG. 2 is a state machine working diagram of the multipath coupled BBR algorithm provided by the embodiment of the present invention;
fig. 3 is a diagram of relations between pacingRate and cwnd according to an embodiment of the present invention;
FIG. 4 is a diagram of the cwnd _ gain calculation process provided by an embodiment of the present invention;
FIG. 5 is a multi-path network topology established by simulation provided by an embodiment of the present invention;
FIG. 6 is a comparison graph of simulated transmission rate effects provided by embodiments of the present invention;
fig. 7 is a comparison diagram of the effects of the simulated round trip delay RTT provided in the embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the present invention is provided in connection with the accompanying drawings and examples. The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.
The embodiment provides a multipath coupling congestion control method for delay sensitive services, which is mainly used for solving the problem of multipath congestion control of delay sensitive services and aims to optimize a multipath transmission control mechanism.
A multipath network environment is first established. The multipath coupling BBR algorithm is mainly applied to service scenarios, such as: the method comprises the steps of high-definition video conference, real-time high-definition video, video diagnosis and treatment and the like, and the conversation type service is sensitive to time delay. The working model scenario of the multipath coupling BBR congestion control mechanism is shown in fig. 1, and the purpose of the working model scenario is to control the sending rhythm of each multipath sub-stream sending end so as to avoid network congestion.
The specific steps of the multipath coupling congestion control method facing the delay sensitive service are as follows.
Startup is the acceleration phase of the congestion control algorithm of the multipath coupling BBR, and in this initial state, each substream is increased by a gain factor of 2/ln2 (greater than 1) to increase the sending rates pacinRate and cwnd, and it is desirable to quickly detect the bottleneck bandwidth. The pacingRate and cwnd are calculated at this stage as follows:
step 1-1, detecting the minimum delay Min RTT and the maximum bandwidth Max BW of each sub-flow, wherein the calculation processes of the minimum delay Min RTT and the maximum bandwidth Max BW are as follows: at any time t, counting time delay:
RTTt＝RTpropt+εt (1)
wherein epsilontRepresenting RTT noise caused by link delay, receiving end confirmation strategy and the like at the time t; RTproptIndicating the path round trip delay at time t which is only related to the path condition. The minimum value of the delay RTT is calculated as follows:
wherein, WRIs the minimum time window of operation.
The multipath coupling BBR congestion control algorithm only needs to calculate the bandwidth according to two values, namely the amount of data Δ delivered to reply and the time Δ t taken to reply to the data. These two values can be obtained from the packet delivery rate, which is an estimate of the bandwidth for each ACK received by the sender, so that the sending rate can be calculated as the ratio of the amount of data transmitted to the time it takes to transmit:
deliveryRatet＝Δdelivered/Δt (3)
the maximum transmission rate, i.e. the maximum bandwidth value, thus detected can be calculated by:
wherein, WBIs a runtime window that collects bandwidth values. RTT is calculated using the time each packet is transmitted over the link, and one RTT is generated for each ACK.
Step 1-2, calculating the time delay bandwidth product BDP according to Min RTT and Max BW for each sub-flow:
BDP＝max BW*min RTT (5)
step 1-3, if the detected Max BW is not increased to more than 25% of the original Max BW for more than 3 times, indicating that the bottleneck bandwidth has arrived, as shown in fig. 2, going to step 2, if an RTT value smaller than or equal to the minimum RTT of the last period is not measured within a time window (min _ RTT _ win _ sec), jumping to step 4, otherwise, continuing to execute step 1-4;
step 1-4, calculating pacingRate and cwnd according to Min RTT, Max BW, BDP and gain coefficient 2/ln2 for each sub-flow:
and step 2, entering a Drain state.
The Max BW detected in step 1-3 does not increase to more than 25% of the original Max BW for more than 3 times, which indicates that the bottleneck bandwidth has been reached at a previous time point, and the bandwidth increases for three times, which undoubtedly increases the number of packets in the network, so that the network cache is filled. Therefore, the multipath coupling enters a Drain state after the StartUP state, extra added flow in the network is drained, and the condition that only messages with the size of BDP exist in the network is ensured. The multipath coupling BBR reduces the transmission rate in this stage, and each substream reduces the transmission rate pacingRate and cwnd by a gain factor of ln 2/2 (less than 1) in order to drain the network buffer caused by the StartUP stage. The pacingRate and cwnd are calculated at this stage of Drain as follows:
step 2-1, detecting the minimum delay Min RTT and the maximum bandwidth Max BW of each sub-flow, and calculating the minimum delay Min RTT and the maximum bandwidth Max BW in the step 1-1;
step 2-2, calculating the time delay bandwidth product BDP according to Min RTT and Max BW of each sub-flow, wherein the BDP is calculated as the step 1-2;
step 2-3, if the flight data packet inflight in the link is BDP, as shown in fig. 2, go to step 3, if no RTT value smaller than or equal to the minimum RTT in the last period is measured within a time window (min _ RTT _ win _ sec), go to step 4, otherwise, continue to execute step 2-4;
step 2-4, calculating pacingRate and cwnd according to Min RTT, Max BW, BDP and gain coefficient ln 2/2 by each sub-flow, and showing the following two formulas:
and 3, entering a ProbeBW state.
This state is a stable state of the multipath coupling BBR algorithm, where the multipath coupling BBR is running most of the time, this stage is with gain coefficients of 8 sequential cycles: 1.25, 0.75, 1, calculating pacingRate, calculating cwnd by a coupling gain coefficient cwnd _ gain, and regulating the competition of each sub-stream for bandwidth by adjusting the coefficient of cwnd. A complete cycle consists of 8 phases, each phase having a duration of one RTprop. With a gain factor of 1.25, more bandwidth can be detected, a further increase in the transmission rate will cause the accumulation of network packets, and in the next RTprop, a rate gain of 0.75 is used to drain the accumulated packets. The gain factor is 1 in order to maintain stability. The pacingRate and cwnd are calculated at this stage of ProbeBW as follows:
step 3-1, detecting the minimum delay Min RTT and the maximum bandwidth Max BW of each sub-flow, and calculating the minimum delay Min RTT and the maximum bandwidth Max BW in the step 1-1;
step 3-2, calculating the time delay bandwidth product BDP according to Min RTT and Max BW of each sub-flow, wherein the BDP is calculated as the step 1-2;
3-3, if no RTT value smaller than or equal to the minimum RTT of the last period is measured in a time window (min _ RTT _ win _ sec), as shown in figure 2, jumping to step 4, otherwise, continuing to execute step 3-4;
and 3-4, circularly calculating the pacingRate of each sub-flow according to Min RTT, Max BW, BDP and 8 gain coefficients:
pacingRate＝max BW*G G∈{1.25，0.75，1，1，1，1，1，1} (10)
the embodiment regulates the competition of each sub-stream for the bandwidth by adjusting the gain coefficient of cwnd. As shown in fig. 3, the multipath coupling BBR separates the rate control calculation from the actual transmission, which is the case when the control plane is typically separated from the data plane. That is, the multipath BBR core module calculates a rate and then throws the packet to the pacifing send engine module, and the pacifing send engine module interacts with the pacifing send engine module through a send buffer when sending is controlled by the pacifing send engine module. cwnd functions to control the amount of packets that are buffered in the transmit buffer for transmission. When multiple sub-streams compete with each other on the bottleneck link, the bandwidth obtained by each sub-stream is proportional to the size of the buffer it occupies in the link queue. Therefore, the competition of each sub-flow for the bandwidth is regulated by adjusting the coefficient of cwnd. The cwnd is calculated by using a coupling gain factor cwnd _ gain, the process of calculating the coupling gain factor cwnd _ gain is shown in fig. 4, and the rule of adjusting the gain factor of the cwnd of each substream by the multipath coupling BBR is to calculate the BDP of each substream respectively, that is:
max1 BW*min1 RTT，max2 BW*min2 RTT，……(11)
the sum of the maximum BDP of the individual substreams is then calculated:
sum＝max1 BW*min1 RTT+max2 BW*min2 RTT+…(12)
finally, the cwnd gain coefficient cwnd _ gain of the cwnd of the nth substream is:
and 4, entering a ProbeRTT state.
The previously reserved Min RTT represents the minimum RTT without queue buffering, and then the increase of the RTT measurement value represents the increase of the buffer queue, and at this time, the multipath coupling BBR considers that the network is congested and needs to enter a ProbeRTT state to empty the network and measure the RTT again. In the ProbeRTT state, cwnd is set to 4 MSSs and RTT is re-measured for 200 ms. The procedure at this stage of ProbeRTT is as follows:
step 4-1, detecting the minimum delay Min RTT and the maximum bandwidth Max BW of each sub-flow, and calculating the minimum delay Min RTT and the maximum bandwidth Max BW in the step 1-1;
after steps 4-2 and 200ms, if the detected Max BW is not increased to more than 25% of the original Max BW for more than 3 times, turning to step 3 as shown in figure 1, otherwise, turning to step 1.
As fig. 5 establishes a multipath network topology, the multipath coupling BBR congestion control algorithm of the present invention is arranged at the sending end, in order to prove the algorithm effectiveness of the present invention. The parameters of the two sub-stream links are set to be 1.5Mbps, the one-way transmission delay is 50ms, and the packet loss rate is 1%, so that experiments are carried out. Starting connection establishment, collecting data obtained by network parameter simulation: sending rate, one-way transmission delay. Meanwhile, the traditional algorithm is simulated, and performance comparison is carried out. The upper stationary line in fig. 6 represents the multipath coupling BBR of the present invention, and the lower line with large square wave motion represents the classical LIA algorithm, which shows that the method of the present invention has higher bandwidth utilization and more stable performance than the LIA algorithm. In the comparison of the one-way transmission delay of fig. 7, the upper line represents the LIA algorithm, and the lower line represents the multi-path coupling BBR delay line, so that it can be seen that the multi-path coupling BBR can realize a lower one-way transmission delay.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; such modifications and substitutions do not depart from the spirit of the corresponding technical solutions and scope of the present invention as defined in the appended claims.
Claims (1)
1. A multipath coupling congestion control method facing to delay sensitive service is characterized in that: the method comprises the following steps:
step 1, entering a StartUP state when end-to-end connection is established;
startup is an acceleration stage of a multipath coupling BBR congestion control algorithm, and each substream in the initial state increases sending rates pacinRate and cwnd by a gain coefficient 2/ln2 larger than 1, and hopes to quickly detect bottleneck bandwidth; the pacingRate and cwnd are calculated at this stage as follows:
step 1-1, detecting minimum delay Min RTT and maximum bandwidth Max BW of each sub-flow;
step 1-2, calculating the time delay bandwidth product BDP according to Min RTT and Max BW of each sub-flow;
step 1-3, if the detected Max BW is not increased to more than 25% of the original Max BW for more than 3 times, indicating that the bottleneck bandwidth has arrived, turning to step 2, if an RTT value smaller than or equal to the minimum RTT of the last period is not measured in a time window, jumping to step 4, otherwise, continuing to execute the step 1-4;
1-4, calculating pacingRate and cwnd for each sub-flow according to Min RTT, Max BW, BDP and a gain coefficient 2/ln 2;
step 2, entering a Drain state;
the multipath coupling BBR congestion control reduces the sending rate in the stage, each sub-flow reduces the sending rate pacinRate and cwnd to ln 2/2 with the gain factor less than 1, which is to drain the network buffer caused by the StartUP stage; the pacingRate and cwnd are calculated at this stage of Drain as follows:
step 2-1, detecting the minimum delay Min RTT and the maximum bandwidth Max BW of each sub-flow;
step 2-2, calculating the time delay bandwidth product BDP according to Min RTT and Max BW of each sub-flow;
step 2-3, if the flight data packet inflight in the link is BDP, going to step 3, if no RTT value smaller than or equal to the minimum RTT in the last period is measured in a time window, jumping to step 4, otherwise, continuing to execute step 2-4;
step 2-4, calculating pacingRate and cwnd according to Min RTT, Max BW, BDP and gain coefficient ln 2/2 for each sub-flow;
step 3, entering a ProbeBW state;
this state is a stable state of the multipath coupled BBR congestion control algorithm, where the majority of the multipath coupled BBR congestion control runs, with a gain factor of 8 sequential cycles: 1.25, 0.75, 1, calculating pacingRate, calculating cwnd by using a coupled gain coefficient cwnd _ gain, and regulating and controlling the competition of each substream for bandwidth by adjusting the coefficient of cwnd; a complete cycle comprises 8 phases, each phase having a duration of Rtprop; the pacingRate and cwnd are calculated at this stage of ProbeBW as follows:
step 3-1, detecting the minimum delay Min RTT and the maximum bandwidth Max BW of each sub-flow;
step 3-2, calculating the time delay bandwidth product BDP according to Min RTT and Max BW of each sub-flow;
3-3, if no RTT value smaller than or equal to the minimum RTT of the last period is measured in a time window, jumping to the step 4, otherwise, continuing to execute the step 3-4;
3-4, circularly calculating pacingRate according to Min RTT, Max BW, BDP and 8 gain coefficients by each sub-flow, and calculating cwnd by using a coupling gain coefficient cwnd _ gain;
step 4, entering a ProbeRTT state;
the previously reserved Min RTT represents the minimum RTT when no queue is cached, so the increase of the RTT measurement value represents the increase of the cache queue, and at the moment, the multipath coupling BBR considers that the network is congested and needs to enter a Probe RTT state to empty the network and measure the RTT again; in the ProbeRTT state, cwnd is set to 4 MSS, and RTT is measured again and lasts for 200 ms; the procedure at this stage of ProbeRTT is as follows:
step 4-1, detecting the minimum delay Min RTT and the maximum bandwidth Max BW of each sub-flow;
after the steps 4-2 and 200ms, if the detected Max BW is not increased to more than 25% of the original Max BW for more than 3 times, turning to the step 3, otherwise, turning to the step 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110480669.5A CN113518040B (en) | 2021-04-30 | 2021-04-30 | Multipath coupling congestion control method for delay sensitive service |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110480669.5A CN113518040B (en) | 2021-04-30 | 2021-04-30 | Multipath coupling congestion control method for delay sensitive service |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113518040A true CN113518040A (en) | 2021-10-19 |
| CN113518040B CN113518040B (en) | 2022-12-09 |
Family
ID=78063870
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110480669.5A Active CN113518040B (en) | 2021-04-30 | 2021-04-30 | Multipath coupling congestion control method for delay sensitive service |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113518040B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114374628A (en) * | 2022-02-09 | 2022-04-19 | 江苏济楚信息技术有限公司 | Model-based audio and video communication low-delay congestion control method |
| CN114629847A (en) * | 2022-03-08 | 2022-06-14 | 西安电子科技大学 | Coupling multi-flow TCP congestion control method based on available bandwidth allocation |
| CN114785757A (en) * | 2022-03-31 | 2022-07-22 | 东北大学 | Multipath transmission control method for real-time session service |
| CN114844837A (en) * | 2022-04-25 | 2022-08-02 | 西安交通大学 | Congestion control mechanism and control device based on time delay in multi-service queue environment |
| CN114915600A (en) * | 2022-06-10 | 2022-08-16 | 贵州大学 | BBRv2 congestion control algorithm under deep buffer |
| CN115412497A (en) * | 2022-11-01 | 2022-11-29 | 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) | Performance optimization method of BBR congestion control algorithm |
| EP4221144A1 (en) * | 2022-01-26 | 2023-08-02 | Deutsche Telekom AG | Determining optimized data rates for a communication path |
| CN117714381A (en) * | 2023-12-14 | 2024-03-15 | 中国科学院合肥物质科学研究院 | Fair congestion control method and device with flow perception under SDN data center network |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016072836A1 (en) * | 2014-11-06 | 2016-05-12 | Universiti Putra Malaysia | A method for tcp congestion in multi-hop wireless network |
| US20180159778A1 (en) * | 2015-06-08 | 2018-06-07 | Samsung Electronics Co., Ltd. | Method and apparatus for tcp-based transmission control in communication system |
| CN109309934A (en) * | 2017-07-27 | 2019-02-05 | 华为技术有限公司 | A kind of jamming control method and relevant device |
| CN109639340A (en) * | 2018-12-11 | 2019-04-16 | 成都天奥信息科技有限公司 | A kind of TCP accelerated method suitable for satellite link |
| WO2019080938A1 (en) * | 2017-10-28 | 2019-05-02 | 华为技术有限公司 | Date transmission method, apparatus and device |
| CN111541626A (en) * | 2020-04-21 | 2020-08-14 | 广州市百果园信息技术有限公司 | Network bandwidth updating method and device, electronic equipment and storage medium |
| CN111683393A (en) * | 2020-05-25 | 2020-09-18 | 华中科技大学 | Adaptive congestion control method for dynamically adjusting gain coefficient |
| WO2021012869A1 (en) * | 2019-07-25 | 2021-01-28 | 广州市百果园信息技术有限公司 | Transmission rate determination method and device, apparatus, and storage medium |
-
2021
- 2021-04-30 CN CN202110480669.5A patent/CN113518040B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2016072836A1 (en) * | 2014-11-06 | 2016-05-12 | Universiti Putra Malaysia | A method for tcp congestion in multi-hop wireless network |
| US20180159778A1 (en) * | 2015-06-08 | 2018-06-07 | Samsung Electronics Co., Ltd. | Method and apparatus for tcp-based transmission control in communication system |
| CN109309934A (en) * | 2017-07-27 | 2019-02-05 | 华为技术有限公司 | A kind of jamming control method and relevant device |
| WO2019080938A1 (en) * | 2017-10-28 | 2019-05-02 | 华为技术有限公司 | Date transmission method, apparatus and device |
| CN109639340A (en) * | 2018-12-11 | 2019-04-16 | 成都天奥信息科技有限公司 | A kind of TCP accelerated method suitable for satellite link |
| WO2021012869A1 (en) * | 2019-07-25 | 2021-01-28 | 广州市百果园信息技术有限公司 | Transmission rate determination method and device, apparatus, and storage medium |
| CN111541626A (en) * | 2020-04-21 | 2020-08-14 | 广州市百果园信息技术有限公司 | Network bandwidth updating method and device, electronic equipment and storage medium |
| CN111683393A (en) * | 2020-05-25 | 2020-09-18 | 华中科技大学 | Adaptive congestion control method for dynamically adjusting gain coefficient |
Non-Patent Citations (1)
| Title |
|---|
| 魏文佳: "基于MPTCP的多路径传输中的耦合拥塞控制和数据调度机制研究", 《中国优秀博硕士学位论文全文数据库（博士）信息科技辑》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP4221144A1 (en) * | 2022-01-26 | 2023-08-02 | Deutsche Telekom AG | Determining optimized data rates for a communication path |
| CN114374628A (en) * | 2022-02-09 | 2022-04-19 | 江苏济楚信息技术有限公司 | Model-based audio and video communication low-delay congestion control method |
| CN114629847A (en) * | 2022-03-08 | 2022-06-14 | 西安电子科技大学 | Coupling multi-flow TCP congestion control method based on available bandwidth allocation |
| CN114629847B (en) * | 2022-03-08 | 2023-09-08 | 西安电子科技大学 | Coupled multi-stream TCP congestion control method based on available bandwidth allocation |
| CN114785757A (en) * | 2022-03-31 | 2022-07-22 | 东北大学 | Multipath transmission control method for real-time session service |
| CN114785757B (en) * | 2022-03-31 | 2023-10-20 | 东北大学 | Multipath transmission control method for real-time conversation service |
| CN114844837A (en) * | 2022-04-25 | 2022-08-02 | 西安交通大学 | Congestion control mechanism and control device based on time delay in multi-service queue environment |
| CN114844837B (en) * | 2022-04-25 | 2023-09-26 | 西安交通大学 | Congestion control method and device based on time delay in multi-service queue environment |
| CN114915600A (en) * | 2022-06-10 | 2022-08-16 | 贵州大学 | BBRv2 congestion control algorithm under deep buffer |
| CN115412497A (en) * | 2022-11-01 | 2022-11-29 | 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) | Performance optimization method of BBR congestion control algorithm |
| CN115412497B (en) * | 2022-11-01 | 2022-12-30 | 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) | Performance optimization method of BBR congestion control algorithm |
| CN117714381A (en) * | 2023-12-14 | 2024-03-15 | 中国科学院合肥物质科学研究院 | Fair congestion control method and device with flow perception under SDN data center network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113518040B (en) | 2022-12-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113518040B (en) | Multipath coupling congestion control method for delay sensitive service | |
| US9088510B2 (en) | Universal rate control mechanism with parameter adaptation for real-time communication applications | |
| US6839767B1 (en) | Admission control for aggregate data flows based on a threshold adjusted according to the frequency of traffic congestion notification | |
| US7284047B2 (en) | System and method for controlling network demand via congestion pricing | |
| CN110572333B (en) | Method for adjusting sending rate in bandwidth detection stage and method for realizing congestion control algorithm | |
| WO2003015355A2 (en) | Method for supporting non-linear, highly scalable increase-decrease congestion control scheme | |
| CN110808884B (en) | Network congestion control method | |
| CN113207147A (en) | MPTCP congestion control method based on deep reinforcement learning and storage medium thereof | |
| Mamatas et al. | Approaches to congestion control in packet networks | |
| Khademi et al. | The new AQM kids on the block: Much ado about nothing? | |
| Nguyen et al. | Multimedia streaming using multiple TCP connections | |
| Matsufuji et al. | Multipath TCP path schedulers for streaming video | |
| Zhang et al. | An evaluation of bottleneck bandwidth and round trip time and its variants | |
| Zhang et al. | An online learning based path selection for multipath real‐time video transmission in overlay network | |
| Yamamoto et al. | Congestion control for reliable multicast achieving TCP fairness | |
| CN114785757B (en) | Multipath transmission control method for real-time conversation service | |
| Nagayama et al. | Path switching schedulers for mptcp streaming video | |
| CN116760777B (en) | Multipath congestion control method based on ABEA3C | |
| Alekseev et al. | ARTCP: Efficient algorithm for transport protocol for packet switched networks | |
| CN115134292B (en) | Path management method for multipath transmission real-time streaming media based on receiving window | |
| Li et al. | Performance Improvement of TFRC Congestion Control Algorithm in Wireless Network Streaming Media Transmission | |
| Li et al. | Fast link adaptation for TFRC after a handover | |
| CN112423342A (en) | LEDBAT-based congestion control window acquisition method | |
| Liu et al. | Achieving high throughput and low delay by accurately regulating link queue length over mobile data network | |
| Nagayama et al. | Managing Path Switching in Multipath Video Streaming |
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 |