CN101056260A - ECN mechanism-based congestion control method in the mixed network - Google Patents

ECN mechanism-based congestion control method in the mixed network Download PDF

Info

Publication number
CN101056260A
CN101056260A CNA2007100349629A CN200710034962A CN101056260A CN 101056260 A CN101056260 A CN 101056260A CN A2007100349629 A CNA2007100349629 A CN A2007100349629A CN 200710034962 A CN200710034962 A CN 200710034962A CN 101056260 A CN101056260 A CN 101056260A
Authority
CN
China
Prior art keywords
congestion
ack
tcp
ecn
probability
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
Application number
CNA2007100349629A
Other languages
Chinese (zh)
Other versions
CN101056260B (en
Inventor
王建新
叶进
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Central South University
Original Assignee
Central South University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Central South University filed Critical Central South University
Priority to CN2007100349629A priority Critical patent/CN101056260B/en
Publication of CN101056260A publication Critical patent/CN101056260A/en
Application granted granted Critical
Publication of CN101056260B publication Critical patent/CN101056260B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The invention discloses a congestion control method based on ECN mechanism in the mixed network, including the following steps: extract final section of ECN feedback information of the router, computer the congestion probability according to the distribution of the congestion mask, the sender can predict the congestion via the change trend of the congestion probability and adjust the congestion control mechanism according to the prediction result. The invention deduces the congestion according to the change rule of the congestion probability, namely only when the congestion probability value shows the increase trend, the congestion can be identified, so it can avoid the instantaneous and blind determination of the network status when the independent ECN feedback is used. The prediction results of the feedback are used to control the response behavior of the TCP packet loss and can accurately eliminate the influences of the non-congestion packet loss on the TCP congestion control in the mixed network.

Description

In the hybrid network based on the jamming control method of ECN mechanism
Technical field
The present invention relates in a kind of hybrid network jamming control method based on ECN mechanism.
Background technology
Different with cable network is, wireless network has higher bit error code, complicated channel fading, the noise of equipment of burst, these characteristics will cause frequently producing packet loss, traditional TCP simply is summed up as network congestion to all packet losses and blindly takes control strategy, thereby has greatly reduced the performance of TCP.Therefore TCP need get rid of the influence of non-congestion packet loss to original congestion control mechanism in the wire/wireless hybrid network.
Mainly improve the TCP performance in the wire/wireless hybrid network by distinguishing congestion packet loss and wireless wrong packet loss.Packet loss differentiating method has the NCPLD based on round-trip delay [4], based on the TCP-Biaz of arrival interval of bag [5], based on the TCP-Spike of ROTT [6], the TCP-Bayes that changes based on RTT [7], based on header check and TCP-HACK [8]Or the like.Mainly there is following problem in existing these methods:
(1) most of differentiating methods adopt and estimate end to end, and they have stronger dynamic in constituting complicated day by day hybrid network, therefore set up these estimate and network state between corresponding relation be inconvenient.This will directly influence the accuracy and the corresponding TCP performance of packet loss differentiating.The TCP-Bayes method has fully been paid attention to these characteristics, set up the HMM model that shows dynamic relationship between RTT and the packet loss character, to distinguish accuracy rate has remained on more than 80%, need learn and train based on historical data but obtain this model, the cost of paying thus will make the application of this method in TCP control be subjected to great restriction.
(2) the present accuracy index that the evaluation of differentiating method is adopted the packet loss incident, and in hybrid network because long delay characteristics does not possess real-time to the response of single packet loss incident, that is to say that the differentiation information that transmit leg obtained is the state of previous network sometime in fact.Reach 100% even distinguish accuracy so, also may have no to help for improving the TCP performance.It is very crucial how effectively utilizing packet loss incident and differentiation information thereof among the TCP.
As Transmission Control Protocol important in the present hybrid network, TCP Westwood, Jersey have adopted the method for packet loss differentiating.
TCP returns ACK speed by monitoring and comes test constantly effective bandwidth, and estimates congestion window and slow enabling gate limit value with current effective bandwidth.This method is taked different control strategies based on the measurement result of system, is actually the implicit overall situation control of the rule of packet loss identification the inside.Its bottleneck is that time delay and cumulative effect that ACK arrives make the accuracy of bandwidth estimation be affected.
TCP Jersey [13]: TCP Jersey has strengthened the ability of packet loss differentiating by means of ECN mechanism on the basis of Westwood, transmit leg detects the ACK feedback when packet loss takes place, if wherein congestion marking position (Congestion Explicit) puts 1, just this packet loss is classified as congested causing, mistake causes otherwise be classified as.The Wireless_ECN method of Fei and Shiduan proposition that Here it is is hereinafter to be referred as WECN.Yet more existing in recent years researchs point out that the packet loss differentiating ability of ECN mechanism is more weak, the levels of precision of the congested perception that places one's entire reliance upon.For this reason, there is a large amount of work to improve packet loss differentiating algorithm from two aspects based on ECN mechanism.
The work of first aspect is intended to improve the precision of ECN mechanism to congested perception/mark.
SpecTCP comes the minimum congestion packet loss by the parameters (cache size, marking-threshold, drop probabilities etc.) of optimizing AQM, but this parameter optimization is to be difficult to guarantee minimizing of congestion packet loss in the network of the many bottlenecks of a plurality of streams owing to be static configuration.
Multilevel ECN is subdivided into two-stage with the marking probability interval, and has designed corresponding marking probability respectively, and this method has improved the accuracy of ECN mark, but has also increased the complexity of router work, and it is responsive more that the setting of parameter simultaneously becomes.
The work of second aspect is intended to reduce the time delay of ECN feedback, guarantees the real-time of TCP response.Accurately mark is congested if ECN mechanism can be in various flow generation models, so should be how the congestion marking of ECN be responded?
The network state that ACK Spoofing/Congestion Signal Cancellation method allows router to upgrade in time in the process of feedback congestion marking, ACK Spoofing is congested to the transmit leg announcement in advance by producing 3 repeat ACKs voluntarily, the then prior buffer memory congestion marking of Congestion Signal Cancellation, wait for next moment router affirmation network state determines whether send this mark again, these two kinds of technology all are in order to allow transmit leg in the nearest network state of the shortest time IT, to have strengthened overhead undoubtedly and implemented difficulty but finish such work by router.
As seen, obtained paying close attention to widely based on the jamming control method of ECN mechanism, existing prioritization scheme makes it demonstrate huge application potential.But on retouch the problem of second aspect, promptly how to utilize the congested feedback of ECN to control the TCP behavior, still need to seek succinct more effectively scheme, also be a key point that improves Transmission Control Protocol performance in the present hybrid network.
Summary of the invention
In order to solve the technical problem that above-mentioned jamming control method based on ECN mechanism exists, the invention provides in a kind of hybrid network jamming control method based on ECN mechanism.This method can combine packet loss incident and congestion probability and carry out the speed regulation and control, thereby has ensured the robustness of TCP under the situation that link error increases.
The technical scheme that the present invention solves the problems of the technologies described above may further comprise the steps: nearest one section ECN feedback information of intercepting router, and calculate congestion probability according to the distribution of congestion marking wherein, transmit leg carries out congestion prediction by the variation tendency of congestion probability, and according to the adjustment congestion control mechanism that predicts the outcome.
In the jamming control method based on ECN mechanism in the above-mentioned hybrid network, described congestion probability is calculated as follows:
CP = Σ i = 0 m - 1 w i * Ack [ i ] - > ecnecho
Ack is the ACK sequence of the nearest arrival of transmit leg collection in the following formula, Ack[i]->ecnecho is the value of CE position during wherein i ACK wraps, w iBe ACK[i] weights, Σ i = 0 m - 1 w i = 1 .
Based on the jamming control method of ECN mechanism, when the variation of congestion probability value presented the trend of increase, transmit leg was inferred as congestion state with current network in the above-mentioned hybrid network.
In the above-mentioned hybrid network based on the jamming control method of ECN mechanism, only when prediction result when being congested TCP send corresponding congestion control mechanisms such as starting reduction of speed just now, otherwise the packet of retransmits lost just.
Technique effect of the present invention is: transmit leg refreshes the ACK array among the present invention when receiving each ACK bag, and calculate new congestion probability value, control according to intrinsic ECN mechanism fully for normal ACK, and for repeat ACK, infer network state according to the congestion probability value, have only when the result who infers is congestion packet loss, just reduce the send window size.The inventive method is carried out congested deduction by the Changing Pattern of congestion probability value, promptly have only when the congestion probability value presents the trend of increase, just be inferred as congested, thereby when having avoided using independent ECN feedback, instantaneity and blindness that network state is judged.Owing to only need to revise the transmit leg code in the invention process, the inventive method can be applied to various Transmission Control Protocol, make it to be adapted to the transmission characteristic of isomery link in the wire/wireless hybrid network, to guarantee that network obtains higher throughput and link utilization, fairness between friendly between maintenance and the original Transmission Control Protocol, and stream simultaneously and the stream.
The present invention is further illustrated below in conjunction with accompanying drawing.
Description of drawings
Fig. 1 is a flow chart of the present invention.
Fig. 2 TCP throughput of the present invention analog simulation environment.
Fig. 3 is applied to performance evaluation design sketch among the Westwood, called after CPECN wherein of the present invention for the present invention.Fig. 3 (a) be different labeling methods throughput ratio; Fig. 3 (b) is that end-to-end packet loss compares; (c) send window situation of change.
Fig. 4 Transmission Control Protocol friendly of the present invention analog simulation environment.
Embodiment
Referring to Fig. 1, Fig. 1 is a flow chart of the present invention.Process is as follows: at first enter startup slowly after TCP starts, when congestion window just enters congestion avoidance phase after starting threshold value slowly.In congestion avoidance phase,, so just restart to start slowly if take place overtimely; If receive new ACK, just carry out Congestion Avoidance by normal TCP; If but received the ACK of repetition, just need carry out congestion prediction.The congestion prediction method is as follows: nearest one section ECN feedback information of intercepting router, and calculate congestion probability according to the distribution of congestion marking wherein, transmit leg carries out congestion prediction by the variation tendency of congestion probability, and according to the adjustment congestion control mechanism that predicts the outcome.If predict the outcome to congested, transmit leg will carry out the speed adjustment and repeating transmission enters fast recovery so; If prediction result is non-congested, fast recovery is only retransmitted and entered to transmit leg so.
The result who shows analysis among the present invention with " congestion probability ".Particularly, gather the feedback information of nearest one section ACK by the TCP transmit leg, adopt EWMA (Exponentially Weighted Moving Average) method congestion probability to be predicted this probability is called " the congested possibility factor ", computational methods are as follows:
CP = Σ i = 0 m - 1 w i * Ack [ i ] - > ecnecho , Formula (1)
Ack[i wherein]->ecnecho={0,1}, Σ i = 0 m - 1 w i = 1 , w i = α , i ∈ [ 0 , m / 2 - 1 ] β , i ∈ [ m / 2 , m - 1 ] ,
M is big more, and the ECN feedback of obtaining is many more, helps the accuracy of prediction probability more, but can influence the real-time of congested control and error control simultaneously.Follow continuous 3 repeat ACKs among the Reno and enter the way that starts slowly, and intercept the state that next ACK wraps to determine intermediate router, therefore wrapping with 4 ACK in the experiment is that unit calculates the P congestion probability, and carries out probability analysis, i.e. m=8 in conjunction with 2 units in front and back.According to the EWMA method, nearest 4 ACK bag is got bigger weights α, before 4 ACK are got less weights β, make the full probability sum of preceding 4 ACK equal the full probability of nearest 1 ACK, so: Σβ = α 4 * α + 4 * β = 1 That is: α=0.2, β=0.05.
In sum, adopt the inventive method not to be subjected to the influence of the recoverable non-congestion packet loss incident of short-term, thereby can strengthen the ability that TCP resists wrong link so that TCP connects.
Utilize NS2.28 network simulation platform, we have realized the inventive method in TCP Westwood, and its performance is tested.The NS network simulator is a kind of general multiprotocol network simulation softward, and it is an open issue (network address: http://www.isi.edu/nsnam/ns), be extensive use of by network research person at present on the Internet.NS2.28 is one of its version.
We adopt the performance of topological structure test the inventive method shown in Figure 4.As shown in Figure 4, single TCP stream sends to R3 from R0, and simulation is through the Internet of the long delay of a plurality of routers between the R1 to R2, and final jump is the Radio Link that contains the mistake packet loss.Particularly, user's incoming end link bandwidth is 100M, and wireless link bandwidth is 10M, and time delay is 3ms, bottleneck bandwidth 5M, and time delay 40ms, the cache size on the bottleneck is set at pipe capacity.We have added the WECN method as a comparison simultaneously in Westwood.For the TCP performance under the test link packet drop, we have introduced a simple wrong packet loss model in the simulated experiment, and with the parameter of link packet drop rate as this model.The acquiescence simulation process continues 100 seconds.The grouping size of all protocol streams all is made as 1Kbyte.
1, the quantitative analysis of handling up
Fig. 3 (a) has provided the comparison of throughput when WECN method and the inventive method are applied to TCPWestwood under the different packet loss rate, suppose to take on the router simple threshold values labeling method, as seen from the figure, no matter how the threshold value on the router is provided with, and uses the WECN method and can both obtain than using the higher throughput of the inventive method.
In order further to analyze the reason that throughput improves, Fig. 3 (b), Fig. 3 (c) have provided marking-threshold and have got 3/4 o'clock link packet drop rate of queue size and the situation of change of tcp window (packet loss all refers to the link packet drop rate of appointment in the error model here, and end-to-end Loss Rate refers to the reception packet of TCP transmit leg statistics and the ratio of transmission packet), Fig. 3 (b) shows the amplitude maximum that the end-to-end packet loss of Westwood increases along with the increase of link packet drop rate, especially after packet loss surpasses 0.005.Use WECN method and the inventive method and do not strengthen end-to-end packet loss, illustrate that the inventive method improves throughput to improve link utilization.
Fig. 3 (c) has compared the window situation of change (getting Loss Rate is 0.01) during 0~30ms, and Westwood has repeatedly to enter slowly and starts, and especially serious shake has appearred in the send window of Westwood near 30ms; After using the WECN method, transmit leg has also entered once at packet loss and after receiving the congestion marking of single ECN feedback and has started slowly; After using the inventive method, since based on congestion probability value variation tendency predict the outcome for " False ", therefore TCP does not enter when packet loss slowly and starts, the judgement different with the WECN method appearred in the inventive method near 22ms equally, as a whole, the send window of the inventive method is more steady.
Load the ON/OFF stream of obeys index distribution in network shown in Figure 2, its speed is 5Mbps, has 25ms to be in " on " state in the time at the RTT of about 100ms, and all the other time slots are in " off " state.Table 1 has been described the result of UDP/TCP mixed flow test.
The TCP/UDP throughput ratio under the table 1 loading on/off stream situation
Packet loss 0.001 0.005 0.01
Agreement TCP UDP TCP UDP TCP UDP
Westwood 1694 936 1457 971 1346 970
Westwood?with?WECN 1437 1114 1393 1041 1359 1039
Westwood with the present invention 1679 994 1618 1039 1482 1055
Table 1 as seen, Westwood possesses preferably the ability with UDP stream shared bandwidth in hybrid network, but along with packet loss is increased to 0.01 from 0.001, the throughput of TCP stream has reduced by 20%; When TCP and UDP stream competitive channel, do not embody bigger superiority after using the WECN method; After using the inventive method, the overall throughput of TCP is optimum, and with the increasing of packet loss, throughput has only descended 11%.
2, fairness and friendly analysis
Fairness and friendly are the important performances of Transmission Control Protocol, a plurality of connections of identical TCP must harmonious coexistences and are rapidly converged to fair point, should coexist with other traditional Transmission Control Protocol with stylish Transmission Control Protocol, and can not cause using the TCP stream of legacy protocol by hungry to death.The fairness of TCP and friendly test when we adopt topology shown in Figure 6 to application the inventive method.
Adopt 10 identical TCP streams to insert bottleneck link in the fairness test, the total bandwidth of 10 streams is 1000Mbps, shares the long delay bottleneck link of 20Mbps, describes the mark of seeing among Fig. 4 in detail.The fairness of TCP Westwood under different packet loss rate situation was all better after the test result of table 2 showed application the present invention.
10 identical TCP stream fairness of table 2 relatively
Loss?Rate Westwood Westwood with WECN Westwood with the present invention
0 0.935 0.962 0.997
0.001 0.998 0.999 0.999
0.005 0.999 1.0 1.0
0.01 0.999 0.998 0.999
In the friendly experiment 10 TCP flow points are not distributed to two quasi-protocols, a class adopts the Westwood agreement of using the inventive method, the Westwood agreement that another kind of employing is original.When table 3 has provided two quasi-protocols and has taken different the connection, the situation of change of average throughput, therefrom as seen, the inventive method in all cases can both friendly coexistence with Westwood.
Table 3 the inventive method and the experiment of Westwood friendly
West wood Source Westwood with Source of the present invention no?loss Loss?Rate:0.001
Westwood Westwood with the present invention Westwood Westwood with the present invention
3 7 1162.00 1174.85 1161.33 1174.14
5 5 1316.80 1331.40 1312.80 1329.40
7 3 1519.28 1536.66 1511 1539.00
West wood Source Westwood with Source of the present invention Loss?Rate:0.005 Loss?Rate:0.01
Westwood Westwood with the present invention Westwood Westwood with the present invention
3 7 1289.66 1314.85 1099.00 1186.57
5 5 1294.80 1319.80 1269.80 1352.2
7 3 1485.28 1508.33 1388.14 1506.00
It seems from above experimental data, the performance of these two aspects of Westwood agreement of application the inventive method is more satisfactory, this is because this Transmission Control Protocol at Radio Link of Westwood, itself has overcome some defectives of traditional Transmission Control Protocol such as Reno, at first, adjusting SSTHRESH that Westwood adopts and the method for CWND are than directly these 2 values being put 1 or the mode that reduces by half among the Reno, can more effective control TCP speed, reduce the influence that link error brought; Secondly, application WECN method is separated error control and congested control by the congestion marking of intermediate router at transmitting terminal, has played certain effect.And the inventive method is on their basis, and the congestion marking of feedback has been carried out more reasonable use, therefore the fairness of self and aspect the friendly of Westwood performance better.

Claims (5)

  1. In the hybrid network based on the jamming control method of ECN mechanism, may further comprise the steps: nearest one section ECN feedback information of intercepting router, and calculate congestion probability according to the distribution of congestion marking wherein, transmit leg carries out congestion prediction by the variation tendency of congestion probability, and according to the adjustment congestion control mechanism that predicts the outcome.
  2. 2. based on the jamming control method of ECN mechanism, described congestion probability is calculated as follows in the hybrid network according to claim 1:
    CP = Σ i = 0 m - 1 w i * Ack [ i ] - > ecnecho
    Ack is the ACK sequence of the nearest arrival of transmit leg collection in the following formula, Ack[i]->ecnecho is the value of CE position during wherein i ACK wraps, w iBe ACK[i] weights, Σ i = 0 m - 1 w i = 1 .
  3. 3. based on the jamming control method of ECN mechanism, when the variation of congestion probability value presented the trend of increase, transmit leg was inferred as congestion state with current network in the hybrid network according to claim 1 and 2.
  4. 4. in the hybrid network according to claim 1 based on the jamming control method of ECN mechanism, only when prediction result when being congested TCP send corresponding congestion control mechanisms such as starting reduction of speed just now, otherwise the packet of retransmits lost just.
  5. 5. in the hybrid network according to claim 3 based on the jamming control method of ECN mechanism, when transmit leg is received an ACK, at first refresh the ACK array and calculate up-to-date congestion probability, if this ACK is a normal numerical value, so no matter what value the current congestion probability be, all equally with intrinsic TCP congestion probability agreement keeps Congestion Avoidance or startup slowly; If this ACK is a repeat ACK and is congested according to the congestion probability prediction result, transmit leg will carry out the speed adjustment and repeating transmission enters fast recovery so; If this ACK is a repeat ACK and is non-congested according to the congestion probability prediction result that fast recovery is only retransmitted and entered to transmit leg so.
CN2007100349629A 2007-05-21 2007-05-21 ECN mechanism-based congestion control method in the mixed network Expired - Fee Related CN101056260B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2007100349629A CN101056260B (en) 2007-05-21 2007-05-21 ECN mechanism-based congestion control method in the mixed network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2007100349629A CN101056260B (en) 2007-05-21 2007-05-21 ECN mechanism-based congestion control method in the mixed network

Publications (2)

Publication Number Publication Date
CN101056260A true CN101056260A (en) 2007-10-17
CN101056260B CN101056260B (en) 2010-07-21

Family

ID=38795875

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2007100349629A Expired - Fee Related CN101056260B (en) 2007-05-21 2007-05-21 ECN mechanism-based congestion control method in the mixed network

Country Status (1)

Country Link
CN (1) CN101056260B (en)

Cited By (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101964755A (en) * 2010-11-03 2011-02-02 中南大学 Explicit congestion control method based on bandwidth estimation in high-bandwidth delay network
CN102006230A (en) * 2010-11-26 2011-04-06 中南大学 Method for controlling congestion control method by fusing three kinds of information in wired/wireless hybrid network
CN102148662A (en) * 2011-03-21 2011-08-10 大唐移动通信设备有限公司 Adjusting method and device for data transmitting speed
CN102255898A (en) * 2011-07-05 2011-11-23 上海悠络客电子科技有限公司 Method for ensuring stability of large data stream by tcp (Transmission Control Protocol)/udp (User Datagram Protocol) double-mode communication
CN102318312A (en) * 2009-02-12 2012-01-11 英国电讯有限公司 Controlling bandwidth share
CN102415060A (en) * 2009-03-31 2012-04-11 英国电讯有限公司 Pre-congestion notification (PCN) as a base for admission control and floe termination
CN103069758A (en) * 2010-09-01 2013-04-24 瑞典爱立信有限公司 Localized congestion exposure
CN105141476A (en) * 2015-08-18 2015-12-09 大唐移动通信设备有限公司 TCP message error information obtaining method and device
WO2016015178A1 (en) * 2014-06-24 2016-02-04 华为技术有限公司 Packet loss detection method, device, and system
CN105515896A (en) * 2015-11-30 2016-04-20 青岛海信移动通信技术股份有限公司 Mobile terminal network congestion determination method and device
CN106375232A (en) * 2015-07-24 2017-02-01 华为技术有限公司 Message transmission method and message transmission system
CN107070804A (en) * 2017-03-15 2017-08-18 清华大学 Join the team and mark and go out the display congestion marking method and device that badge note is combined
CN107084732A (en) * 2011-02-03 2017-08-22 通腾发展德国公司 The method for producing expected average gait of march
CN110391956A (en) * 2019-07-23 2019-10-29 中国工商银行股份有限公司 The identification monitoring method and device of network service processes state
CN111095882A (en) * 2017-06-29 2020-05-01 华为技术有限公司 System and method for predicting flows in a network
CN111163013A (en) * 2020-03-02 2020-05-15 西南交通大学 Semi-reliable data transmission congestion control method based on UDP
CN112511323A (en) * 2019-09-16 2021-03-16 华为技术有限公司 Method and related apparatus for handling network congestion
CN113141314A (en) * 2021-04-20 2021-07-20 上海卓易科技股份有限公司 Congestion control method and equipment
CN115514709A (en) * 2022-11-23 2022-12-23 北京云脉芯联科技有限公司 Congestion control event queue scheduling method, device, equipment and storage medium

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1226844C (en) * 2003-09-01 2005-11-09 中国科学院计算技术研究所 Adaptive jamming control method suitable of mixed wired/radio network

Cited By (33)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102318312B (en) * 2009-02-12 2014-10-22 英国电讯有限公司 Controlling bandwidth share
CN102318312A (en) * 2009-02-12 2012-01-11 英国电讯有限公司 Controlling bandwidth share
CN102415060A (en) * 2009-03-31 2012-04-11 英国电讯有限公司 Pre-congestion notification (PCN) as a base for admission control and floe termination
CN102415060B (en) * 2009-03-31 2015-06-17 英国电讯有限公司 Pre-congestion notification (PCN) as a base for admission control and floe termination
CN103069758B (en) * 2010-09-01 2016-06-08 瑞典爱立信有限公司 Local congestion's process for exposing and node
CN103069758A (en) * 2010-09-01 2013-04-24 瑞典爱立信有限公司 Localized congestion exposure
CN101964755A (en) * 2010-11-03 2011-02-02 中南大学 Explicit congestion control method based on bandwidth estimation in high-bandwidth delay network
CN102006230A (en) * 2010-11-26 2011-04-06 中南大学 Method for controlling congestion control method by fusing three kinds of information in wired/wireless hybrid network
CN107084732A (en) * 2011-02-03 2017-08-22 通腾发展德国公司 The method for producing expected average gait of march
CN102148662B (en) * 2011-03-21 2014-01-15 大唐移动通信设备有限公司 Adjusting method and device for data transmitting speed
CN102148662A (en) * 2011-03-21 2011-08-10 大唐移动通信设备有限公司 Adjusting method and device for data transmitting speed
CN102255898A (en) * 2011-07-05 2011-11-23 上海悠络客电子科技有限公司 Method for ensuring stability of large data stream by tcp (Transmission Control Protocol)/udp (User Datagram Protocol) double-mode communication
US10419314B2 (en) 2014-06-24 2019-09-17 Huawei Technologies Co., Ltd. Method, apparatus, and system for packet loss detection
WO2016015178A1 (en) * 2014-06-24 2016-02-04 华为技术有限公司 Packet loss detection method, device, and system
CN106375232B (en) * 2015-07-24 2019-11-26 华为技术有限公司 A kind of message transmitting method and system
CN106375232A (en) * 2015-07-24 2017-02-01 华为技术有限公司 Message transmission method and message transmission system
CN105141476A (en) * 2015-08-18 2015-12-09 大唐移动通信设备有限公司 TCP message error information obtaining method and device
CN105141476B (en) * 2015-08-18 2018-12-18 大唐移动通信设备有限公司 A kind of acquisition methods and device of TCP message error message
CN105515896B (en) * 2015-11-30 2019-02-15 青岛海信移动通信技术股份有限公司 A kind of judgment method and device of mobile terminal network obstruction
CN105515896A (en) * 2015-11-30 2016-04-20 青岛海信移动通信技术股份有限公司 Mobile terminal network congestion determination method and device
CN107070804B (en) * 2017-03-15 2019-09-20 清华大学 It joins the team label and badge remembers the display congestion marking method and device that combines out
CN107070804A (en) * 2017-03-15 2017-08-18 清华大学 Join the team and mark and go out the display congestion marking method and device that badge note is combined
CN111095882A (en) * 2017-06-29 2020-05-01 华为技术有限公司 System and method for predicting flows in a network
CN110391956A (en) * 2019-07-23 2019-10-29 中国工商银行股份有限公司 The identification monitoring method and device of network service processes state
CN110391956B (en) * 2019-07-23 2021-08-13 中国工商银行股份有限公司 Method and device for identifying and monitoring state of network service process
CN112511323A (en) * 2019-09-16 2021-03-16 华为技术有限公司 Method and related apparatus for handling network congestion
CN112511323B (en) * 2019-09-16 2022-06-14 华为技术有限公司 Method and related apparatus for handling network congestion
US11991082B2 (en) 2019-09-16 2024-05-21 Huawei Technologies Co., Ltd. Network congestion processing method and related apparatus
CN111163013A (en) * 2020-03-02 2020-05-15 西南交通大学 Semi-reliable data transmission congestion control method based on UDP
CN111163013B (en) * 2020-03-02 2022-04-29 西南交通大学 Semi-reliable data transmission congestion control method based on UDP
CN113141314A (en) * 2021-04-20 2021-07-20 上海卓易科技股份有限公司 Congestion control method and equipment
CN115514709A (en) * 2022-11-23 2022-12-23 北京云脉芯联科技有限公司 Congestion control event queue scheduling method, device, equipment and storage medium
CN115514709B (en) * 2022-11-23 2023-02-28 北京云脉芯联科技有限公司 Congestion control event queue scheduling method, device, equipment and storage medium

Also Published As

Publication number Publication date
CN101056260B (en) 2010-07-21

Similar Documents

Publication Publication Date Title
CN101056260B (en) ECN mechanism-based congestion control method in the mixed network
CN107171842B (en) Multipath transmission protocol congestion control method based on reinforcement learning
CN107634911B (en) Adaptive congestion control method based on deep learning in information center network
CN106160953B (en) A kind of transmission method based on learning-oriented energy efficiency model
US6901593B2 (en) Active queue management with flow proportional buffering
CN102006230B (en) Method for controlling congestion control by fusing three kinds of information in wired/wireless hybrid network
CN110572333B (en) Method for adjusting sending rate in bandwidth detection stage and method for realizing congestion control algorithm
CN1327677C (en) ECN based congestion control method with prediction verification
US7729249B2 (en) Systems and methods for improving TCP-friendliness of delay-based congestion control
US20050083850A1 (en) Method for adjusting a transmission rate to obtain the optimum transmission rate in a mobile ad hoc network environment
CN101969408B (en) Active queue management method based on packet DSCP (Differentiated Services Code Point) marks
CN102291389A (en) Cross-layer congestion control method in satellite network
CN109873773B (en) Congestion control method for data center
CN104980365A (en) TCP transmission acceleration method based on continuous packet losing congestion judgment
CN112822720A (en) Cross-layer congestion control method based on MAC (media Access control) layer link quality in unmanned aerial vehicle networking technology
CN110891027B (en) Named data network transmission control method, device and equipment based on queue scheduling
Domzal et al. Congestion control in wireless flow-aware networks
Nabeshima et al. Improving the convergence time highspeed TCP
CN116760777B (en) Multipath congestion control method based on ABEA3C
Cheng et al. An threshold-based congestion control mechanism for Vegas TCP over heterogeneous wireless networks
Liu et al. Modeling and quantifying the impact of P2P file sharing traffic on traditional internet traffic
CN114785757B (en) Multipath transmission control method for real-time conversation service
CN114553783B (en) Load balancing method for self-adaptive regulation of cell granularity of data center network
Hamamreh et al. Adaptive Control of Congestion in Tough Wireless Environments Using DCM+ Algorithm
CN1949729A (en) Method for controlling RLAN performance optimizing

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20100721

Termination date: 20180521