CN108882302A - A kind of unrelated double jamming control methods of agreement - Google Patents
A kind of unrelated double jamming control methods of agreement Download PDFInfo
- Publication number
- CN108882302A CN108882302A CN201811062808.7A CN201811062808A CN108882302A CN 108882302 A CN108882302 A CN 108882302A CN 201811062808 A CN201811062808 A CN 201811062808A CN 108882302 A CN108882302 A CN 108882302A
- Authority
- CN
- China
- Prior art keywords
- congestion
- network
- wavelength
- function
- rtt
- 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
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/0289—Congestion control
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention belongs to network technique field, the unrelated double jamming control methods of specially a kind of agreement.The method of the present invention includes：The network congestion probability Q under traditional congestion control algolithm independent role is determined according to network current congestion state parametern(t)；According to Qn(t) and Pn(mn) inverse relation determine the congestion avoidance algorithm independent role based on wavelength change under network congestion probability Pn(mn)；Qn(t) and Pn(mn) co- controlling network congestion state.Advantage of the invention is that：The changeability of wavelength can reduce the congestion probability of network, and in the case where meeting some requirements, and can not influence or be promoted the overall throughput and bandwidth availability ratio of network；Traditional network congestion avoidance algorithm, which only needs to do, seldom to be improved to be adapted to Pn(mn), guarantee the flexibility and scalability of this method.Using the present invention, traditional congestion control algolithm can be optimized, promote network throughput and bandwidth availability ratio.
Description
Technical field
The invention belongs to network technique fields, and in particular to a kind of method for controlling network congestion.
Background technique
Congestion control is always the research hotspot in network field, and main cause is that change constantly occurs for the state of network
Change, and out-of-date congestion avoidance algorithm is not suitable with current network environment, or is not optimal, so needing to improve.Traditional
Congestion avoidance algorithm, such as TCP reno, TCP new reno, HSTCP, HTCP, UDT, BIC, CUBIC, BBR, QUIC etc. are all
It is that adjusting is carried out based on packet loss, RTT or available bandwidth.These algorithms excessively rely on the procedure parameter in congestion, such as NAK
(packet loss feedback signal), ACK (answer signal), RTT (round-trip delay), Bandwidth etc., and many parameters are difficult standard here
Really measurement, along with network state is ever-changing, so traditional network congestion avoidance algorithm is extremely difficult to global optimum.Especially
Under BDP very big wide area network, bandwidth availability ratio and throughput have to be hoisted.
Irregular saw tooth wave shape is presented in network flow, can be analogous to sine wave and go to analyze.Can with mathematics and
It is demonstrated experimentally that the wavelength of adjustment sine wave, that is, business increase window speedup or drop the acceleration of window reduction of speed, network can be reduced
Congestion probability, and promote network overall throughput and bandwidth availability ratio.The wavelength of sine wave, with one in congestion avoidance algorithm
A little variables are related, such as gain factor in BBR etc., adjust wavelength, are equivalent to the value for adjusting these gain factors.On the one hand it is
Traditional congestion control algolithm adjusts network congestion condition, is on the other hand that the method based on wavelength further improves congestion shape
State, so last network congestion probability can be relatively low.For example, can not only drop window when network will or congestion occur
Reduction of speed can also increase wavelength, to further improve network congestion condition.The adjusting of wavelength can be with the ginseng in congestion
Number is unrelated, only with whether congestion is related, and is also that agreement is unrelated.
Summary of the invention
The purpose of the present invention is to provide a kind of unrelated double jamming control methods of agreement, thus promoted network throughput and
Bandwidth availability ratio.
The unrelated double jamming control methods of agreement provided by the invention are by by traditional congestion control algolithm and wavelength tune
The method of section combines adjustment network congestion, and then reduces the probability of network congestion, promotes network throughput and bandwidth benefit
With rate.
The unrelated double jamming control methods of agreement provided by the invention, basic ideas are：By traditional congestion control algolithm into
Row segmentation, determines every section of argument of function, using the independent variable as the wavelength for influencing this section of function；Monitor throughput, CWND
Or the curve of packet sending speed these variable changes, the curve can analogize to the probability point of traditional congestion algorithm lower network congestion
Cloth function；The probability density function of traditional congestion control algolithm (being denoted as Q) is found out according to distribution function, and is acquired in turn based on wave
The probability density function of the long congestion avoidance algorithm (being denoted as P) adjusted, then determines wavelength m according to optimization methodn, QP-Wave
It is according to mnTo carry out dynamic regulation to network congestion condition in real time.
The unrelated double jamming control methods of agreement provided by the invention, the specific steps are：
(1) sinusoidal fluctuation function is converted by network flow to go to analyze.Using Fourier transformation by the time domain of network flow
It is transformed into frequency domain, acquires the data such as throughput, rate, the CWND of network flow in real time；By traditional congestion control algolithm according to letter
It counts continuity itself, raising and lowering stage and congestion phases to be segmented, and extracts every section of argument of function x, such as
RTT, NAK, Loss rate, ACK, Bandwidth etc.；
(2) distribution function is sought.The characteristics of according to different congestion avoidance algorithms, finds the parameter for influencing congestion state, so
After extract the relevant distribution function with it.Such as the congestion state of UDT is mainly determined by packet loss, so being distributed to UDT
The extraction of function can focus on the packet sending speed or throughput situation of change of packet loss state lower network.It is discrete due to RTT
Type and uncertainty, the distribution function be it is discrete, traditional congestion control algolithm (being denoted as Q) can be determined by the distribution function
Probability density function and the congestion avoidance algorithm (being denoted as P) based on wavelength probability density function；
(3) Q is acquired by above-mentioned distribution function, P is acquired according to the inverse relation of Q and P；According to the congestion based on wavelength regulation
Probability density function and congestion probability limits value PlimitDetermine wavelength mn, by wavelength mnNet is adjusted jointly with traditional congestion algorithm
Network congestion state；In real time according to distribution function update probability density function, and congestion state parameter feedback is given and is based on wavelength tune
The congestion probability density function of section updates wavelength mn.By this traditional congestion control algolithm (being denoted as Q) and based on wavelength regulation
Congestion avoidance algorithm (being denoted as P) combines, according to mnThe method for carry out dynamic regulation to network congestion condition in real time, is denoted as QP-
Wave。
In step (1) of the present invention, analysis network flow is gone using sinusoidal fluctuation function and Fourier transformation, will be gathered around to network
The research of plug control provides new direction, and is also beneficial to carry out quantitative analysis to web impact factor.
In step (1) of the present invention, distribution function can be piecewise function, and distribution function is measured and locates according to real time data
Reason obtains, such as according to the handling capacity of network interface card, the packet sending speed of business, packet loss function call number and phase even in program
Variable change value is closed to determine.
In step (1) of the present invention, the extraction to independent variable x mainly considers congestion control algorithm f (mnX) with what parameter
It is measured as congestion；Adjust wavelength mn, can be so that f (mnX) curve controls business as horizontal axis broadens or narrows
Lead to the probability of network congestion.
In step (2) of the present invention, the extraction of distribution function can determine Q and P, and the extraction of distribution function depends on reality
When the data tested.
In step (3) of the present invention, the probability density function of QP-Wave can cooperate other parameters to determine current time most
Excellent wavelength mn；By Q and P commonly through adjusting wavelength mnTo realize double congestion control QP-Wave.
Advantage of the invention is that：
It, can be with 1. the changeability of wavelength will will be greatly reduced the congestion probability of network, and in the case where meeting some requirements
Do not influence or promoted the overall throughput and bandwidth availability ratio of network；
It seldom improves 2. traditional network congestion avoidance algorithm only needs to do to be adapted to Pn(mn) parameter interface, ensure that this
The flexibility and scalability of method；
3. network flow is analogous to sinusoidal wave function analysis, it will provided newly to the optimization of traditional congestion control algolithm
Thinking.
Using the present invention, traditional congestion control algolithm can be optimized, promote network throughput and bandwidth availability ratio, and be net
Network congestion and the optimization problem of other field provide new solution.
Detailed description of the invention
Fig. 1 the method for the present invention frame diagram.
Theoretical modeling of Fig. 2 present invention to network flow.
Q and P compares the influence that data packet is transmitted in Fig. 3 present invention.
P utilizes the compensation of bandwidth in Fig. 4 present invention.
The advantage of the bis- jamming control method QP-Wave of Fig. 5.
Fig. 6 Qn(t, NAK, ACK, RTT...) and Pn(mn) relation curve.
Fig. 7 QP-Wave implementation flow chart.
Fig. 8 not adjusting wavelength when UDT packet loss curve.
Packet loss curve when Fig. 9 adjusting wavelength is 1.005.
The relationship of Figure 10 wavelength shift and file transmission time.
Specific embodiment
It is convenient for statement, traditional congestion control algolithm is known as Q below, the congestion avoidance algorithm based on wavelength is known as P, Q
It is independent from each other with P.Realization frame of the invention is as shown in Figure 1, wherein Qn(t, NAK, ACK, RTT...) is that Q effect is off line
The distribution function F of network flown(t, NAK, ACK, RTT...) derivation；Wherein, t is the time, and NAK is packet loss feedback signal,
ACK is answer signal, and RTT is round-trip delay.Here independent variable determines according to Q itself, such as based on packet loss, Q just only with
NAK or packet loss are related；Based on delay, Q is just only with RTT correlation.Pn(mn) be P congestion probability, it only with wavelength mnHave
It closes.Qn(t, NAK, ACK, RTT...) regulates and controls network congestion condition, P in itselfn(mn) further pass through control wavelength mnTo regulate and control
Network congestion, so the present invention is double congestion controls to network.
The present invention is described in detail with reference to the accompanying drawing.
1, the theoretical modeling of network flow and analysis
Traditional congestion control algolithm is confined to RTT to the analysis of network flow, bandwidth, the discrete amount such as packet loss, and this hair
It is bright that the utilizing thoughts of Fourier transformation come, network flow is modeled as sine wave and goes to analyze, as shown in Figure 2.Data packet passes
Defeated rate can be segmented into v by sinusoidal wave function approximationn(t)≈an+An|sinmnT |, and handling capacity is its piecewise function
Integral：
At this moment, network flow frequency domain can be transformed into from time-domain analysis to analyze.The drop window reduction of speed and increasing window speedup of Q
Process, regardless of how rate and throughput change, how is network congestion, and acceleration is generally consistent, i.e., all
Business all substantially concentrates on some Frequency point in frequency domain, unless Q itself can automatically adjust this acceleration, and actually
This acceleration of Q or gain are fixed value or very limited adjusting, are thus easy to cause network congestion.P algorithm
A possibility that main thought is that acceleration or gain dynamic change are broken up Frequency point, and the business resonance of reduction generates congestion.And
The adjusting of P will depend on Pn(mn) m that is determinedn, and Pn(mn) by Qn(t, NAK, ACK, RTT...) and PlimitIt determines.It can be with
Mathematical proof, in mnIn the case that mutual size relation meets least common multiple, the adjusting of P will not influence theoretical network and handle up
Amount, but will be greatly reduced network congestion probability, and in fact, the reduction of network congestion probability, it will make goodput or band
Wide utilization rate rises.
To the analysis of the data transmission procedure of Q with P as shown in figure 3, the acceleration due to Q is consistent, cause statistically
There is consistent rising behavior in data packet, is easy to cause congestion, and consistent decline behavior after congestion, and results in bandwidth wave
Take；And thought of the P based on wavelength change, the acceleration of data packet can be set to difference, the probability of their congestions in this way is reduced,
Even if there is congestion, too big bandwidth waste will not be caused.That is, P has certain compensating action to bandwidth, such as scheme
Shown in 4, and Q can slattern some bandwidth, performance it is excessively radical.QP-Wave combines the advantages of both P and Q, such as Fig. 5 institute
Show, when congestion occurs in Q, P adjusts the Congestion Level SPCC of Q by wavelength control；When there is congestion in P, Q also can according to itself
Congestion mechanism adjusts the Congestion Level SPCC of P；By QP by the common regulating networks congestion of wave Wave, to make network congestion probability most
Small, this is also the advantage of QP-Wave.
2, the distribution function of Q is sought
In order to seek the probability density function Q of Qn(t, NAK, ACK, RTT...), we can first seek the distribution function of Q
Fn(t, NAK, ACK, RTT...), the function are that real-time measurement throughput or rate obtain.QP-Wave is that agreement is unrelated
, it is meant that can be adapted for various Q algorithms, but the congestion control state parameter of every kind of Q is different, key see Q mainly with
That factor is related, for example for Cubic, Q is mainly related with packet loss, then distribution function is exactly the letter about packet loss independent variable
Number；For BBR, Q has certain relationship with available bandwidth, packet loss, RTT, then the specification of variables of distribution function, will just consider
These factors.Distribution function is probability density that is real-time and dynamic, being only used in prediction a period of time.Distribution function because
Variable can be handling capacity, packet sending speed and CWND etc..
3, the probability density function and m of Q and P are soughtn
The probability density function Q of Qn(t, NAK, ACK, RTT...) can be acquired by the distribution function of Q, and seek the general of P
Rate density function, mainly seeks mn, because P can be exported by above-mentioned Sinusoid Model and mathematical proofn(mn)=mn/
(m1m2...mn), wherein m is the natural wavelength of Q, mnFor the variable adjustment wavelength of P, it can be seen that Pn(mn) only with mnIt is related to m,
With NAK, ACK, RTT etc. are unrelated, and it is unrelated to be equivalent to agreement.Assuming that the probability that a business causes network congestion to occur is A, due to
P and Q is independent event, so the probability P that n business causes network congestion to occursumMeet bi-distribution.To sum up institute
It states, there is following relationship establishment：
1-(1-Pn(mn))(1-Qn(t, NAK, ACK, RTT...))=A
In order to acquire mn, need to set an optimization problem：We always want to the bandwidth usage for improving total business as far as possible
Rate or throughput.It is assumed that the link throughput upper limit is Tmax, actual throughput rates Tactual, then the optimization problem is denumerable
It is expressed as on：
Minimize Δ T=Tmax-Tactual
Subject to 1-(1-Pn(mn))(1-Qn(t, NAK, ACK, RTT...))=A
Pn(mn)min≤Pn(mn)≤Pn(mn)max
Pn(mn)∝1/Qn(t, NAK, ACK, RTT...) (inverse relation is as shown in Figure 6)
Boundary value in above-mentioned constraint condition can go to determine according to the boundary value of distribution function.It can be excellent by this
Change problem finally acquires the m under current real-time conditionsn.Q passes through mnDe-regulation P, the final double congestion controls for realizing that agreement is unrelated
QP-Wave, implementation flow chart are as shown in Figure 7.
It is above-mentioned about description of the invention only for the detailed description of embodiments of the present invention, not due to limitation
Protection scope of the present invention, corresponding field technical staff it should be appreciated that all without departing from equivalent made by the technology of the present invention essence
Embodiment or change should be all included within the scope of the present invention.
Illustrate realization process of the invention by taking a kind of UDT (congestion avoidance algorithm) algorithm as an example below, other congestions are calculated
The realization process of method is similar.
We analyze the congestion avoidance algorithm feature of UDT first, it is the testing mechanism using packet loss as congestion state,
NAK is its packet loss detection signal.So Qn(t, NAK, ACK, RTT...) can be expressed as Qn(t,NAK).Congestion avoidance algorithm
Having and increase window speedup before congestion occurs, the process of drop window reduction of speed after congestion occurs, these processes have corresponded to different piecewise functions,
Each piecewise function can have the wavelength of oneself.Since UDT is the signal using packet loss as detection congestion state, it is possible to
Wavelength is adjusted after congestion, observes bandwidth availability ratio situation.It is exactly to modify void CUDTCC in corresponding UDT code::onLoss
Some algorithm in (const int32_t*losslist, int) function, such as m_dPktSndPeriod=ceil (m_
DPktSndPeriod*wavelength the wavelength in).We further control net by adjusting wavelength
The probability of network congestion, so that the algorithm of UDT itself be cooperated to realize the two supports match control, key is to choose wavelength most here
Good value.
Secondly, first seeking distribution function F to seek the optimum value of wavelengthn(t, NAK, ACK, RTT...), by
In distribution function only with packet loss correlation, it is possible to be reduced to Fn(t,NAK).Distribution function is for probability to be calculated
Density, and then obtain the optimum value of wavelength.Since the distribution function is with packet loss correlation, so we want extract real-time
The packet drop of business.Static variable can be set in program, and number, time and the transmission speed of packet loss occur for record traffic
Rate.The relationship of packet loss number and time is fitted to line chart, and it is smooth, finally just obtains distribution function.
Finally, pair distribution function derivation can obtain probability density function QnP is arranged in (t, NAK)limit=20%, net
Cassette tape width is 11M or so, and packet loss is arranged with software.So 1- (1-Pn(mn))(1-Qn(t, NAK))=20%.It needs to find out
Δ T=11-TactualMinimum value.Then different packet loss is set, analyzed, be distributed using wireshark packet capturing in real time
Function, finally acquiring optimal wavelength is wavelength=1.005 or so, this also relatively coincide with actual performance.Test
Correlation curve is as follows：
Fig. 8 not adjusting wavelength when UDT packet loss curve.The curve describes the test curve of original UDT, original to mean
Keep the fixed wave length wavelength=1.125 of UDT constant.The case where testing four kinds of different packet loss rates altogether, it is found that
File is sent completely required time longest when packet loss is 1% (Packet loss=0.01), and packet sending speed is minimum, to gathering around
The adaptation of plug is poorer, and congestion probability is higher, and drop window reduction of speed is more frequent, wastes fractional bandwidth.
Packet loss curve when Fig. 9 adjusting wavelength is 1.005.Compared to the curve in Fig. 8, no matter under which kind of packet loss,
File transmission time shorten to original one third when wavelength=1.005, and transmission fluctuation is smaller, more flat
Surely, illustrate under identical packet loss congested environment, this wavelength is stronger to the adaptation of congestion, and congestion avoidance algorithm judges that network is gathered around
The probability is relatively small for plug, and drop window reduction of speed improves bandwidth availability ratio than more gentle.
The relationship of Figure 10 wavelength shift and file transmission time.The figure is fixed packet loss Packet loss=0.01's
In the case of, the change of test wavelength and the relationship of file transmission time, it can be found that wavelength is bigger from figure, when file transmits
Between it is shorter, wavelength is smaller, and file transmission time is bigger.But wavelength is too big, be easy to cause network congestion, influences other business biography
Defeated, wavelength is too small, and packet sending speed is caused to reduce, and bandwidth availability ratio is poor, so their optimum value is the intersection point in figure, just
It is located near 1.005 well, the conclusion that optimal wavelength is 1.005 when this also just demonstrates above-mentioned theory analysis demonstrates this hair
Bright validity.
The parameter of table 1 final test and calculating.
| Parameter | Value |
| Practical optimal wavelength | 1.005 |
| File size | 208MB |
| Packet loss | 1% |
| Transmission time | 63s |
| Average packet sending speed | 3.3Mbps |
| Performance improvement | Nearly 2 times |
| Maximum bandwidthBWmax | 11Mbps |
| Theoretical optimal wavelength | About 1.005 |
| Plimit | 20% |
The table is the partial parameters in experiment.As can be seen that UDT congestion avoidance algorithm is in packet loss environment using the present invention
Under, that is, when congestion may occur for network, nearly twice of performance boost.For other congestion avoidance algorithms, the present invention
It is also effectively, because of all congestion avoidance algorithms, basic process is similar, that is, drop window reduction of speed when congestion, do not have
Increase window speedup when congestion, traditional congestion algorithm is substantially coincident to the increase or reduction rate that increase window speedup and drop window reduction of speed
, and present invention trial goes to change this rate, that is, wavelength, has good effect as the result is shown.And the present invention is only
It is only to change wavelength, there is no congestion avoidance algorithm itself is changed, so being double congestion controls.
Claims (5)
1. a kind of unrelated double jamming control methods of agreement, which is characterized in that the specific steps are：
(1) sinusoidal fluctuation function is converted by network flow to go to analyze；The time domain of network flow is converted using Fourier transformation
To frequency domain, throughput, the rate, CWND data of network flow are acquired in real time；By traditional congestion control algolithm according to function itself
Continuity, raising and lowering stage and congestion phases are segmented, and extract every section of argument of function x, and independent variable x is
RTT, NAK, Loss rate, ACK and/or Bandwidth；
(2) distribution function is sought；The characteristics of according to different congestion avoidance algorithms, finds the parameter for influencing congestion state, then mentions
Take the relevant distribution function with it；Wherein, the congestion state of UDT is mainly determined by packet loss, so being distributed to UDT
The extraction of function can focus on the packet sending speed or throughput situation of change of packet loss state lower network；It is discrete due to RTT
Type and uncertainty, the distribution function be it is discrete, traditional congestion control algolithm (being denoted as Q) can be determined by the distribution function
Probability density function and the congestion avoidance algorithm (being denoted as P) based on wavelength probability density function；
(3) Q is acquired by above-mentioned distribution function, P is acquired according to the inverse relation of Q and P；According to the congestion probability based on wavelength regulation
Density function and congestion probability limits value PlimitDetermine wavelength mn, by wavelength mnNetwork is adjusted jointly with traditional congestion algorithm to gather around
Plug-like state；The probability density function of Q is updated according to distribution function in real time, and by congestion state parameter feedback to the probability density of P
Updated optimal wavelength m is finally calculated in functionn；
This traditional congestion control algolithm (Q) and the congestion avoidance algorithm (P) based on wavelength regulation are combined, according to mnCome real-time
To the method that network congestion condition carries out dynamic regulation, it is denoted as QP-Wave.
2. the unrelated double jamming control methods of agreement according to claim 1, which is characterized in that in step (1), be distributed letter
Number is piecewise function, and distribution function is measured according to real time data and processing obtains.
3. the unrelated double jamming control methods of agreement according to claim 1, which is characterized in that in step (1), to change certainly
The extraction for measuring x, mainly considers congestion control algorithm f (mnWhat x) measured using parameter as congestion；Adjust wavelength mn, can so that
f(mnX) curve is as horizontal axis broadens or narrows, and then the probability that regulating networks congestion occurs.
4. the unrelated double jamming control methods of agreement according to claim 1, which is characterized in that in step (2), be distributed letter
To determine Q and P, the extraction of distribution function depends on the data that real-time testing arrives for several extractions.
5. the unrelated double jamming control methods of agreement according to claim 1, which is characterized in that in step (3), seek Q
With the probability density function and m of Pn, realize that the detailed process of the unrelated double congestion control QP-Wave of agreement is as follows：
If the distribution function of network flow is Fn(t, NAK, ACK, RTT...), the probability density function of Q are Qn(t,NAK,ACK,
RTT...), wherein t is the time, and NAK is packet loss feedback signal, and ACK is answer signal, and RTT is round-trip delay；Qn(t,NAK,
ACK, RTT...) it is the distribution function F that Q acts on lower network flown(t, NAK, ACK, RTT...) derivation and obtain；And seek P
Probability density function, mainly seek mn, because P can be exported by above-mentioned Sinusoid Model and mathematical proofn(mn)=
mn/(m1m2...mn), wherein m is the natural wavelength of Q, mnFor the variable adjustment wavelength of P, P heren(mn) only with mnIt is related to m, with
NAK, ACK, RTT are unrelated；Assuming that the probability that a business causes network congestion to occur is A, since P and Q is independent thing
Part, so the probability P that n business causes network congestion to occursumMeet bi-distribution；There is following relationship establishment：
1-(1-Pn(mn))(1-Qn(t, NAK, ACK, RTT...))=A
In order to acquire mn, need to set an optimization problem：People always want to as far as possible improve total business bandwidth availability ratio or
Throughput；It is assumed that the link throughput upper limit is Tmax, actual throughput rates Tactual, then the optimization problem is mathematically stated
For：
Minimize Δ T=Tmax-Tactual
Subject to 1-(1-Pn(mn))(1-Qn(t, NAK, ACK, RTT...))=A
Pn(mn)min≤Pn(mn)≤Pn(mn)max
Pn(mn)∝1/Qn(t,NAK,ACK,RTT...)
Boundary value in above-mentioned constraint condition goes to determine according to the boundary value of distribution function；It is finally asked by this optimization problem
Obtain the m under current real-time conditionsn；Q passes through mnDe-regulation P, the final double congestion control QP-Wave for realizing that agreement is unrelated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811062808.7A CN108882302B (en) | 2018-09-12 | 2018-09-12 | Protocol-independent double-congestion control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811062808.7A CN108882302B (en) | 2018-09-12 | 2018-09-12 | Protocol-independent double-congestion control method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108882302A true CN108882302A (en) | 2018-11-23 |
| CN108882302B CN108882302B (en) | 2021-04-30 |
Family
ID=64324079
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811062808.7A Active CN108882302B (en) | 2018-09-12 | 2018-09-12 | Protocol-independent double-congestion control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108882302B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090067335A1 (en) * | 2007-07-06 | 2009-03-12 | Telefonaktiebolaget Lm Ericsson (Publ) | Congestion control in a transmission node |
| CN102377631A (en) * | 2010-08-06 | 2012-03-14 | 北京乾唐视联网络科技有限公司 | Flow-control-based data transmission method and communication system |
| CN108259364A (en) * | 2016-12-29 | 2018-07-06 | 中国移动通信集团浙江有限公司 | A kind of network congestion determines method and device |
-
2018
- 2018-09-12 CN CN201811062808.7A patent/CN108882302B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090067335A1 (en) * | 2007-07-06 | 2009-03-12 | Telefonaktiebolaget Lm Ericsson (Publ) | Congestion control in a transmission node |
| CN102377631A (en) * | 2010-08-06 | 2012-03-14 | 北京乾唐视联网络科技有限公司 | Flow-control-based data transmission method and communication system |
| CN108259364A (en) * | 2016-12-29 | 2018-07-06 | 中国移动通信集团浙江有限公司 | A kind of network congestion determines method and device |
Non-Patent Citations (3)
| Title |
|---|
| SHAVETA RANI 等: "Survivability strategy with congestion control in WDM optical networks", 《2007 INTERNATIONAL SYMPOSIUM ON HIGH CAPACITY OPTICAL NETWORKS AND ENABLING TECHNOLOGIES》 * |
| 孙中凯: "光突发交换网络中拥塞控制方法的研究", 《中国优秀硕士学位论文全文数据库(信息科技I辑)》 * |
| 葛一楠 等: "一种改进的高速TCP拥塞控制算法", 《计算机科学》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108882302B (en) | 2021-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8904025B2 (en) | Systems and methods for traffic management | |
| CN105281981B (en) | The data traffic monitoring method and device of network service | |
| EP1836813B1 (en) | Method of controlling packet flow | |
| Inoue | Analysis of the age of information with packet deadline and infinite buffer capacity | |
| WO2021057382A1 (en) | Abnormality detection method and apparatus, terminal, and storage medium | |
| CN104734916B (en) | A kind of high-efficiency multi-stage anomalous traffic detection method based on Transmission Control Protocol | |
| EP3110090B1 (en) | System and method for heuristic control of network traffic management | |
| CN113055923B (en) | Mobile network traffic prediction method, device and equipment | |
| CN104702623B (en) | IP blockage method and system | |
| CN102883352A (en) | GSM (global system for mobile communications) cell parameter optimization method based on traffic modeling and traffic prediction | |
| CN106254394B (en) | A kind of recording method and device of attack traffic | |
| WO2020093502A1 (en) | Nominal bandwidth adjusting method and device | |
| CN114615294B (en) | Electric power internet of things gateway edge calculation method | |
| CN110278611A (en) | A kind of resource allocation methods in the mobile edge calculations system of wireless power | |
| CN110336799A (en) | A kind of network defense method and its terminal towards edge calculations terminal | |
| CN109257252A (en) | A method of judging network congestion condition | |
| CN106714189A (en) | Method and device for analyzing cell over coverage | |
| CN108882302A (en) | A kind of unrelated double jamming control methods of agreement | |
| DE102021109229A1 (en) | BURST PACKET Biasing for Dev Estimation with Rate Limiters | |
| CN112988892A (en) | Distributed system hot spot data management method | |
| Lv et al. | A traffic interval prediction method based on ARIMA | |
| CN113055333A (en) | Network flow clustering method and device capable of self-adaptively and dynamically adjusting density grids | |
| DE102022131778A1 (en) | ADAPTIVE WIRELESS CONNECTIONS IN MULTIPLE OPERATING MODE DEVICES | |
| CN101834843B (en) | Method for establishing relation between fire bulkhead throughput rate and number and sort order of filer rules | |
| WO2020233438A1 (en) | Transaction link decision-making method, apparatus and device, and computer-readable storage medium |
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 |