WO2013044438A1 - Procédé et appareil d'évitement de congestion - Google Patents

Procédé et appareil d'évitement de congestion Download PDF

Info

Publication number
WO2013044438A1
WO2013044438A1 PCT/CN2011/080189 CN2011080189W WO2013044438A1 WO 2013044438 A1 WO2013044438 A1 WO 2013044438A1 CN 2011080189 W CN2011080189 W CN 2011080189W WO 2013044438 A1 WO2013044438 A1 WO 2013044438A1
Authority
WO
WIPO (PCT)
Prior art keywords
threshold
queue
preset
offset
congestion
Prior art date
Application number
PCT/CN2011/080189
Other languages
English (en)
Chinese (zh)
Inventor
王爽
Original Assignee
中兴通讯股份有限公司
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 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Priority to PCT/CN2011/080189 priority Critical patent/WO2013044438A1/fr
Publication of WO2013044438A1 publication Critical patent/WO2013044438A1/fr

Links

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
    • H04L47/30—Flow control; Congestion control in combination with information about buffer occupancy at either end or at transit nodes
    • 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
    • H04L47/32—Flow control; Congestion control by discarding or delaying data units, e.g. packets or frames
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00—Packet switching elements
    • H04L49/90—Buffering arrangements
    • H04L49/9084—Reactions to storage capacity overflow

Definitions

  • the present invention mainly relates to the field of network transmission control, and more particularly to a method and apparatus for avoiding congestion of a router having adaptive characteristics. Background technique
  • Network congestion is likely to cause degradation in quality of service (QoS) performance such as delay and throughput, which is a key factor affecting resource utilization such as bandwidth and system cache. Therefore, an effective congestion avoidance mechanism is important for improving router performance.
  • QoS quality of service
  • Active queue management Including RED and its derivative algorithms. For example, Weighted Random Early Detection (WRED) algorithm. By checking the usage of the queue, the enqueue packets are discarded randomly when the congestion is intensified.
  • WRED Weighted Random Early Detection
  • FIG. 1 is a schematic structural diagram of a congestion avoidance system based on a TD and WRED mechanism in the prior art.
  • the WRED mechanism or the TD mechanism is selected according to the discard policy configured by the user.
  • the TD mechanism is relatively simple. If the sum of the new packet length and the queue depth of the queue to which the packet belongs is greater than the TD threshold, the new packet is discarded, and the queue depth is not updated; otherwise, the new packet is enqueued, and the system cache and the queue to be updated are simultaneously updated. Queue depth (plus new packet length).
  • the WRED mechanism is relatively complex.
  • the average queue depth avg_q_len is calculated based on the packet length and the queue depth of the queue to which it belongs.
  • the average queue depth is compared to the discard threshold. If the average queue is deep If the degree is less than the lower threshold threshold wred_min-th, the packet is directly enqueried; if the average queue depth is greater than or equal to the upper threshold threshold wred_max-th, the packet is directly discarded.
  • the discarding probability p of the packet is calculated according to the packet length and the average queue depth.
  • Both the WRED and TD policies are based on pre-set queue drop thresholds. Although the implementation is convenient, it lacks adaptability.
  • the main parameters that affect the packet discard judgment are the maximum discard threshold wred max th and the minimum discard threshold wred min th.
  • the maximum discard threshold wred max th and the minimum discard threshold wred min th When the average queue depth avg q len is greater than wred—max—th, 4 ⁇ is discarded; when the average queue depth is less than wred—min—th, the text is enqueued.
  • These two thresholds are critical to determining if a message is enqueued. When the traffic is small and the cache space is large, you should ensure that the packets of the current queue can be enqueued. If the value of wred max th and wred min th is set to ' J, the pre-set value will cause some queues to be discarded.
  • the queue with high priority should be ensured. Try to discard the packets of the superior low-level first. If the values of wred-max-th and wred-min-th are set in advance, many messages will continue to be queued, and the cache will eventually overflow. The message could not be entered.
  • the TD mechanism has the same problem, assuming that the system can support a number of queues, and all queues share the system cache. When all queues are active, you can make full use of the system cache space. If only a small number of queues are active (assuming only 2 queues), the fixed discard threshold limits the enqueue of packets from the above queues, making these queues enter the discarding phase prematurely, and the system cache is sufficient. Therefore, the fixed drop threshold severely affects the utilization of the system cache.
  • the present invention provides a congestion avoiding method and apparatus for solving the problem that the existing congestion avoidance mechanism cannot fully utilize the current system congestion degree to discard the packet discarding, and cannot avoid the buffer overflow risk caused by the congestion problem. And the problem of low cache utilization.
  • the present invention provides a congestion avoiding method, where the method includes: a) obtaining a congestion level and a congestion effective indication according to a system cache occupancy condition;
  • the method for obtaining the congestion effective indication is: when the system buffer is less than the minimum cache threshold or greater than the maximum cache threshold, the congestion indication is invalid, otherwise the congestion indication is valid.
  • selecting the amplification factor according to the congestion level and the congestion effective indication is specifically: bl) when the congestion indication is valid, selecting different amplification factors according to the obtained congestion level; b2) When the congestion level is invalid, the amplification factor is 0.
  • the discarding threshold of the queue to which the preset new incoming packet belongs is obtained, and the step value of the queue is calculated as follows:
  • the discarding threshold of the queue to which the preset new incoming packet belongs is obtained, and the step value of the queue is calculated as follows:
  • the product of the preset offset percentage for the queue and the preset maximum discard threshold is used as the maximum step value for adjusting the current maximum discard threshold, which will be for the queue a product of a preset offset percentage and the preset minimum discard threshold as a minimum step value for adjusting a current minimum discard threshold;
  • the step values for adjusting the current maximum discard threshold and for adjusting the current minimum discard threshold are equal to the preset maximum discard threshold or the minimum discard threshold and for the queue The product of the preset offset percentage.
  • the calculating the threshold offset according to the step value and the amplification factor is specifically: obtaining the threshold offset by multiplying the step value and the amplification factor.
  • the discarding threshold of the queue to which the new incoming packet belongs is adjusted by the threshold offset:
  • the sum of the discarding threshold of the queue to which the preset new incoming packet belongs and the threshold offset is calculated, and the adjusted discarding threshold is obtained, and the discarding threshold is adjusted by the adjusted discarding threshold.
  • the present invention further provides a congestion avoidance apparatus, where the apparatus includes: a congestion level determination module, configured to detect a system cache depth, generate a congestion level, and effectively indicate a congestion;
  • a step value calculation module configured to obtain a discarding threshold of a queue to which a preset new incoming message belongs, and calculate a step value of the queue
  • a threshold offset module configured to: after receiving the congestion effective indication and the congestion level sent by the congestion level determining module, search for a corresponding amplification factor, and calculate a threshold offset according to the amplification factor and the step value sent by the step value calculation module; The sum of the preset discarding threshold and the threshold offset is calculated to obtain an adjusted discarding threshold, and the current discarding threshold of the queue to which the incoming packet belongs is adjusted.
  • the step value calculation module further includes: a first threshold value obtaining module, configured to acquire, according to a queue number carried in the new incoming message, a preset discard of the queue to which the new incoming message belongs Threshold value
  • the first step value calculation module is configured to calculate the product of the offset percentage preset for the queue to which the new incoming message belongs and the preset discarding threshold to obtain the step value.
  • the step value calculation module further includes: a second threshold value obtaining module, configured to acquire, according to a queue number carried in the newly forwarded message, a preset maximum queue to which the new incoming message belongs Discard threshold and minimum discard threshold;
  • a second step value calculation module configured to calculate a maximum step value for adjusting a current maximum discard threshold and a minimum step for adjusting a current minimum discard threshold in a single offset mode or an overall offset mode
  • the maximum step value used to adjust the current maximum discard threshold is the product of the preset offset percentage and the preset maximum discard threshold, and is used to adjust the current minimum discard threshold.
  • the minimum step value is the product of the preset offset percentage and the preset minimum discard threshold; in the overall offset mode, it is used to adjust the current maximum discard threshold and to adjust the current minimum discard threshold.
  • the step value is equal to the product of the preset maximum discard threshold or the minimum discard threshold and the offset percentage preset for the queue.
  • the threshold offset module includes:
  • An amplification factor selection module configured to select an amplification factor corresponding to the congestion level according to the congestion effective indication and the congestion level;
  • a threshold offset calculation module configured to calculate a product of a large factor and a step value obtained from the step value calculation module, and output a threshold offset
  • the adjustment module is configured to calculate a sum of a reference drop threshold of the system preset and the threshold offset, and obtain an adjusted discard threshold, and adjust the current discard threshold of the queue by the adjusted discard threshold.
  • the present invention adaptively adjusts the discarding threshold according to the cache usage of system level maintenance. If the system cache is sufficient, increase the discarding threshold and reduce the probability of packet discarding. This ensures that more packets are enqueued and the system cache is fully utilized. When the system cache is small, the discarding threshold is decreased. The probability of large packets being dropped reduces the number of packets enqueued, thus protecting the system cache from overflow.
  • the technical solution of the invention improves the performance of system congestion avoidance and improves the whole system Cache utilization.
  • the device adopts a percentage offset method to maintain the quality of service between the queues.
  • the discarding threshold of the high-priority queue is increased by a larger amount than the low one, and the difference in QoS is clearly reflected.
  • Figure 2 is a congestion avoidance device structure with adaptive characteristics
  • FIG. 1 Schematic diagram of congestion level judgment
  • the basic idea of the present invention is to provide a router congestion avoiding method and apparatus with adaptive characteristics.
  • the method and apparatus adopt a dynamic threshold mechanism to dynamically adjust the queue discarding threshold according to the cache usage of the system level maintenance.
  • the adjustable discard threshold includes, but is not limited to, the tail discard threshold td-th in the TD mechanism, the maximum discard threshold wred_max-th and the minimum discard threshold wred_min-th in the WRED mechanism, and other congestion avoidance algorithms Dynamic thresholding mechanisms can be supported.
  • the discarding thresholds (such as td-th, wred-max-th and wred-min-th) are increased to reduce the probability of packet discarding, thereby ensuring that more packets are enqueued;
  • reduce the discard thresholds (such as td-th, wred-max-th and wred-min-th), increase the phase-to-no rate of discarding, and reduce the enrollment. Quantity.
  • FIG. 1 is a schematic diagram of the application of the congestion avoidance device and the scenario shown in FIG. 1 according to the embodiment of the present invention.
  • the congestion avoidance device 200 provided by the present invention includes the following modules: a congestion level determination module 201, a step value calculation module 202, and a threshold. Offset module 203.
  • the congestion level determining module 201 is configured to detect a system cache depth, generate a congestion level, and generate a congestion indication.
  • the congestion state determination module 201 monitors the system cache usage by detecting the cache depth of the system, determines the severity of the network congestion, and generates a congestion level indication signal; when the new packet enters the system, the The module reads the current system cache value sys_q_len(System Queue Length). If the system cache has exceeded the pre-configured minimum cache threshold and is less than the maximum cache threshold, the module will generate a congestion valid indication cg_vld (Congestion Valid), and determine the congestion level cg_class (Congestion Class) according to the interval in which the system cache is located. ), the range is 0 ⁇ N-1. If the system cache has not reached the minimum cache threshold configured by the user or the maximum cache threshold has been exceeded, the congestion valid indication eg_vld is set to 0, and the congestion valid indication and the congestion level are passed to the threshold offset module 203.
  • the step value calculation module 202 is configured to obtain a preset discarding threshold of the queue to which the new incoming message belongs. After the new packet arrives, the preset discarding threshold of the queue to be queued is searched according to the queue number of the packet. And calculate the step value of the queue to which the new incoming message belongs.
  • the step value calculation module 202 includes:
  • the first threshold obtaining module is configured to obtain, according to the queue number carried in the new incoming packet, a discarding threshold of the queue to which the new incoming message belongs;
  • the first step value calculation module is configured to calculate the step value by multiplying a preset percentage of the offset of the queue to which the new incoming message belongs and the preset discarding threshold.
  • the step value calculation module 202 includes:
  • the second threshold obtaining module is configured to obtain, according to the queue number carried in the new incoming packet, a preset maximum discarding threshold and a minimum discarding threshold of the queue to which the new incoming packet belongs;
  • a second step value calculation module configured to calculate a maximum step value for adjusting a current maximum discard threshold and a minimum step for adjusting a current minimum discard threshold in a single offset mode or an overall offset mode Value; the maximum step used to adjust the current maximum discard threshold in separate offset mode
  • the value is a product of a preset offset percentage and the preset maximum discard threshold.
  • the minimum step value for adjusting the current minimum discard threshold is a preset offset percentage and the preset minimum discard.
  • the product of the threshold is discarded; in the overall offset mode, the step values for adjusting the current maximum discard threshold and for adjusting the current minimum discard threshold are equal to a preset maximum discard threshold or a minimum discard threshold and The product of the offset percentage of the queue preset.
  • the threshold offset module 203 is configured to: after receiving the congestion effective indication and the congestion level sent by the congestion level determining module 201, search for a corresponding amplification factor, and calculate a gate according to the amplification factor and the step value sent by the step value calculation module 202. Limit the offset, add the preset discard threshold to the calculated threshold offset, and obtain the adjusted discard threshold. Then use the adjusted discard threshold to re-adjust the current use of the congestion avoidance algorithm. The threshold is discarded to achieve the purpose of adaptively adjusting the discard threshold.
  • the threshold offset module 203 searches for the corresponding amplification factor i_amply according to the congestion effective indication eg_val and the congestion level eg-class sent by the congestion level determination module 201. If the congestion valid indication eg_vld is 0, the amplification factor i_amply is directly set to zero. At the same time, the threshold offset of each discard threshold is calculated according to the step value step input by the step value calculation module 202. Finally, a new discard threshold is obtained, including a new TD discard threshold td th new, a new WRED discard threshold wred max th new and a new minimum drop threshold wred min th new 0
  • the judgment is still performed according to the normal flow. If the user selects the TD drop policy, the packet length is added to the queue depth of the queue to which it belongs. If the value is greater than the adjusted discard threshold td_th_new, the message is discarded; otherwise, the message is judged to be enqueued, and the system buffer and the queue depth of the queue (plus the packet length) are updated. If the user selects the WRED drop policy, the judgment method is the same as the original method, except that the adjusted discard thresholds wred max th new and wred min th ne w 0 are adopted.
  • Step 1 the new message enters, reads the current system level cache, according to the system level cache occupancy If the system level cache is less than the minimum cache threshold or greater than the maximum cache threshold, the congestion indication eg-vld is invalid, eg-class is set to 0; otherwise, the congestion indication eg-vld is valid.
  • Step 2 Detect whether the congestion indication eg-vld is valid. If valid, select different amplification factors i-amply according to the obtained congestion level. Otherwise, the amplification factor i-amply is 0, where different congestion levels correspond to different amplification factors. .
  • Step 3 Find the preset discarding threshold of the queue according to the queue number carried in the new incoming packet.
  • the step value step is calculated according to the user-configured offset percentage parameter and the offset mode.
  • the step value is a system preset reference discard threshold multiplied by an offset percentage
  • the congestion avoidance algorithm used includes a maximum discarding threshold and a minimum discarding threshold (for example, the WRED algorithm)
  • the two offset modes are included, respectively, an individual offset mode and an overall offset mode; in the separate offset mode, the The step value is divided into a maximum step value and a minimum step value, which are respectively used to adjust the current maximum discard threshold and the minimum discard threshold of the queue, where the maximum step value is equal to the system default maximum discard threshold multiplied by the bias. a percentage shift, the minimum step value being equal to a system preset minimum discard threshold multiplied by an offset percentage;
  • the step value is equal to the system default maximum discard threshold or the minimum discard threshold multiplied by the offset percentage, and the obtained step value is used to adjust the current maximum discard threshold and the minimum of the queue. Discard the threshold.
  • Step 5 Adjust the discarding threshold of the queue to which the new incoming packet belongs by using the threshold offset th-delta; if the WRED policy is selected, adjust the WRED maximum discarding threshold to wred_max_th_new, and the minimum discarding threshold is wred — min—th — new; If the TD policy is selected, the adjustment discard threshold is td th newo Step 6.
  • the congestion avoiding device discards the packet. If it is determined that the packet is enqueued (not discarded), the system cache is updated, and the current system cache is added with the length of the packet; otherwise, the packet is discarded, and the system cache is not updated.
  • Embodiment 1 Embodiment 1
  • the user configures two parameters, namely a maximum cache threshold Wth_h and a minimum cache threshold Wth-1.
  • the parameters Wth - h (window threshold high) and Wth - l (window threshold low) define a window used by the cache.
  • the maximum cache threshold Wth-h can be taken as 90% of the total cache.
  • the minimum cache threshold Wth-1 is greater than or equal to 0% of the total cache. This value is generally small to maintain a large cache space.
  • the system-level congestion level is divided into N levels, and the cache amount of each level is W/N, and the maintained cache window is divided into N intervals.
  • Each interval corresponds to a congestion level eg-class, and the higher the level eg-class, the deeper the congestion.
  • the step value calculation module 202 a percentage step calculation method is employed.
  • the step value step is a relative value indicating the percentage of the threshold increase for each level. The original difference between the queues can be guaranteed to some extent. Because the pre-set high-priority queue drop threshold is greater than the low-priority drop threshold, the step value of the high-priority queue is greater than the ste value of the drop threshold of the low-priority queue.
  • the WRED policy there is a minimum discard threshold wred min th and a maximum discard threshold wred_max-th.
  • the two preset discard thresholds are used as the base value, multiplied by the offset percentage step_per, and the following assumption is that step_per is x%.
  • the step value is unique, x%xtd-th (x supports user-matchable);
  • Figure 4 illustrates the step An implementation of the value calculation module, the maximum step value max_step corresponding to wred_max-th is x% X wred max th , and the minimum step value corresponding to wred min th is min x x X wred min th .
  • the threshold offset module 203 receives the congestion level eg-class and the congestion indication eg-vld obtains different amplification factors.
  • Figure 5 illustrates an implementation of the threshold offset module 203, eg-class and eg-vld as strobe signals for the selector, such that the selector MUX output and the corresponding amplification factor, the amplification factor i-amply ranges from [0, M-1] , ⁇ configured by the user. When i-amply is configured to 0, it is in fixed mode and the discard threshold remains unchanged.
  • Minimum discard threshold wred—min—th threshold offset is
  • Adding the discard thresholds to their respective th-deltas yields a new discard threshold.
  • the new WRED drop curve is shown in Figure 6. Since the pre-set wred_max-th is greater than wred min th , the offset of wred max th is an offset greater than wred min th. The separate offset makes the system cache more significant for the WRED drop algorithm, which is reflected in the discard curve for the discard slope to change, and the discard curve is smoother.
  • the threshold offset is
  • the adaptively adjusted TD discard threshold is not td-th + td-th-delta.
  • the first step and the second step are the same as in the first embodiment.
  • the system-level congestion level e-class is determined; then, different amplification factors i-amply are selected according to the obtained congestion level.
  • a percentage step calculation method is also employed. The difference is that in the overall offset mode, only one (wred_max_th or wred_min_th) is selected, so the base value is multiplied by the step value step, and the maximum and minimum discard threshold increase amounts are the same.
  • the TD discarding threshold td-th and the WRED maximum discarding threshold wred_max-th are selected. Used as a benchmark for the threshold shift of the two algorithms.
  • the td_step value is x%xtd-th; for the WRED mechanism, the step value is x%xwred_max tho
  • Threshold offset is
  • the adaptively adjusted WRED discard thresholds are wred—max—th+ wred—th—delta and wred—min—th+ wred—th—delta.
  • the discarding slope is constant, but the discarding curve is shifted to the right as a whole, which ensures the discarding characteristics of the WRED algorithm to some extent.
  • the threshold offset is
  • the adaptively adjusted TD discard threshold is not td-th + td-th-delta.
  • the congestion avoiding method and apparatus provided by the present invention adaptively adjust the discarding threshold according to the cache usage maintained in the system level.
  • the adjustable discard threshold includes, but is not limited to, the tail discard threshold td-th of the TD policy, the maximum discard threshold wred_max_th and the minimum discard threshold wred_min_th in the WRED policy, and other congestion avoidance algorithms may be used. stand by. If the system cache is sufficient, increase the discarding threshold and reduce the probability of packet discarding. This ensures that more packets are enqueued and the system cache is fully utilized. When the system cache is small, the discarding threshold is decreased. The probability of dropping large messages, minus Reduce the number of incoming messages, thus protecting the system cache from overflow.
  • the technical solution of the invention improves the performance of system congestion avoidance and improves the utilization of the entire system cache.
  • the device adopts a percentage offset method to maintain the quality of service between the queues.
  • the discarding threshold of the high-priority queue is increased by a larger amount than the low one, and the difference in QoS is clearly reflected.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

L'invention porte sur un procédé et un appareil d'évitement de congestion, qui sont utilisés pour résoudre le problème selon lequel : le mécanisme d'évitement de congestion existant ne peut pas exploiter entièrement la classe de congestion de système courante pour effectuer une détermination d'abandon de message, et ne peut pas éviter le risque de dépassement de capacité de cache et le faible taux d'utilisation de cache qui sont occasionnés par une congestion. L'invention ajuste un seuil d'abandon d'une manière adaptative en fonction de la condition d'utilisation du cache système, par augmentation du seuil d'abandon quand le cache système est adéquat et par réduction du seuil d'abandon quand le cache système est insuffisant. La solution technique de l'invention améliore les performances d'évitement de congestion de système, et augmente le taux d'utilisation du cache système entier.
PCT/CN2011/080189 2011-09-26 2011-09-26 Procédé et appareil d'évitement de congestion WO2013044438A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2011/080189 WO2013044438A1 (fr) 2011-09-26 2011-09-26 Procédé et appareil d'évitement de congestion

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2011/080189 WO2013044438A1 (fr) 2011-09-26 2011-09-26 Procédé et appareil d'évitement de congestion

Publications (1)

Publication Number Publication Date
WO2013044438A1 true WO2013044438A1 (fr) 2013-04-04

Family

ID=47994108

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2011/080189 WO2013044438A1 (fr) 2011-09-26 2011-09-26 Procédé et appareil d'évitement de congestion

Country Status (1)

Country Link
WO (1) WO2013044438A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019140757A1 (fr) * 2018-01-16 2019-07-25 厦门亿联网络技术股份有限公司 Procédé de commande de mémoire-tampon destiné à la transmission d'un flux de messages intersystème

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101388833A (zh) * 2008-07-18 2009-03-18 广东工业大学 基于自适应阈值机制的网络控制方法
CN101562566A (zh) * 2009-06-04 2009-10-21 吉林大学 基于实时路由器缓存占有率的主动队列管理方法
WO2010056158A1 (fr) * 2008-11-11 2010-05-20 Telefonaktiebolaget L M Ericsson (Publ) Procédé et dispositif pour permettre une indication de congestion dans un réseau de télécommunications

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101388833A (zh) * 2008-07-18 2009-03-18 广东工业大学 基于自适应阈值机制的网络控制方法
WO2010056158A1 (fr) * 2008-11-11 2010-05-20 Telefonaktiebolaget L M Ericsson (Publ) Procédé et dispositif pour permettre une indication de congestion dans un réseau de télécommunications
CN101562566A (zh) * 2009-06-04 2009-10-21 吉林大学 基于实时路由器缓存占有率的主动队列管理方法

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019140757A1 (fr) * 2018-01-16 2019-07-25 厦门亿联网络技术股份有限公司 Procédé de commande de mémoire-tampon destiné à la transmission d'un flux de messages intersystème

Similar Documents

Publication Publication Date Title
US10230645B1 (en) Dynamic adjustment of quality of service parameters
US8761012B2 (en) Packet relay apparatus and method of relaying packet
KR101451150B1 (ko) 헤드룸 감소
US7069356B2 (en) Method of controlling a queue buffer by performing congestion notification and automatically adapting a threshold value
US7675857B1 (en) Method and apparatus to avoid network congestion
US9444749B2 (en) Apparatus and method for selectively delaying network data flows
JP5575845B2 (ja) 分散通信システムのための協同的な自律的なスケジュールされたリソース割当
US7414973B2 (en) Communication traffic management systems and methods
US8509074B1 (en) System, method, and computer program product for controlling the rate of a network flow and groups of network flows
AU2014252266B2 (en) Voip bandwidth management
RU2450468C2 (ru) Система и способ адаптации к перегрузке сети
US20140347997A1 (en) Backward congestion notification
US20100098047A1 (en) Setting a data rate of encoded data of a transmitter
US8995268B2 (en) Method for controlling a flow in a packet switching network
KR20090130394A (ko) 네트워크 액세스 기술을 선택하기 위한 시스템 및 방법
WO2011162650A1 (fr) Procédé et configuration de détection d'encombrement dans un réseau de communications
WO2018218957A1 (fr) Procédé, dispositif et système de commande de trafic
WO2021083160A1 (fr) Procédé et appareil de transmission de données
WO2015024400A1 (fr) Procédé et appareil d'évitement d'encombrement de routeur
WO2014026554A1 (fr) Procédé et appareil de traitement de la congestion des paquets
Kamiyama et al. Simple and accurate identification of high-rate flows by packet sampling
JP4729413B2 (ja) パケット通信装置
US20050185581A1 (en) Active flow management with hysteresis
WO2013044438A1 (fr) Procédé et appareil d'évitement de congestion
Alemu et al. Dynamic configuration of RED parameters [random early detection]

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 11873178

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 11873178

Country of ref document: EP

Kind code of ref document: A1