CN113507420B - Congestion control method and device - Google Patents
Congestion control method and device Download PDFInfo
- Publication number
- CN113507420B CN113507420B CN202110898650.2A CN202110898650A CN113507420B CN 113507420 B CN113507420 B CN 113507420B CN 202110898650 A CN202110898650 A CN 202110898650A CN 113507420 B CN113507420 B CN 113507420B
- Authority
- CN
- China
- Prior art keywords
- network data
- data packet
- congestion control
- network
- congestion
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/12—Avoiding congestion; Recovering from congestion
- H04L47/127—Avoiding congestion; Recovering from congestion by using congestion prediction
Abstract
The invention provides a congestion control method and a congestion control device, wherein the method comprises the following steps: identifying each received network data packet to obtain marking bit information of each network data packet, wherein the marking bit information is determined according to the network data packet sending state in the next preset time period in the network flow corresponding to each network data packet; and generating a corresponding congestion control strategy according to the marking bit information and a preset congestion signal type. According to the invention, by predicting the congestion state between the switch and the end side, on the premise of not causing network packet loss and higher queues, the congestion control mechanism of the end side is improved, the accuracy of congestion condition judgment and the bandwidth utilization rate of the whole network are improved, and the performance indexes of saving network cost, optimizing throughput, time delay and the like of upper layer application are achieved.
Description
Technical Field
The present invention relates to the field of network communication technologies, and in particular, to a congestion control method and apparatus.
Background
Among various communication protocols with reliable Transmission, such as Transmission Control Protocol (TCP), congestion Control is an important basic design. In the existing public network environment, the classic designs of TCP NewReno, TCP Vegas, TCP Cubic, XCP and the like which have been born for about twenty years exist; in a data center network-oriented environment, there are also specialized designs such as DCTCP, DCQCN, HPCC, which have appeared in the last decade. These designs can be used in congestion control of the classical TCP protocol, in the roce (rdma over ethernet) protocol, or even in various new transport layer protocols in the future.
The basic ideas of the above designs are consistent, that is, the switch directly feeds back a Congestion signal to an end side (for example, a mobile device) according to a Congestion state of a current Network link, and the end side performs heuristic or precise adjustment according to the feedback signal, where the Congestion signal used In the existing Network Congestion control algorithm includes packet loss, time delay, display Congestion Notification (ECN), In-Network-telecommunications (INT), and the like. However, these designs only consider the current network link condition, and in modern networks, especially data center networks, because of frequent flow start and end, congestion control is performed by simply using the existing algorithm, so that the end side often judges the congestion condition by mistake, unnecessary rate reduction is generated, and waste of network bandwidth resources is caused.
Therefore, a method and an apparatus for congestion control are needed to solve the above problems.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a congestion control method and a congestion control device.
The invention provides a congestion control method, which comprises the following steps:
identifying each received network data packet to obtain marking bit information of each network data packet, wherein the marking bit information is determined according to the network data packet sending state in the next preset time period in the network flow corresponding to each network data packet;
and generating a corresponding congestion control strategy according to the marking bit information and a preset congestion signal type.
According to the congestion control method provided by the invention, the generating of the corresponding congestion control strategy according to the flag bit information and the preset congestion signal type comprises the following steps:
acquiring the mark type of each network data packet according to the mark bit information, wherein the mark type comprises a first mark type and a second mark type, and the first mark type indicates that the transmission of other network data packets exists in a network flow corresponding to the network data packet in the next preset time period; the second mark type indicates that no other network data packet is sent in the network flow corresponding to the network data packet in the next preset time period;
determining a congestion control algorithm according to the preset congestion signal type, if the congestion control algorithm is a congestion control algorithm based on a mark, acquiring the total byte number of the messages of the queue at the current moment according to the mark type and the byte number of each network data packet, and generating a target network data packet containing network congestion information according to the total byte number of the messages and a preset byte number threshold value;
and sending the target network data packet to a receiving end, and sending receiving confirmation information fed back by the receiving end to a sending end so that the sending end can carry out congestion control according to the receiving confirmation information.
According to the congestion control method provided by the present invention, the generating a corresponding congestion control policy according to the flag bit information and a preset congestion signal type further includes:
if the congestion control algorithm is based on network random packet loss, respectively placing network data packets with different mark types in corresponding queues to generate congestion control strategies according to the network data packets in the different queues;
if the congestion control algorithm is based on the time delay information, inserting the first marked type network data packet in the queue in front of the second marked type network data packet so as to generate a congestion control strategy according to the data of the network data packet in the new queue;
and sending the congestion control strategy to a receiving end, and sending receiving confirmation information fed back by the receiving end to a sending end so that the sending end can carry out congestion control according to the receiving confirmation information.
According to the congestion control method provided by the invention, the generating of the target network data packet containing the network congestion information according to the total byte number of the message and the preset byte number threshold value comprises the following steps:
and if the total byte number of the message is larger than a preset byte number threshold value, generating network congestion information, and writing the network congestion information into a network data packet to be sent to a receiving end at the current moment to obtain a target network data packet.
The invention also provides a congestion control method, which comprises the following steps:
generating marker bit information of a network data packet according to the network data packet sending state of a network flow to which the network data packet to be sent belongs in the next preset time period;
sending the network data packet and the marking bit information to a server side, so that the server side generates a corresponding congestion control strategy according to the marking bit information and a preset congestion signal type, and sends the congestion control strategy to a receiving end;
and acquiring the receiving confirmation information fed back by the receiving end, and carrying out congestion control according to the receiving confirmation information.
According to a congestion control method provided by the present invention, generating flag bit information of a network data packet according to a network data packet transmission state of a network flow to which the network data packet to be transmitted belongs in a next preset time period includes:
judging the network data packet sending state of the network flow to which the network data packet to be sent belongs in the next preset time period, and marking the marking bit information of a first marking type in the network data packet to be sent if the network flow to which the network data packet to be sent belongs has the sending of other network data packets in the next preset time period;
and if no other network data packet is sent in the network flow to which the network data packet to be sent belongs within the next preset time period, marking the marking bit information of the second marking type in the network data packet to be sent.
The present invention also provides a congestion control device, including:
the system comprises a mark information identification module, a mark information identification module and a data transmission module, wherein the mark information identification module is used for identifying each received network data packet to obtain mark bit information of each network data packet, and the mark bit information is determined according to the network data packet transmission state in the next preset time period in a network flow corresponding to each network data packet;
and the congestion control strategy generation module is used for generating a corresponding congestion control strategy according to the marking bit information and the preset congestion signal type.
The present invention also provides a congestion control device, including:
the system comprises a marking information generation module, a marking information generation module and a marking module, wherein the marking information generation module is used for generating marking bit information of a network data packet according to the network data packet transmission state of a network flow to which the network data packet to be transmitted belongs in the next preset time period;
the sending module is used for sending the network data packet and the marking bit information to a server end, so that the server end generates a corresponding congestion control strategy according to the marking bit information and a preset congestion signal type, and sends the congestion control strategy to a receiving end;
and the congestion control module is used for acquiring the receiving confirmation information fed back by the receiving end and carrying out congestion control according to the receiving confirmation information.
The present invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the steps of the congestion control method as described in any of the above when executing the program.
The invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the congestion control method as any one of the above.
According to the congestion control method and device provided by the invention, by predicting the congestion state between the switch and the end side, on the premise of not causing network packet loss and higher queues, the congestion control mechanism of the end side is improved, the accuracy of judging the congestion state and the bandwidth utilization rate of the whole network are improved, and the performance indexes such as network cost saving, throughput and time delay optimization of upper-layer application are achieved.
Drawings
In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed for the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a congestion control method according to the present invention;
fig. 2 is a second schematic flow chart of a congestion control method according to the present invention;
FIG. 3 is a schematic diagram of a network topology provided by the present invention;
fig. 4 is a schematic diagram of a switch receiving a network packet according to the present invention;
FIG. 5 is a schematic diagram of a switch sending network packets according to the present invention
Fig. 6 is a schematic structural diagram of a congestion control device according to the present invention;
fig. 7 is a second schematic structural diagram of a congestion control device according to the present invention;
fig. 8 is a schematic structural diagram of an electronic device provided in the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The existing network congestion control algorithm does not consider the assumption that each flow in a network link has unlimited data to be sent, and cannot fully utilize network bandwidth resources aiming at the characteristic that network flow in the existing network is started and stopped frequently, so that the following functional defects exist: firstly, judging the congestion state only according to the current network state, and if the state after the flow is not considered, carrying out misjudgment on the congestion state, namely, feeding back a congestion signal to the end sides by mistake, so that some end sides generate misspeed reduction judgment; the average throughput rate of the end-side is reduced due to the error speed reduction of the end-side, so that higher flow completion time is caused, and further, the performance of the throughput rate, the time delay and the like of the upper-layer service to which the flow belongs is reduced; thirdly, the reduction of the average throughput rate of the traffic can cause the shortage of the network bandwidth utilization rate, which wastes the network bandwidth resources on one hand, and on the other hand, when the same service requirement is met, more network devices and network links such as switches need to be deployed, which brings the increase of the fixed cost of the devices and more expenditure of the power consumption of the devices.
Based on the defects of the existing network congestion control algorithm, the invention provides a congestion control method, which is characterized in that network data packets sent in the last round of each flow are identified at the end side, and the network data packets are specially marked; then, the switch feeds back the statistical result to the end side (from the opposite end side to the switch) according to the information of the marked network data packet (i.e. the network data packet sent by the last round of each flow) and the unmarked network data packet (or the network data packet of the last round and the network data packet of the non-last round can be respectively marked into different types for distinguishing); the end side can more accurately judge whether the current network and the future network are in the congestion state according to the feedback, and can better control and adjust the congestion. The invention can be applied to a data center network and an enterprise network, provides stable and expandable network service for users, improves the utilization rate of network links, saves the network operation cost, reduces the flow completion time of upper-layer application, and improves the performance of the upper-layer application.
Fig. 1 is a schematic flow diagram of a congestion control method provided by the present invention, and as shown in fig. 1, the present invention provides a congestion control method, including:
In the present invention, a server-side switch is explained as an execution subject. The switch counts and summarizes the network data packet information flowing through the device, and carries out classification, identification and calculation according to whether the packet head of each network data packet contains extra marking bit information. In the present invention, the flag bit information of the network packet is written at the end side (for example, the mobile device as the sending end), specifically, before the switch receives the network packet, the network terminal device at the end side first determines whether a certain network flow is about to end or stops sending data. Wherein, the network flow is finished, that is, no network data packet needs to be sent to the network from the network card of the network terminal device in the network flow; the network flow stopping sending data means that the network flow will be at T0After a time, T, the network data is sent again0>RTT, RTT is Round-Trip Time (Round Trip Time), tableThe total time from the beginning of sending data by the sending end to the time when the sending end receives the confirmation information from the receiving end is shown.
Furthermore, the exchanger identifies the received network data packet, and the invention distinguishes the network data packet by two kinds of marked information. Specifically, the time when any network data packet leaves the network card (i.e. is sent to the switch by the sending end) is t0The network flow where the network data packet is located is the network flow f, if at [ t0+RTT，t0+RTT+T]In the time period, if no network data packet is sent any more by the network flow f, identifying that the network data packet is last-RTT pkt and marking the network data packet as a first mark type; otherwise, the network data packet is normal pkt, wherein T is an adjustable parameter, and the invention takes T as RTT.
And 102, generating a corresponding congestion control strategy according to the flag bit information and a preset congestion signal type.
In the invention, a network switch reads the marking bit information in a network data packet sent by network equipment (the marking bit information is added to the packet head of the network data packet), respectively carries out statistical calculation aiming at normal pkt and last-RTT pkt, and judges the current congestion state and the future congestion state of the network equipment by combining the current queue length.
Specifically, the present invention assumes that the RTTs of all network flows flowing through the current link are equal, and are all T1(ii) a Each network flow maintains its congestion window unchanged and no new flows are added. At [0, T1]During the time period, the network data packet leaving from the link has m last-RTT pkts and n normal pkts, and is in [ T ]1,2T1]During the time period, the network packet entering the link has only n pkts at most (the n pkts may all be normal pkts, and there may be a part of last-RTT pkt). Because no new network data packet can re-enter the network after last-RTT pkt leaves the network, if the link has a queue at 0 moment, there are m network data packets in the queue, if the switch does not mark the network data packet with a congestion signalTo feed back the downgrade to the end side, but to let each flow keep its congestion window, at 2T1The queue of the switch will naturally drop to 0 and just let the port keep the full speed send state. In the invention, the core of the classification calculation of the network switch is to reduce the interference of last-RTT pkt on the congestion judgment, namely, when judging whether congestion occurs, the core of the invention is to separate and calculate normal pkt and last-RTT pkt, aiming at different preset congestion signal types, different congestion control strategies are adopted, for example, the congestion signal for displaying the congestion notification is to carry out congestion control according to the total byte number of a message; the congestion signal based on the time delay adjusts the positions of the network data packets with different mark types in the queue to realize congestion control.
According to the congestion control method provided by the invention, by predicting the congestion state between the switch and the end side, on the premise of not causing network packet loss and higher queues, the congestion control mechanism of the end side is improved, the accuracy of judging the congestion state and the bandwidth utilization rate of the whole network are improved, and the performance indexes such as network cost saving, throughput and time delay optimization of upper-layer application are achieved.
On the basis of the above embodiment, the generating a corresponding congestion control policy according to the flag bit information and a preset congestion signal type includes:
acquiring the mark type of each network data packet according to the mark bit information, wherein the mark type comprises a first mark type and a second mark type, and the first mark type indicates that the transmission of other network data packets exists in a network flow corresponding to the network data packet in the next preset time period; the second mark type indicates that no other network data packet is sent in the network flow corresponding to the network data packet in the next preset time period;
determining a congestion control algorithm according to the preset congestion signal type, if the congestion control algorithm is a congestion control algorithm based on a mark, acquiring the total byte number of the messages of the queue at the current moment according to the mark type and the byte number of each network data packet, and generating a target network data packet containing network congestion information according to the total byte number of the messages and a preset byte number threshold value;
and sending the target network data packet to a receiving end, and sending receiving confirmation information fed back by the receiving end to a sending end so that the sending end can carry out congestion control according to the receiving confirmation information.
In the invention, as for a congestion control algorithm based on marking, an ECN marking algorithm is taken as an example, and the existing ECN marking algorithm is to judge a threshold value according to the number of messages in a queue or the total byte number Q of the messages. According to the preset values of high and low waterlines K1 and K2(K1 < K2), if Q < K1, the network is considered not to be congested; if Q > K2, the network is considered congested; and if K1 is less than Q and less than K2, randomly judging whether the network is congested. The invention does not need to modify the judgment formula of the ECN marking algorithm, and only needs to re-measure the total byte number Q of the message. In the invention, the total byte number formula of the message is as follows:
Q＝∑normal pkts+λ∑lastRTT pkts；
wherein 0 < λ < 1. For last-RTT pkt, the invention specially underestimates the message number/byte number, so that the nature of natural speed reduction brought by last-RTT pkt can be taken into account. The parameter lambda is a preset parameter which can be adjusted by network operation and maintenance personnel to control the additional speed reduction of the network, and the larger lambda represents the existence of the additional speed reduction which can be used for dealing with the newly started flow in the network. Due to the existence of extra speed reduction, the utilization rate of a network link is often not in a 100% state, so that the reserved partial bandwidth can ensure that even if newly started traffic exists, a queue on a switch cannot be suddenly increased too high; and a smaller lambda represents that network operation and maintenance personnel are more concerned about the network bandwidth utilization rate, and the sudden size of the network queue is regarded as a secondary factor. Other similar marking algorithms can be used in schemes such as HPCC and XCP which use extra headers for accurate congestion control. Taking HPCC as an example, when the switch inserts the INT header, the number/byte number of normal pkt and last-RTT pkt may be respectively measured and fed back to the end side, and then the end side performs rate adjustment. The essential idea is that the size of last-RTT pkt is "underestimated" to reduce the number of times of "misjudging congestion occurs". And finally, calculating the total byte number of the message at the current moment, if the congestion condition is judged to occur, writing the congestion information into the network data packet to be sent to the receiving end by the switch at the moment so as to generate a target network data packet to be sent to the receiving end, reading the congestion information in the target network data packet after the receiving end interprets the target data packet, and feeding back confirmation information to the switch, and sending the receiving confirmation information fed back by the receiving end to the sending end by the switch so that the sending end can carry out congestion control according to the receiving confirmation information.
On the basis of the above embodiment, the generating a target network data packet including network congestion information according to the total byte number of the packet and a preset byte number threshold includes:
and if the total byte number of the message is larger than a preset byte number threshold value, generating network congestion information, and writing the network congestion information into a network data packet to be sent to a receiving end at the current moment to obtain a target network data packet.
In the invention, if a congestion signal such as ECN or INT which needs the switch to actively modify the header information for feedback is selected, after the congestion state which can occur subsequently is judged through the embodiment, the corresponding feedback information is modified according to the definition of ECN/INT on congestion. Taking ECN as an example, if the network packet currently leaving the queue is determined to be congested, the ECN flag is modified to 11.
On the basis of the above embodiment, the generating a corresponding congestion control policy according to the flag bit information and a preset congestion signal type further includes:
if the congestion control algorithm is based on network random packet loss, respectively placing network data packets with different mark types in corresponding queues to generate congestion control strategies according to the network data packets in the different queues;
if the congestion control algorithm is based on the time delay information, inserting the first marked type network data packet in the queue in front of the second marked type network data packet so as to generate a congestion control strategy according to the data of the network data packet in the new queue;
and sending the congestion control strategy to a receiving end, and sending receiving confirmation information fed back by the receiving end to a sending end so that the sending end can carry out congestion control according to the receiving confirmation information.
In the invention, taking TCP Cubic, a congestion control algorithm based on network random packet loss as an example, a network switch can respectively place last-RTT pkt and normal pkt in different queues, so that packet loss judgment of normal pkt is not interfered by last-RTT pkt through TCP Cubic, and the problem of misjudgment can not occur.
Further, taking an algorithm such as Swift or TCP Vegas, which performs congestion determination based on delay information as an example, if there is a queue on the switch, normal pkt in the queue can be always inserted in front of last-RTT pkt, so as to reduce the occurrence of misdetermination.
Fig. 2 is a second schematic flow chart of the congestion control method provided by the present invention, and as shown in fig. 2, the present invention provides a congestion control method, which includes:
In the present invention, a transmitting end is explained as an execution subject. The end of the network flow occurs, usually because the upper layer service has completed the service, no corresponding data interaction is needed any more; the reason that the network stops sending data is that the opposite end needs to process data for a long time, and the sending end and the opposite end do not need data information interaction for a period of time. Therefore, there are two main methods for the network terminal device (sending end) to determine the end/stop of the network flow to send data: one is that when the upper end service of the end side issues the data to be sent, the upper end service explicitly notifies the transport layer of the network protocol stack; the other is a transport layer protocol stack at the end side, which judges that the network flow is about to end according to the comparison between the size of the data to be sent and the size of the congestion window of the current network.
Further, tongAnd the network card or the operating system at the end side adds additional marking information to the packet head of the network data packet. In the invention, the time for any network data packet to leave the network card (namely, to be sent to the switch by the sending end) is t0The network flow where the network data packet is located is the network flow f, if at [ t0+RTT，t0+RTT+T]In the time period, if no network data packet is sent any more by the network flow f, identifying that the network data packet is last-RTT pkt and marking the network data packet as a first mark type; otherwise, the network data packet is normal pkt, wherein T is an adjustable parameter, and the invention takes T as RTT. The time for the network data packet to leave the network card is t0When t is0And when the current time is, the congestion judgment belongs to the prediction of the future condition of the network flow.
Specifically, the generating the flag bit information of the network data packet according to the network data packet transmission state of the network flow to which the network data packet to be transmitted belongs in the next preset time period includes:
judging the network data packet sending state of the network flow to which the network data packet to be sent belongs in the next preset time period, and marking the marking bit information of a first marking type in the network data packet to be sent if the network flow to which the network data packet to be sent belongs has the sending of other network data packets in the next preset time period;
and if no other network data packet is sent in the network flow to which the network data packet to be sent belongs within the next preset time period, marking the marking bit information of the second marking type in the network data packet to be sent.
In the invention, the information is additionally marked on the packet head of the network data packet in order to distinguish whether the network data packet belongs to last-RTT pkt or normal pkt. The marking information can be added by an extra header customized by network personnel, and can also be directly multiplexed at certain positions of the existing network data packet header, wherein the multiplexing positions comprise but are not limited to an ECN flag bit of an IP packet header, optional options of the IP packet header and a Reserved bit of a BTH packet header, the ECN flag bit is 2 bits, 01 and 10 are not distinguished in the original RFC design of ECN, the ECN bit of last-RTT pkt can be marked as 01, and the ECN bit of normal pkt can be marked as 10, so that the purpose of being compatible with the congestion control scheme based on ECN is achieved. The invention divides the network data packets into two types, and can distinguish the two types of network data packets only by 1bit information. It should be noted that 2 bits or even more bits can be used for the distinguishing bits due to design considerations such as compatibility with other schemes.
In the present invention, the end side (sender) adjusts the rate of the corresponding network stream according to the feedback from the receiver. Specifically, after receiving a network data packet sent from the end side, the opposite end extracts congestion information from the network data packet and actively feeds back the congestion information to the end side in the form of ACK; and after the end side receives the feedback, the sending rate of the regulator is regulated according to a preset congestion window regulation formula. In the present invention, the end side can use the AIMD method of the classical DCTCP: that is, each time feedback with congestion flags is received, the congestion window is decreased by β, and if not, the congestion window is increased by α/cw (cw is the congestion window), where α and β are available parameters, or other Cubic constant rate adjustment formulas may be used.
According to the congestion control method provided by the invention, by predicting the congestion state between the switch and the end side, on the premise of not causing network packet loss and higher queues, the congestion control mechanism of the end side is improved, the accuracy of judging the congestion state and the bandwidth utilization rate of the whole network are improved, and the performance indexes such as network cost saving, throughput and time delay optimization of upper-layer application are achieved.
In an embodiment, a general description is given of a congestion control method provided by the present invention. Fig. 3 is a schematic diagram of a network topology provided by the present invention, and as shown in fig. 3, the network topology includes 3 terminals, i.e., H1, H2, and H3, and 1 forwarding device S1 (switch), where there are multiple network flows between H1 and H3, and there are multiple network flows between H2 and H3. If both H1 and H2 transmit data at full rate, congestion will result at S1 (the total rate of transmission of H1 and H2 will exceed the link rate between S1 and H3). The invention adopts RoCEv2 protocol communication, and the scheme takes DCTCP protocol as a basic example and can be applied to cloud data centers and enterprise networks.
Fig. 4 is a schematic diagram of a switch provided by the present invention receiving a network packet, fig. 5 is a schematic diagram of a switch provided by the present invention sending a network packet, and referring to fig. 3, fig. 4 and fig. 5, when each network flow on H1 sends a network traffic packet, each network flow knows the total byte number B to be sent due to the RoCEv2 protocol characteristicstNumber of bytes currently finished being sent B1Number of bytes sent to the network that have not been acknowledged B2And the congestion window cw for that network flow. If B is present1+B2+cw>BtThen sending end H1The network packet sent out is marked as the "last round".
For example, if the network flow needs to send 500KB of data in total, the sender now knows that the peer has received 300KB, sends out but does not know whether the peer has received 100KB of data, and the congestion window is 50 KB. When the network flow is to continue sending network traffic packets, the network traffic packets will be considered normal network traffic packets, not marked as "last round", since 300KB +100KB +50KB < 500 KB; if the sender already knows that the peer received 370KB, at which time 370KB +100KB +50KB > 520KB, the network packet is marked as "last round".
Further, when the switch S1 receives each network packet, it first determines which queue the network packet should be put into according to the forwarding rule. For example, switch S1 received pkt1, found that it should be placed in queue Q1, and if pkt1 is a packet marked as "last round", the count of Q1.C1(C1 is last-RTT pkt) would be incremented by 1; if pkt1 is not marked as "last round", the count of Q1.C2(C2 is normal pkt) is increased by 1.
In the present invention, the switch S1 polls each queue according to the queue management policy, and when the Q1 is polled, the switch S1 takes out the pkt2, calculates X ═ λ C1+ C2, and then determines whether the congestion state is present. If λ is set to 0.5, C1 is 20, C2 is 30, the two pipelines for setting ECN are both 45, and X is 0.5 × 20+30 is 40 < 45, which means that when 20 packets in the "last round" are buffered in the switch, 30 packets in the normal round "are buffered, and the standard of congestion is 45, because more network packets in the switch belong to the network packets in the" last round ", the queue will decrease quickly, it can be considered that there is no congestion in the future, and the network packet pkt2 is not marked; if C2 is now 40, and X is calculated to be 0.5 × 20+40 ═ 50>45, it is considered that the queue is already too high now, and there is a risk of too high a future queue, and if pkt2 is a normal network packet, it will be ECN marked to be fed back to the end side for speed reduction. However, if pkt2 is the "last round" network packet, it will not be marked (note that, no matter it is marked, the end-side will not send any more data until the feedback signal reaches the end-side).
Further, when H3 receives the network packet, it copies the ECN flag in the network packet and feeds back the ECN flag to H1 via an ACK network packet.
Finally, after receiving the fed back ACK packet, H1 adjusts the congestion window according to the fed back ECN flag, specifically, if the ECN flag is 11, the congestion window is decreased by 0.5 KB; if the ECN is marked as 10 or 01, its congestion window is increased by (1 KB. times. pkt-size)/cw.
In the invention, when a plurality of flows exist between H1 and H3, and between H2 and H3, the running network flow can timely detect that other network flows are about to finish sending, and quickly preempt the residual bandwidth.
In another embodiment, the present invention is compared to the HPCC technique in a ns3 simulation environment. Specifically, a 1:1 layer 3 Fat-Tree network is formed by 160 servers, all link bandwidths are 100Gbps, and link time delay is 1 us. Loads used in the environment randomly generate flow according to a flow size distribution curve of the loads, under the condition that the average utilization rate of a link is 50%, the median (50% delay) and tail distribution (99% tail delay) of flow completion time of the flows with different sizes are respectively counted, and the flow completion time of smaller flows (namely the flows with the sizes of S, M and L) is not greatly increased, which means that a lower switch queue is maintained; the flow completion time of a larger flow (i.e. the flow size is XL) is reduced by more than 25%, which means that the present invention utilizes the network bandwidth resources more fully, and the specific data can refer to table 1:
TABLE 1
The invention provides a network congestion control method, which relates to the judgment of 'network flow is about to end or will stop sending' by an end side, extra marking information is added to the packet header of a network data packet through an end side network card or an operating system, a network switch counts and summarizes the network packet information flowing through the equipment, and then classification calculation is carried out according to whether each packet header contains the extra marking information, so that the network switch writes network congestion information into the network data packet, and the end side adjusts the speed of the corresponding network flow according to feedback. The invention ensures the low queue of the switch, improves the utilization rate of the network link, saves the network resource and reduces the network cost; for the network flow (large data sending quantity) of a larger network, the completion time can be obviously reduced, and various congestion signals can be compatible. The invention only needs to modify the existing congestion control algorithm a little, and the realization cost is low.
Fig. 6 is a schematic structural diagram of a congestion control apparatus provided by the present invention, and as shown in fig. 6, the present invention provides a congestion control apparatus, including a tag information identifying module 601 and a congestion control policy generating module 602, where the tag information identifying module 601 is configured to identify each received network data packet, and obtain tag bit information of each network data packet, where the tag bit information is determined according to a network data packet sending state in a next preset time period in a network flow corresponding to each network data packet; the congestion control policy generating module 602 is configured to generate a corresponding congestion control policy according to the flag bit information and a preset congestion signal type.
The congestion control device provided by the invention can improve the congestion control mechanism at the end side by predicting the congestion state between the switch and the end side on the premise of not causing network packet loss and higher queues, improve the accuracy of judging the congestion state and the bandwidth utilization rate of the whole network, and achieve the performance indexes of saving the network cost, optimizing the throughput, time delay and the like of upper-layer application.
Fig. 7 is a second schematic structural diagram of the congestion control apparatus provided by the present invention, and as shown in fig. 7, the present invention provides a congestion control apparatus, including a tag information generating module 701, a sending module 702, and a congestion control module 703, where the tag information generating module 701 is configured to generate tag bit information of a network data packet according to a network data packet sending state of a network flow to which the network data packet to be sent belongs in a next preset time period; the sending module 702 is configured to send the network data packet and the flag bit information to a server, so that the server generates a corresponding congestion control policy according to the flag bit information and a preset congestion signal type, and sends the congestion control policy to a receiving end; the congestion control module 703 is configured to obtain the reception confirmation information fed back by the receiving end, and perform congestion control according to the reception confirmation information.
The congestion control device provided by the invention can improve the congestion control mechanism at the end side by predicting the congestion state between the switch and the end side on the premise of not causing network packet loss and higher queues, improve the accuracy of judging the congestion state and the bandwidth utilization rate of the whole network, and achieve the performance indexes of saving the network cost, optimizing the throughput, time delay and the like of upper-layer application.
The apparatus provided by the present invention is used for executing the above method embodiments, and for details and flow, reference is made to the above embodiments, which are not described herein again.
Fig. 8 is a schematic structural diagram of an electronic device provided in the present invention, and as shown in fig. 8, the electronic device may include: a processor (processor)801, a communication interface (communication interface)802, a memory (memory)803 and a communication bus 804, wherein the processor 801, the communication interface 802 and the memory 803 complete communication with each other through the communication bus 804. The processor 801 may invoke logic instructions in the memory 803 to perform a congestion control method comprising: identifying each received network data packet to obtain marking bit information of each network data packet, wherein the marking bit information is determined according to the network data packet sending state in the next preset time period in the network flow corresponding to each network data packet; generating a corresponding congestion control strategy according to the flag bit information and a preset congestion signal type;
or generating the marker bit information of the network data packet according to the network data packet sending state of the network flow to which the network data packet to be sent belongs in the next preset time period; sending the network data packet and the marking bit information to a server side, so that the server side generates a corresponding congestion control strategy according to the marking bit information and a preset congestion signal type, and sends the congestion control strategy to a receiving end; and acquiring the receiving confirmation information fed back by the receiving end, and carrying out congestion control according to the receiving confirmation information.
In addition, the logic instructions in the memory 803 may be implemented in the form of software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-only memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
In another aspect, the present invention also provides a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, enable the computer to perform the congestion control method provided by the above methods, the method comprising: identifying each received network data packet to obtain marking bit information of each network data packet, wherein the marking bit information is determined according to the network data packet sending state in the next preset time period in the network flow corresponding to each network data packet; generating a corresponding congestion control strategy according to the flag bit information and a preset congestion signal type;
or generating the marker bit information of the network data packet according to the network data packet sending state of the network flow to which the network data packet to be sent belongs in the next preset time period; sending the network data packet and the marking bit information to a server side, so that the server side generates a corresponding congestion control strategy according to the marking bit information and a preset congestion signal type, and sends the congestion control strategy to a receiving end; and acquiring the receiving confirmation information fed back by the receiving end, and carrying out congestion control according to the receiving confirmation information.
In yet another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored, the computer program being implemented by a processor to perform the congestion control method provided in the above embodiments, the method including: identifying each received network data packet to obtain marking bit information of each network data packet, wherein the marking bit information is determined according to the network data packet sending state in the next preset time period in the network flow corresponding to each network data packet; generating a corresponding congestion control strategy according to the flag bit information and a preset congestion signal type;
or generating the marker bit information of the network data packet according to the network data packet sending state of the network flow to which the network data packet to be sent belongs in the next preset time period; sending the network data packet and the marking bit information to a server side, so that the server side generates a corresponding congestion control strategy according to the marking bit information and a preset congestion signal type, and sends the congestion control strategy to a receiving end; and acquiring the receiving confirmation information fed back by the receiving end, and carrying out congestion control according to the receiving confirmation information.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (6)
1. A congestion control method, comprising:
identifying each received network data packet to obtain marking bit information of each network data packet, wherein the marking bit information is determined according to the network data packet sending state in the next preset time period in the network flow corresponding to each network data packet;
generating a corresponding congestion control strategy according to the flag bit information and a preset congestion signal type;
generating a corresponding congestion control strategy according to the flag bit information and a preset congestion signal type, wherein the congestion control strategy comprises the following steps:
acquiring the mark type of each network data packet according to the mark bit information, wherein the mark type comprises a first mark type and a second mark type, and the first mark type indicates that the transmission of other network data packets exists in a network flow corresponding to the network data packet in the next preset time period; the second mark type indicates that no other network data packet is sent in the network flow corresponding to the network data packet in the next preset time period;
determining a congestion control algorithm according to the preset congestion signal type, if the congestion control algorithm is a congestion control algorithm based on a mark, acquiring the total byte number of the messages of the queue at the current moment according to the mark type and the byte number of each network data packet, and generating a target network data packet containing network congestion information according to the total byte number of the messages and a preset byte number threshold value;
sending the target network data packet to a receiving end, and sending receiving confirmation information fed back by the receiving end to a sending end so that the sending end can carry out congestion control according to the receiving confirmation information;
generating a target network data packet containing network congestion information according to the total byte number of the message and a preset byte number threshold value, wherein the method comprises the following steps:
if the total byte number of the message is larger than a preset byte number threshold value, generating network congestion information, and writing the network congestion information into a network data packet to be sent to a receiving end at the current moment to obtain a target network data packet;
generating a corresponding congestion control strategy according to the flag bit information and a preset congestion signal type, further comprising:
if the congestion control algorithm is based on network random packet loss, respectively placing network data packets with different mark types in corresponding queues to generate congestion control strategies according to the network data packets in the different queues;
if the congestion control algorithm is based on the time delay information, inserting the first marked type network data packet in the queue in front of the second marked type network data packet so as to generate a congestion control strategy according to the data of the network data packet in the new queue;
and sending the congestion control strategy to a receiving end, and sending receiving confirmation information fed back by the receiving end to a sending end so that the sending end can carry out congestion control according to the receiving confirmation information.
2. A congestion control method, comprising:
generating marker bit information of a network data packet according to the network data packet sending state of a network flow to which the network data packet to be sent belongs in the next preset time period;
sending the network data packet and the marking bit information to a server side, so that the server side generates a corresponding congestion control strategy according to the marking bit information and a preset congestion signal type, and sends the congestion control strategy to a receiving end;
acquiring receiving confirmation information fed back by the receiving end, and carrying out congestion control according to the receiving confirmation information;
the server side generates a corresponding congestion control strategy according to the marker bit information and a preset congestion signal type, and sends the congestion control strategy to a receiving end, and the method comprises the following steps:
if the congestion control algorithm is based on network random packet loss, respectively placing network data packets with different mark types in corresponding queues to generate congestion control strategies according to the network data packets in the different queues;
if the congestion control algorithm is based on the time delay information, inserting the first marked type network data packet in the queue in front of the second marked type network data packet so as to generate a congestion control strategy according to the data of the network data packet in the new queue;
sending the congestion control strategy to a receiving end, and sending receiving confirmation information fed back by the receiving end to a sending end so that the sending end can carry out congestion control according to the receiving confirmation information;
the generating the flag bit information of the network data packet according to the network data packet transmission state of the network flow to which the network data packet to be transmitted belongs in the next preset time period includes:
judging the network data packet sending state of the network flow to which the network data packet to be sent belongs in the next preset time period, and marking the marking bit information of a first marking type in the network data packet to be sent if the network flow to which the network data packet to be sent belongs has the sending of other network data packets in the next preset time period;
and if no other network data packet is sent in the network flow to which the network data packet to be sent belongs within the next preset time period, marking the marking bit information of the second marking type in the network data packet to be sent.
3. A congestion control apparatus, comprising:
the system comprises a mark information identification module, a mark information identification module and a data transmission module, wherein the mark information identification module is used for identifying each received network data packet to obtain mark bit information of each network data packet, and the mark bit information is determined according to the network data packet transmission state in the next preset time period in a network flow corresponding to each network data packet;
the congestion control strategy generating module is used for generating a corresponding congestion control strategy according to the marking bit information and a preset congestion signal type;
the congestion control policy generation module is specifically configured to:
acquiring the mark type of each network data packet according to the mark bit information, wherein the mark type comprises a first mark type and a second mark type, and the first mark type indicates that the transmission of other network data packets exists in a network flow corresponding to the network data packet in the next preset time period; the second mark type indicates that no other network data packet is sent in the network flow corresponding to the network data packet in the next preset time period;
determining a congestion control algorithm according to the preset congestion signal type, if the congestion control algorithm is a congestion control algorithm based on a mark, acquiring the total byte number of the messages of the queue at the current moment according to the mark type and the byte number of each network data packet, and generating a target network data packet containing network congestion information according to the total byte number of the messages and a preset byte number threshold value;
sending the target network data packet to a receiving end, and sending receiving confirmation information fed back by the receiving end to a sending end so that the sending end can carry out congestion control according to the receiving confirmation information;
generating a target network data packet containing network congestion information according to the total byte number of the message and a preset byte number threshold value, wherein the method comprises the following steps:
if the total byte number of the message is larger than a preset byte number threshold value, generating network congestion information, and writing the network congestion information into a network data packet to be sent to a receiving end at the current moment to obtain a target network data packet;
the congestion control policy generation module is further configured to:
if the congestion control algorithm is based on network random packet loss, respectively placing network data packets with different mark types in corresponding queues to generate congestion control strategies according to the network data packets in the different queues;
if the congestion control algorithm is based on the time delay information, inserting the first marked type network data packet in the queue in front of the second marked type network data packet so as to generate a congestion control strategy according to the data of the network data packet in the new queue;
and sending the congestion control strategy to a receiving end, and sending receiving confirmation information fed back by the receiving end to a sending end so that the sending end can carry out congestion control according to the receiving confirmation information.
4. A congestion control apparatus, comprising:
the system comprises a marking information generation module, a marking information generation module and a marking module, wherein the marking information generation module is used for generating marking bit information of a network data packet according to the network data packet transmission state of a network flow to which the network data packet to be transmitted belongs in the next preset time period;
the sending module is used for sending the network data packet and the marking bit information to a server end, so that the server end generates a corresponding congestion control strategy according to the marking bit information and a preset congestion signal type, and sends the congestion control strategy to a receiving end;
the congestion control module is used for acquiring the receiving confirmation information fed back by the receiving end and carrying out congestion control according to the receiving confirmation information;
the server side generates a corresponding congestion control strategy according to the marker bit information and a preset congestion signal type, and sends the congestion control strategy to a receiving end, and the method comprises the following steps:
if the congestion control algorithm is based on network random packet loss, respectively placing network data packets with different mark types in corresponding queues to generate congestion control strategies according to the network data packets in the different queues;
if the congestion control algorithm is based on the time delay information, inserting the first marked type network data packet in the queue in front of the second marked type network data packet so as to generate a congestion control strategy according to the data of the network data packet in the new queue;
sending the congestion control strategy to a receiving end, and sending receiving confirmation information fed back by the receiving end to a sending end so that the sending end can carry out congestion control according to the receiving confirmation information;
the tag information generation module is specifically configured to:
judging the network data packet sending state of the network flow to which the network data packet to be sent belongs in the next preset time period, and marking the marking bit information of a first marking type in the network data packet to be sent if the network flow to which the network data packet to be sent belongs has the sending of other network data packets in the next preset time period;
and if no other network data packet is sent in the network flow to which the network data packet to be sent belongs within the next preset time period, marking the marking bit information of the second marking type in the network data packet to be sent.
5. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the congestion control method according to any of claims 1 to 2 when executing the computer program.
6. A non-transitory computer readable storage medium, having stored thereon a computer program, wherein the computer program, when being executed by a processor, is adapted to carry out the steps of the congestion control method according to any of the claims 1 to 2.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110898650.2A CN113507420B (en) | 2021-08-05 | 2021-08-05 | Congestion control method and device |
| PCT/CN2022/106690 WO2023011179A1 (en) | 2021-08-05 | 2022-07-20 | Congestion control method and apparatus |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110898650.2A CN113507420B (en) | 2021-08-05 | 2021-08-05 | Congestion control method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113507420A CN113507420A (en) | 2021-10-15 |
| CN113507420B true CN113507420B (en) | 2022-03-15 |
Family
ID=78015819
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110898650.2A Active CN113507420B (en) | 2021-08-05 | 2021-08-05 | Congestion control method and device |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN113507420B (en) |
| WO (1) | WO2023011179A1 (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113507420B (en) * | 2021-08-05 | 2022-03-15 | 清华大学 | Congestion control method and device |
| CN114760252A (en) * | 2022-03-24 | 2022-07-15 | 北京邮电大学 | Data center network congestion control method and system |
| CN115119068B (en) * | 2022-06-21 | 2023-07-18 | 广州市奥威亚电子科技有限公司 | Network congestion processing method and system |
| CN114938354A (en) * | 2022-06-24 | 2022-08-23 | 北京有竹居网络技术有限公司 | Congestion control method, device, equipment and storage medium |
| CN115118663B (en) * | 2022-06-27 | 2023-11-07 | 西安电子科技大学 | Method for obtaining network congestion information by combining in-band network telemetry |
| CN115865827B (en) * | 2023-02-09 | 2023-05-23 | 合肥综合性国家科学中心人工智能研究院(安徽省人工智能实验室) | Event-driven cross-layer end network cooperative congestion control method |
| CN116886622B (en) * | 2023-09-08 | 2023-11-24 | 合肥综合性国家科学中心人工智能研究院(安徽省人工智能实验室) | Network congestion control method, device, equipment and storage medium |
| CN116915706B (en) * | 2023-09-13 | 2023-12-26 | 合肥综合性国家科学中心人工智能研究院(安徽省人工智能实验室) | Data center network congestion control method, device, equipment and storage medium |
| CN117395206B (en) * | 2023-12-12 | 2024-03-12 | 南京邮电大学 | Rapid and accurate congestion feedback method for lossless data center network |
| CN117544567B (en) * | 2024-01-09 | 2024-03-19 | 南京邮电大学 | Memory transfer integrated RDMA data center congestion control method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106533970A (en) * | 2016-11-02 | 2017-03-22 | 重庆大学 | Differential flow control method and device for cloud computing data center network |
| CN112468405A (en) * | 2020-11-30 | 2021-03-09 | 中国人民解放军国防科技大学 | Data center network congestion control method based on credit and reaction type |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102882803B (en) * | 2012-10-11 | 2015-12-16 | 四川大学 | A kind of mixing jamming control method based on packet loss and time delay |
| EP2869514A1 (en) * | 2013-10-30 | 2015-05-06 | Alcatel Lucent | Method and system for queue management in a packet-switched network |
| CN106330761B (en) * | 2015-06-30 | 2020-09-15 | 中兴通讯股份有限公司 | Congestion control method and device based on queue time delay |
| CN105357138B (en) * | 2015-09-30 | 2018-06-05 | 四川大学 | A kind of adaptive method for reliable transmission based on low priority |
| CN106160953B (en) * | 2016-07-06 | 2019-04-16 | 四川大学 | A kind of transmission method based on learning-oriented energy efficiency model |
| CN108259383B (en) * | 2016-12-29 | 2021-10-01 | 北京华为数字技术有限公司 | Data transmission method and network equipment |
| US10931587B2 (en) * | 2017-12-08 | 2021-02-23 | Reniac, Inc. | Systems and methods for congestion control in a network |
| CN109067665B (en) * | 2018-09-25 | 2022-01-11 | 华为技术有限公司 | Congestion control method and network equipment |
| CN113141313A (en) * | 2020-01-19 | 2021-07-20 | 华为技术有限公司 | Congestion control method, device and system and storage medium |
| CN113162862A (en) * | 2020-01-23 | 2021-07-23 | 华为技术有限公司 | Congestion control method and device |
| CN113507420B (en) * | 2021-08-05 | 2022-03-15 | 清华大学 | Congestion control method and device |
-
2021
- 2021-08-05 CN CN202110898650.2A patent/CN113507420B/en active Active
-
2022
- 2022-07-20 WO PCT/CN2022/106690 patent/WO2023011179A1/en unknown
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106533970A (en) * | 2016-11-02 | 2017-03-22 | 重庆大学 | Differential flow control method and device for cloud computing data center network |
| CN112468405A (en) * | 2020-11-30 | 2021-03-09 | 中国人民解放军国防科技大学 | Data center network congestion control method based on credit and reaction type |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2023011179A1 (en) | 2023-02-09 |
| CN113507420A (en) | 2021-10-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113507420B (en) | Congestion control method and device | |
| US6026075A (en) | Flow control mechanism | |
| EP2890065B1 (en) | Systems and methods for managing congestion | |
| US6934256B1 (en) | Method of detecting non-responsive network flows | |
| US11677665B2 (en) | Systems and methods for identifying candidate flows in data packet networks | |
| US9325628B2 (en) | Packet handling method, forwarding device and system | |
| CN110943933B (en) | Method, device and system for realizing data transmission | |
| CN114938350B (en) | Congestion feedback-based data stream transmission control method in lossless network of data center | |
| CN109120544A (en) | The transfer control method of Intrusion Detection based on host end flow scheduling in a kind of data center network | |
| CN116527584B (en) | Congestion control method and device, communication network and computer storage medium | |
| US7027395B2 (en) | Method for marking packets of a data transmission flow and marker device performing this method | |
| CN104113884B (en) | Real-time multimedia transfer rate control mechanism in a kind of wireless network | |
| EP3267639B1 (en) | Congestion control within a communication network | |
| CN111386683A (en) | System and method for accelerating or decelerating a data transport network protocol based on real-time transport network congestion conditions | |
| CN105450441A (en) | TCP-protocol-based weak network acceleration method and apparatus | |
| CN106789709B (en) | Load balancing method and device | |
| CN116055416A (en) | Method and device for adjusting transmission rate in long-distance communication network scene | |
| US11622028B2 (en) | Explicit notification of operative conditions along a network path | |
| CN108322402B (en) | Message processing method, device and system | |
| CN114828081A (en) | Cooperative hybrid congestion control method based on path recovery | |
| CN115733799A (en) | Network congestion control method and related device | |
| CA2429416A1 (en) | Apparatus and method for traffic control for data transmissions in a tcp/ip data transmission network | |
| CN116527585B (en) | Congestion control method for stream length perception | |
| CN116260773A (en) | Congestion control method and related equipment | |
| Usman et al. | A transport layer approach to improve energy efficiency |
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 |