CN114401230A - Sending rate control method and device based on cross-data center network communication - Google Patents
Sending rate control method and device based on cross-data center network communication Download PDFInfo
- Publication number
- CN114401230A CN114401230A CN202111617446.5A CN202111617446A CN114401230A CN 114401230 A CN114401230 A CN 114401230A CN 202111617446 A CN202111617446 A CN 202111617446A CN 114401230 A CN114401230 A CN 114401230A
- Authority
- CN
- China
- Prior art keywords
- data center
- link
- rate
- bandwidth utilization
- sending
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/10—Flow control; Congestion control
- H04L47/25—Flow control; Congestion control with rate being modified by the source upon detecting a change of network conditions
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention relates to the technical field of network communication, in particular to a sending rate control method and a sending rate control device based on cross-data center network communication. The invention firstly calculates the bandwidth utilization rate of each data center link under the historical sending rate, then calculates the change information of each bandwidth utilization rate in a period of time, and finally obtains the position of the bottleneck of the whole link network according to the change information and the bandwidth utilization rate, namely, judges whether the bandwidth utilization rate of the data center link is close to the bottleneck bandwidth utilization rate of the data center link or the bandwidth utilization rate of the wide area network link is close to the bottleneck bandwidth utilization rate of the wide area network link, and then the sending rate can be adjusted in a targeted manner so as to prevent the whole link network from being congested. The invention can adjust the sending rate before congestion occurs to avoid congestion, and simultaneously, the invention can also maximize the bandwidth utilization rate of the whole link network by adjusting the sending rate while avoiding congestion.
Description
Technical Field
The invention relates to the technical field of network communication, in particular to a sending rate control method and a sending rate control device based on cross-data center network communication.
Background
The cross-data center network communication is that two or more data center links communicate through a wide area network link, and both the bandwidth utilization rate of the data center link and the bandwidth utilization rate of the wide area network link can cause network congestion.
In order to effectively utilize network bandwidth resources, reduce network congestion and improve user experience, a large number of congestion control algorithms are proposed to effectively control the sending rate, and the sending rate is controlled within a proper range, so that the congestion degree of the network is reduced while the whole network communication efficiency is improved. In order to achieve the above purpose, the congestion degree of the network can be obtained through two kinds of feedback information by counting packet loss information and delay information in the network. The prior art also adjusts the sending rate by calculating the bandwidth utilization of the data center links. However, in the prior art, whether the sending rate is adjusted according to the post feedback information or the bandwidth utilization rate of the data center link, it is difficult to effectively prevent the network congestion phenomenon.
In summary, it is difficult for the prior art to effectively prevent the occurrence of congestion in a communication network across a data center network.
Thus, there is a need for improvements and enhancements in the art.
Disclosure of Invention
In order to solve the technical problems, the invention provides a sending rate control method and a sending rate control device based on cross-data center network communication, and solves the problem that the prior art is difficult to effectively prevent the occurrence of the cross-data center network communication network congestion phenomenon.
In order to achieve the purpose, the invention adopts the following technical scheme:
in a first aspect, the present invention provides a sending rate control method based on cross-data center network communication, including:
calculating the bandwidth utilization rate of each data center link;
calculating utilization rate change information corresponding to the bandwidth utilization rate of each data center link, wherein the utilization rate change information is used for representing the increase amplitude or the decrease amplitude of the bandwidth utilization rate;
obtaining link positions corresponding to bottleneck bandwidth utilization rates of a whole link network according to the bandwidth utilization rates of the data center links and utilization rate change information of the data center links, wherein the whole link network comprises the data center links serving as a sending end, the data center links serving as a receiving end and a wide area network link, and the wide area network link is used for enabling the data center links to realize cross communication;
and controlling the sending rate of the data center link as a sending end according to the link position corresponding to the bottleneck bandwidth utilization rate.
In an implementation manner, the obtaining, according to the bandwidth utilization of each data center link and the utilization change information of each data center link, a link position corresponding to a bottleneck bandwidth utilization of an entire link network, where the entire link network includes the data center link serving as a sending end, the data center link serving as a receiving end, and a wide area network link, and the wide area network link is used to enable cross communication between the data center links, includes:
obtaining an increase rate corresponding to the bandwidth utilization rate in the utilization rate change information according to the utilization rate change information;
and when the bandwidth utilization rate of each data center link is smaller than a set value and the growth rate corresponding to the bandwidth utilization rate is smaller than a preset value, obtaining that the link position corresponding to the bottleneck bandwidth utilization rate of the whole link network is positioned on the wide area network link.
In an implementation manner, the controlling, according to a link position corresponding to the bottleneck bandwidth utilization, a sending rate of the data center link as a sending end includes:
obtaining the link position corresponding to the bottleneck bandwidth utilization rate of the whole link network on the wide area network link according to the link position corresponding to the bottleneck bandwidth utilization rate;
acquiring historical sending speed v1 and historical sending speed v2 of the data center link as a sender, wherein the historical sending speed v1 is different from the historical sending speed v 2;
calculating a forwarding rate V1 aiming at a historical sending rate V1 in each data center link, wherein the forwarding rate is the data volume forwarded by the data center link in unit time;
calculating a forwarding rate V2 aiming at the historical sending rate V2 in each data center link;
and controlling the sending rate of the data center link as a sender at the next moment according to the historical sending rate V1, the historical sending rate V2, the forwarding rate V1 and the forwarding rate V2.
In one implementation, the controlling the sending rate of the data center link as the sender at the next time according to the historical sending rate V1, the historical sending rate V2, the forwarding rate V1 and the forwarding rate V2 includes:
obtaining a maximum forwarding rate V1max in the forwarding rates V1 according to the forwarding rate V1;
obtaining a maximum forwarding rate V2max in the forwarding rates V2 according to the forwarding rate V2;
and controlling the sending rate of the data center link as a sending end at the next moment according to the maximum forwarding rate V1max, the historical sending rate V1, the maximum forwarding rate V2max and the historical sending rate V2.
In an implementation manner, the obtaining, according to the bandwidth utilization of each data center link and the utilization change information of each data center link, a link position corresponding to a bottleneck bandwidth utilization of an entire link network, where the entire link network includes the data center link serving as a sending end, the data center link serving as a receiving end, and a wide area network link, and the wide area network link is used to enable cross communication between the data center links, includes:
and when the bandwidth utilization rate of each data center link is greater than a set value, obtaining that the link position corresponding to the bottleneck bandwidth utilization rate of the whole link network is positioned on the data center link.
In an implementation manner, the controlling, according to a link position corresponding to the bottleneck bandwidth utilization, a sending rate of the data center link as a sending end includes:
obtaining the link position corresponding to the bottleneck bandwidth utilization rate of the whole link network on the data center link according to the link position corresponding to the bottleneck bandwidth utilization rate;
acquiring bottleneck bandwidth utilization rate of each data center link;
acquiring a historical sending rate v' of the data center link as a sending end;
calculating the bandwidth utilization rate f corresponding to each data center link and the historical sending rate v';
and controlling the sending rate of the data center link as a sending end at the next moment according to the bottleneck bandwidth utilization rate, the bandwidth utilization rate f and the historical sending rate v' of each data center link.
In one implementation, the controlling a sending rate of the data center link as a sending end at a next time according to the bottleneck bandwidth utilization rate, the bandwidth utilization rate f, and the historical sending rate v' of each data center link includes:
calculating a difference value between the bottleneck bandwidth utilization rate and the bandwidth utilization rate f according to the bottleneck bandwidth utilization rate and the bandwidth utilization rate f;
and controlling the sending rate of the data center link as the sending end at the next moment according to the difference and the historical sending rate v'.
In one implementation, the calculating the bandwidth utilization of each data center link includes:
calculating a forwarding rate corresponding to each data center link, wherein the forwarding rate is a data volume forwarded by the data center link in unit time;
calculating a queue utilization rate corresponding to each data center link, wherein the queue is located in the data center link, and the queue utilization rate is used for representing the occupied degree of the queue;
and obtaining the bandwidth utilization rate of each data center link according to the forwarding rate and the queue utilization rate of each data center link.
In one implementation, the obtaining the bandwidth utilization of each data center link according to the forwarding rate and the queue utilization of each data center link includes:
and adding the forwarding rate and the queue utilization rate to obtain the bandwidth utilization rate of each data center link.
In one implementation, the calculating a forwarding rate corresponding to each of the data center links, where the forwarding rate is a data amount forwarded by the data center link in a unit time, includes:
acquiring timestamp information and switch data processing capacity corresponding to the timestamp information through a confirmation packet, wherein the confirmation packet is confirmation information sent to a data center link of a receiving end after the data center link serving as the receiving end receives data sent by the data center link serving as a sending end, and the switch is located in the data center link;
and obtaining the forwarding rate corresponding to each data center link according to the timestamp information and the data processing amount.
In one implementation, the calculating a queue utilization corresponding to each of the data center links, where the queue is located in the data center link, and the queue utilization is used to characterize the degree of occupation of the queue, and includes:
acquiring the queuing length of the queue through a confirmation packet, wherein the confirmation packet is confirmation information sent to the data center link of a receiving end after the data center link of the receiving end receives data sent by the data center link of a sending end, and the switch is positioned in the data center link;
obtaining the round trip delay corresponding to the whole link network;
acquiring a bandwidth corresponding to the switch;
calculating the bandwidth times the round trip delay to obtain a product result;
and dividing the queuing length by the product result to obtain the queue utilization rate corresponding to each data center link.
In a second aspect, an embodiment of the present invention further provides an apparatus for a sending rate control method based on cross-data center network communication, where the apparatus includes the following components:
the bandwidth utilization rate calculation module is used for calculating the bandwidth utilization rate of each data center link;
a bandwidth utilization rate change information calculation module, configured to calculate utilization rate change information corresponding to the bandwidth utilization rate of each data center link, where the utilization rate change information is used to represent an increase or decrease of the bandwidth utilization rate;
a bottleneck position calculating module, configured to obtain a link position corresponding to a bottleneck bandwidth utilization ratio of an entire link network according to a bandwidth utilization ratio of each data center link and utilization ratio change information of each data center link, where the entire link network includes the data center link serving as a sending end, the data center link serving as a receiving end, and a wide area network link, and the wide area network link is used to implement cross communication between the data center links;
and the sending rate calculation module is used for controlling the sending rate of the data center link as a sending end according to the link position corresponding to the bottleneck bandwidth utilization rate.
In a third aspect, an embodiment of the present invention further provides a terminal device, where the terminal device includes a memory, a processor, and a sending rate control program based on cross-data center network communication, where the sending rate control program based on cross-data center network communication is stored in the memory and is executable on the processor, and when the processor executes the sending rate control program based on cross-data center network communication, the step of implementing the sending rate control method based on cross-data center network communication is implemented.
In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, where a sending rate control program based on cross-data center network communication is stored on the computer-readable storage medium, and when the sending rate control program based on cross-data center network communication is executed by a processor, the steps of the sending rate control method based on cross-data center network communication described above are implemented.
Has the advantages that: the invention firstly calculates the bandwidth utilization rate of each data center link under the historical sending rate, then calculates the change information of each bandwidth utilization rate in a period of time, and finally obtains the position of the bottleneck of the whole link network according to the change information and the bandwidth utilization rate, namely, the position of the bottleneck of the bandwidth utilization rate of the whole link network can be obtained by judging whether the bandwidth utilization rate of the data center link is close to the bottleneck bandwidth utilization rate of the data center link or the bandwidth utilization rate of the wide area network link is close to the bottleneck bandwidth utilization rate of the wide area network link, and the sending rate can be adjusted in a targeted manner only to prevent the whole link network from being congested. The invention can adjust the sending rate before congestion occurs to avoid congestion, and simultaneously, the invention can also maximize the bandwidth utilization rate of the whole link network by adjusting the sending rate while avoiding congestion.
Drawings
FIG. 1 is an overall flow chart of the present invention;
FIG. 2 is a cross-data center communication topology diagram in an embodiment;
FIG. 3 is a flow chart in an embodiment;
fig. 4 is a schematic block diagram of an internal structure of a terminal device according to an embodiment of the present invention.
Detailed Description
The technical scheme of the invention is clearly and completely described below by combining the embodiment and the attached drawings of the specification. 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.
Research shows that the cross-data center network communication is that two or more data center links communicate through a wide area network link, and the network congestion phenomenon can be caused by the bandwidth utilization rate of the data center link and the bandwidth utilization rate of the wide area network link. In order to effectively utilize network bandwidth resources, reduce network congestion and improve user experience, a large number of congestion control algorithms are proposed to effectively control the sending rate, and the sending rate is controlled within a proper range, so that the congestion degree of the network is reduced while the whole network communication efficiency is improved. In order to achieve the above purpose, the congestion degree of the network can be obtained through two kinds of feedback information by counting packet loss information and delay information in the network. The prior art also adjusts the sending rate by calculating the bandwidth utilization of the data center links. However, in the prior art, whether the sending rate is adjusted according to the post feedback information or the bandwidth utilization rate of the data center link, it is difficult to effectively prevent the network congestion phenomenon.
In order to solve the technical problems, the invention provides a sending rate control method and a sending rate control device based on cross-data center network communication, and solves the problem that the prior art is difficult to effectively prevent the occurrence of the cross-data center network communication network congestion phenomenon. In specific implementation, the position of the bottleneck of the whole link network is obtained according to the bandwidth utilization rate and the change information of the bandwidth utilization rate, and then the sending rate is adjusted to avoid the occurrence of congestion.
For example, two data center links and a wan link in fig. 2 form the whole link network, where one data center link serves as a transmitting end, the other data center link serves as a receiving end, and the transmitting end and the receiving end perform data transmission through the wan link. And calculating the bandwidth utilization rate of the two data center links at a first sending rate, and the bandwidth utilization rate of the two data center links at a second sending rate, wherein the first sending rate is different from the second sending rate, but the bandwidth utilization rates corresponding to the two sending rates do not change greatly, so that the bottleneck can be judged to be positioned on the wide area network link, otherwise, the bottleneck is positioned on one of the data center links. Knowing on which link the bottleneck is located (which link is the determining factor for whether congestion will occur), the rate of occurrence can be adjusted according to the bandwidth utilization corresponding to the link to avoid congestion.
Exemplary method
The sending rate control method based on cross-data center network communication of the embodiment can be applied to terminal equipment, and the terminal equipment can be a terminal product with a control function. In this embodiment, as shown in fig. 1, the sending rate control method based on cross-data center network communication specifically includes the following steps:
and S100, calculating the bandwidth utilization rate of each data center link.
Step S100 includes two parts: and calculating the queue utilization rate and the forwarding rate of each data center link, wherein the queue utilization rate and the forwarding rate are the bandwidth utilization rate. Step S100 specifically includes the following steps S101, S102, S103, S104, S105, S106, S107, and S108:
s101, time stamp information and switch data processing amount corresponding to the time stamp information are obtained through a confirmation packet, the confirmation packet is confirmation information sent to a data center link of a receiving end after the data center link serving as the receiving end receives data sent by the data center link serving as a sending end, and the switch is located in the data center link.
S102, obtaining the queuing length of the queue through a confirmation packet, wherein the confirmation packet is confirmation information sent to the data center link of the receiving end after the data center link serving as the receiving end receives data sent by the data center link serving as the sending end, and the switch is located in the data center link.
And S103, acquiring the round trip delay corresponding to the whole link network.
The timestamp information (dequeued timestamp), switch data throughput, and queue length in steps S101-S103 constitute the intra-network data for each data center link in fig. 2. In this embodiment, the intra-network data is obtained by a confirmation packet, each P4 switch in fig. 2 first stores the timestamp information, the data processing amount, and the queue length of the current time locally (obtains intra-network data), and when the confirmation packet arrives at each P4 switch, each P4 switch writes the above three data in the header of the confirmation packet (adds intra-network data to the header). Therefore, steps S101 to S103 include acquiring intra-network data and adding intra-network data to the packet header, and the specific processes of these two parts are described below:
acquiring data in the network: queue length and timestamp information can be obtained directly from the P4 switch encapsulated data. And the amount of data handled by the switch needs to be obtained using a data structure of registers in the switch.
The structure of the register in the P4 switch can count the data packets processed in a certain direction of the switch, and the principle is that when a data packet passes through, the register will automatically accumulate the count. Since our scheme obtains the data in the network by modifying the header of the acknowledgment packet, the amount of data (the number of processed data packets) processed by the switch returning from the receiving end to the transmitting end downlink is obtained by counting in the downlink using only one register. In fact, the present embodiment is expected to count the amount of data (the number of processed packets) sent from the sending end to the receiving end in the uplink, because the amount of data processed by the uplink switch can reflect the current utilization of network resources. The required result is indeed obtained if only one register is used to count the uplink, but this value cannot be passed on to the downlink acknowledgment packet, since the register held by the uplink is not triggered in the downlink and the value of the register is not available when the acknowledgment packet passes through, which is also not feasible. Therefore, in this embodiment, two registers are used, one is to count the amount of data processed by the uplink and downlink, and the other is to count the amount of data processed by the downlink switch, and first, both registers are triggered in the downlink where the acknowledgement packet passes through, so as to ensure that the register values are available. We then subtract the values of these two registers to get the amount of data we wish to have uplink processed.
As can be seen from the above description, in this embodiment, two registers are provided in each P4 switch, where one register is used to count the amount of data in the direction (uplink) in which the sender sends data to the receiver; the other register is used for counting the data amount when the confirmation packet arrives at the register in the process of moving from the receiving end to the transmitting end (downlink). The subtraction of the two statistical data amounts is the data amount processed by the uplink.
Adding in-network data to the header: when a packet including a confirmation packet arrives at the switch, it is first necessary to determine whether the packet is a communication packet or a confirmation packet, and the determination can be made based on the sender IP and the receiver IP of the packet. If the data packet is judged to be normal, the switch only needs to forward the packet to a specific port, and if the data packet is confirmed, the switch also needs a TCP option of a packet header. When an acknowledgement packet arrives at the switch, the header is first parsed, including the IP layer, TCP layer, etc. After the header is parsed, three in-network elements (timestamp information of the switch, data throughput, queue length) are sequentially added to the newly defined option field of TCP during dequeuing, and a new TCP option identifier is set to 253. When the switch recognizes that the TCP header has the new option defined, the switch writes the in-network field into the TCP header. After the addition is completed, the TCP header field is modified, so that the check code put by the TCP needs to be modified, and the acknowledgement packet is forwarded after the update is completed.
S104, obtaining the forwarding rate corresponding to each data center link according to the timestamp information and the data processing amount:
wherein ack.l [ i ]. timestamp-L [ i ]. timestamp represents the difference between the timestamps of two arriving ack packet options, ack.l [ i ]. pkg _ count-L [ i ]. pkg represents the difference between the data volumes (packets) of the two ack packet options, pkg _ size is the number of bytes of the packet, txRate is the forwarding rate, and the forwarding rate divided by the bandwidth ack.l [ i ]. B of the switch is the forwarding rate of the embodiment.
And S105, acquiring the bandwidth corresponding to the switch.
And S106, calculating the bandwidth multiplied by the round trip delay to obtain a product result.
And S107, dividing the queuing length by the product result to obtain the queue utilization rate corresponding to each data center link.
Steps S105 to S107 are to calculate the queue utilization, and the formula for calculating the queue utilization is as follows:
wherein, ack.l [ i ] qlen is the queue length (queuing length, i.e. how many packets are queued in the queue) in the current acknowledgment packet option, L [ i ] qlen is the queue length of the link recorded at the last time, the minimum value of the two queue lengths is taken to reduce the influence of network jitter as much as possible, and baseRTT is the latest round trip delay.
In this embodiment, the base RTT of the entire link network always changes because the base RTT is communicated across the data center network, and the RTT needs to be involved in calculating the bandwidth utilization. Therefore, in order to obtain more accurate bandwidth utilization, the embodiment designs a link base RTT detection module. Base RTT refers to the end-to-end delay when the link has no queuing delay. Since the embodiment can acquire the data in the network in the data center, the queue length is included. So that the queue length information can be used. When the queue length information extracted from the acknowledgement packet is close to 0, a threshold value low _ qlen is given, and the RTT at this time is the base RTT. Therefore, it is not necessary to detect the base RTT by sending a packet, and accurate base RTT information can be acquired.
And S108, adding the forwarding rate and the queue utilization rate to obtain the bandwidth utilization rate of each data center link.
The timestamp information (dequeue timestamp), the switch data processing amount, and the queuing length contained in the in-network data may be directly written into the header of the acknowledgment packet, but since the option space reserved by TCP is 40 bytes, at most, only the in-network data of three P4 switches can be written, and in order to write more in-network data of P4 switches, the in-network data of the switches is written into the header of the acknowledgment packet in the following manner in this embodiment:
when network data is written into a plurality of switches, three data (time stamp information, switch data processing amount and queuing length) are not directly written, partial results are obtained through local preliminary calculation by utilizing the calculation capacity of the P4 switch, and then the results are written into a packet header. Specifically, the utilization rate of the link is calculated according to the queue length, the timestamp and the forwarded data amount, and the utilization rate comprises the utilization rate of the queue and the utilization rate of forwarding. The utilization rate of the queue needs the information of the queue length, the bandwidth information of the node and the base RTT of the link to be obtained together. The forwarding utilization rate needs the timestamp and the forwarded data volume information of the two acknowledgement packets, the forwarding rate can be obtained by dividing the difference value of the forwarded data volumes of the two acknowledgement packets by the difference value of the timestamp, and the forwarding utilization rate can be obtained by dividing the forwarding rate by the bandwidth of the forwarding node. It can be seen that the utilization of the queue needs to be obtained by relying on base RTT obtained by end-side probing, and therefore cannot be directly calculated in the P4 switch.
But the forwarding utilization can be calculated directly in the switch. The time stamp and the processed data volume when the confirmation packet arrives are maintained locally at the switch, when the next packet arrives, the information of the two packets can be directly utilized to obtain the forwarding rate, and the bandwidth of the node is a constant and known, so the forwarding utilization rate can be obtained by dividing the forwarding rate by the bandwidth of the node. Note, however, that division is not supported in P4 switches, and therefore, forwarding utilization cannot be directly obtained by division. However, the P4 switch supports shifting, and shifting left by one corresponds to dividing by 2, so it can be realized by shifting. Therefore, in deployment, when the forwarding rate is calculated, the forwarding rate is calculated not according to the interval of every two acknowledgement packets, but the data amount processed in a period of time is counted at intervals of time t, so that the data amount is divided by t and then divided by the node bandwidth. Since both t and node bandwidth are constants, to divide the amount of data processed by this constant, the amount of data processed may be divided by the nth power of 2 and then the nth power of 2 by this constant, n being a given constant. The division of the processed data amount by 2 can be directly realized by displacement, and the nth power of 2 divided by a constant is a constant value and can be kept locally. In this way, the utilization of forwarding can be calculated locally.
After the forwarding utilization rate is obtained, at each P4 switch node, the data to be written in the packet header includes the queue length and the locally calculated forwarding utilization rate, so that only two variables, which are 8 bytes, need to be written in, and therefore 40 reserved TCP option bytes can be written in the data of five nodes, which can meet the multi-node communication requirements in many scenarios.
As can be seen from the above description, in this embodiment, the forwarding rate and the queue utilization rate corresponding to the switch are first calculated by using the computing capability of the switch itself through three intra-network data including the timestamp information (dequeuing timestamp), the data processing amount of the switch, and the queuing length, and then two data, that is, the forwarding rate and the queue utilization rate, are written into the packet header of the acknowledgment packet, so that 40 reserved TCP option bytes can be written into the forwarding rates and the queue utilization rates of five switches.
And S200, calculating utilization rate change information corresponding to the bandwidth utilization rate of each data center link, wherein the utilization rate change information is used for representing the increase amplitude or the decrease amplitude of the bandwidth utilization rate.
In this embodiment, the utilization rate change information corresponding to the bandwidth utilization rate includes an increase rate or a decrease rate of the bandwidth utilization rate.
And S300, obtaining a link position corresponding to a bottleneck bandwidth utilization rate of the whole link network according to the bandwidth utilization rate of each data center link and the utilization rate change information of each data center link, wherein the whole link network comprises the data center link serving as a sending end, the data center link serving as a receiving end and a wide area network link, and the wide area network link is used for realizing cross communication among the data center links.
The bottleneck in bandwidth utilization may be located on the switches at the data center links, or may be located on the switches at the wide area network links. The bottleneck of the bandwidth utilization ratio is that a certain sending rate sends data to occupy bandwidth, the larger the sending rate is, the larger the bandwidth utilization ratio is, and when the bandwidth utilization ratio is close to the maximum bandwidth utilization ratio (namely, the bottleneck is a fixed value), the sending rate is increased, and the whole link network is congested. The embodiment adopts the following method to obtain the bottleneck position of the bandwidth utilization rate:
and when the bandwidth utilization rate of each data center link is smaller than a set value and the growth rate corresponding to the bandwidth utilization rate is smaller than a preset value, the bottleneck bandwidth utilization rate of the whole link network is positioned on the wide area network link.
And when the bandwidth utilization rate of each data center link is greater than a set value, the bottleneck bandwidth utilization rate of the whole link network is positioned on the data center link.
The above-mentioned link where the bottleneck is obtained is based on the following principle:
if data packets continue to be transmitted throughout the link network, the bandwidth utilization may increase over a period of time if the current bandwidth utilization is low. Thus, if the calculated bandwidth utilization in the data center is still low and the growth rate is low for a certain period of time, the bottleneck bandwidth utilization of the wan link is presumably higher than the bottleneck bandwidth utilization of the data center link (i.e., the wan link limits the transmission of data packets throughout the link network). Conversely, if the bandwidth utilization within the data center is greater than 1, the bottleneck point of the entire link should be within the data center link rather than the wide area network link (i.e., the data center link limits the transmission of data packets throughout the link network).
The specific meaning of the above principle is: and when the calculated bandwidth utilization rate in the data center link is lower than low _ u, the data center link is considered to have low bandwidth utilization rate, and the data center link with the bandwidth utilization rate higher than high _ u is considered to be overloaded. A queue, the size of which is update _ size, is used to store u and txRate, etc. related history data. When the bandwidth utilization rate increase rate of the uplink _ size continuous acknowledgement packets is smaller than low _ u _ growth, the bandwidth utilization rate increase rate is considered to be too low, that is, a data packet is blocked on the wan link.
In this embodiment, it is determined whether the bottleneck point is located on the wan link according to the bandwidth utilization rate and the increase rate of the data center link, rather than directly determining whether the bottleneck point is located on the wan link according to the bandwidth utilization rate on the wan link, because the wan link cannot deploy the P4 switch, the in-network data of the wan link cannot be acquired.
And S400, controlling the sending rate of the data center link as a sending end according to the link position corresponding to the bottleneck bandwidth utilization rate.
When the bottleneck bandwidth utilization of the whole link network obtained according to step S300 is located on the wan link, the step S400 of controlling the transmission rate includes the following steps:
s401, according to the link position corresponding to the bottleneck bandwidth utilization rate, obtaining that the link position corresponding to the bottleneck bandwidth utilization rate is located on the wide area network link.
S402, acquiring historical sending speed v1 and historical sending speed v2 of the data center link as a sender, wherein the historical sending speed v1 is different from the historical sending speed v 2.
And S403, calculating a forwarding rate V1 aiming at the historical sending rate V1 in each data center link, wherein the forwarding rate is the data volume forwarded by the data center link in unit time.
When the data packet is sent by using the historical sending speed V1, a forwarding speed V1 of the switch corresponding to the historical sending speed V1 is obtained.
S404, calculating a forwarding rate V2 aiming at the historical sending rate V2 in each data center link.
When the data packet is sent by using the historical sending speed V2, a forwarding speed V2 of the switch corresponding to the historical sending speed V1 is obtained. The method of calculating the forwarding rate in steps S403 and S404 is the same as the method of calculating the forwarding rate in step S104.
S405, obtaining the maximum forwarding rate V1max in the forwarding rates V1 according to the forwarding rate V1.
S406, obtaining the maximum forwarding rate V2max in the forwarding rates V2 according to the forwarding rate V2.
S407, controlling the sending rate of the data center link as the sending end at the next moment according to the maximum forwarding rate V1max, the historical sending rate V1, the maximum forwarding rate V2max and the historical sending rate V2.
For example, when the maximum forwarding rate V1max corresponding to the historical sending rate V1 is close to 1, the sending rate at the next time should be adjusted to be less than V1 to avoid congestion. The same is true for v 2.
When V1 is smaller than V2, V1max is smaller than V2max, and neither V1max nor V2max is close to 1, the sending rate at the next moment can be continuously increased on the premise that congestion can be avoided, so that the transmission rate of the whole link network is increased.
When the bottleneck bandwidth utilization of the whole link network is located on the wan link, the present embodiment adjusts the sending rate based on the following principle:
when the utilization rate of the bottleneck point of the wide area network is greater than that of the data center, the effective forwarding rate of the data center link is the bottleneck point of the whole link network, so that the bottleneck bandwidth at the wide area network link can be updated by using the current forwarding rate of the data center link.
When the bottleneck bandwidth utilization rate of the whole link network obtained according to the step S300 is located on the data center link, the step S400 of controlling the sending rate includes the following steps S408, S409, S4010, S4011, S4012, and S4013:
s408, according to the link position corresponding to the bottleneck bandwidth utilization rate, obtaining that the link position corresponding to the bottleneck bandwidth utilization rate is positioned on the data center link.
As shown in step S300, when the bandwidth utilization is greater than 1, the bottleneck point of the data of the entire link is located on the data center link.
And S409, acquiring the bottleneck bandwidth utilization rate of each data center link.
S4010, obtaining a historical sending rate v' of the data center link as a sending end.
S4011, calculating bandwidth utilization rate f corresponding to each data center link and the historical sending rate v'.
S4012, calculating a difference between the bottleneck bandwidth utilization rate and the bandwidth utilization rate f according to the bottleneck bandwidth utilization rate and the bandwidth utilization rate f.
S4013, controlling the sending rate of the data center link as the sending end at the next moment according to the difference value and the historical sending rate v'.
In this embodiment, the method for calculating the bandwidth utilization ratio f is the same as the method for calculating the bandwidth utilization ratio in step S108. When the difference is small, it indicates that sending the data packet at the sending rate v 'will make the bandwidth utilization close to the bottleneck bandwidth utilization, and if sending the data packet at a sending rate larger than the sending rate v' again will cause the whole link network to be congested, at this moment, the sending rate should be reduced to avoid the congestion. When the difference is large, it indicates that sending the data packet at the sending rate v 'does not cause the bandwidth utilization to approach the bottleneck bandwidth utilization, so the data packet can be sent at a sending rate greater than the sending rate v' to achieve full utilization of the bandwidth and increase the transmission speed of the entire link network.
Taking fig. 3 as an example, the overall process of controlling the sending rate according to the present invention is described:
in this embodiment, an algorithm for avoiding congestion by adjusting the sending rate needs coordination between the end-side server protocol stack and the network-side switch, and when an acknowledgment packet reaches the P4 switch, the P4 switch needs to parse the packet header, obtain intra-network information, and write the intra-network information into the packet header. When the confirmation packet reaches the end-side server, the packet header is analyzed, the in-network information is extracted, on one hand, the in-network information is used for calculating the bottleneck point in the data center, on the other hand, the in-network information can also be used for estimating the bottleneck point in the wide area network according to the information, and then the bandwidth utilization rate of the bottleneck point of the whole link network is comprehensively obtained. The sending rate can be calculated by algorithm 1 and the congestion control window adjusted. As can be seen from the above description, in this embodiment, the sending rate capable of avoiding the occurrence of congestion is obtained through three parts, namely, the data center link bottleneck bandwidth utilization detection, the wide area network link bottleneck bandwidth utilization detection, and the link base RTT detection. The bottleneck bandwidth utilization rate can be calculated through the algorithm 2, and the base RTT can be updated through the algorithm 3.
Algorithm 1 rate control mechanism
If the acknowledgement packet arrives:
change＝false；change＝false
U＝CalculateUtilization(ack)
if ack.seq > lastUpdateSeq:
Wc＝W；lastUpdateSeq＝ack.seq
if changeB is true:
ProbeLowerBoundBW(txRate_queue)
if change T is true:
ProbeRTT(rtt_queue)
algorithm 2 link utilization calculation mechanism
CalculateUtilization(ack)
maxu＝0；max_queue＝0；
For each forwarding node i corresponding link L:
if u > maxu:
maxu＝u
if qlen > max _ qlen:
max_qlen＝qlen
storing u, txRate, max _ qlen to u _ queue, txRate _ queue,
in qlen _ queue, the queue size is update _ size, and old data is deleted
If qlen < low _ qlen and changT:
changeT＝true
changeB＝true
for each of the u _ queues:
if u _ queue > low _ u or u _ queue < high _ u:
changeB＝false
for each adjacent two values in u _ queues:
if u _ queue[i]＜low_u and u_queue[i+1]＜low_u
and
changeB＝false
returnmaxu
Algorithm 3 link bottleneck and RTT update mechanism
ProbeLowerBoundBW(txRate_queue)
lowerBW＝avg(txRate_queue)
Storing the lowerBW into a lowerBW _ queue, the queue size being history _ size,
and delete old data
lowerBW＝max(lowerBW_queue)；max_u＝0
For each hop P4 switch i that the acknowledgement packet passes through:
if u [ i ] > maxu:
maxu＝u[i]；index＝i
L[links[index]].B＝lowerBW
ProbeRTT(rtt_queue)
storing RTT into RTT _ queue, queue size is history _ size, and deleting old data
baseRTT＝min(rtt_queue)
The embodiment includes a data center link bottleneck bandwidth utilization rate detection module, a wide area network link bottleneck bandwidth utilization rate detection module, and a link base RTT detection module, and the following focuses on a specific working process of the wide area network link bottleneck bandwidth utilization rate detection module and the link base RTT detection module:
a wide area network link bottleneck bandwidth utilization detection module: in order to realize the detection of the bottleneck bandwidth of the wide area network, data detected by the bottleneck bandwidth of the data center, namely the forwarding rate txRate, is required. When the bottleneck point of the whole link is positioned in the data center link, the link utilization rate of the bottleneck point can be calculated in the process of detecting the bottleneck bandwidth of the data center link. When the bottleneck point of the whole link is located in the wide area network, the utilization rate of the bottleneck bandwidth of the data center calculated in the previous step is greater than the bandwidth utilization rate of the wide area network, and therefore the bottleneck point of the wide area network needs to be calculated to represent the bottleneck point of the whole link network. In this case, it can be reasonably concluded that the wide area network is the point of congestion. Therefore, in this embodiment, when the consecutive history _ size acknowledgment packets reflect that the bottleneck utilization of the data center is lower than the given threshold low _ u and the growth rate of the utilization is lower than low _ u _ growth, the bottleneck point of the wide area network needs to be updated. The present embodiment uses txRate as the forwarding rate of the bottleneck point of the wide area network and is used to update the lowest bandwidth of the entire link. When the next acknowledgement packet arrives, when the forwarding rate of the bottleneck point is calculated, the forwarding rate can be obtained by dividing the calculated forwarding rate by the lowest bandwidth, and at this time, the forwarding rate is updated, which represents the utilization rate of the bottleneck bandwidth of the wide area network and also represents the bandwidth utilization rate of the bottleneck of the whole link.
Link baseRTT detection module: to calculate accurate queue utilization, baseRTT needs to be calculated at the end side. To reduce the overhead of frequent probing, queue information in the acknowledgement packet option is utilized. When the queue length extracted by the packet header is smaller than low _ qlen, which is a given constant close to zero, the queue length at this time is considered to be approximately equal to zero, and the RTT at this time is recorded. An RTT array is maintained of length size RTT, and when the amount of data is greater than this amount, old data is deleted. When calculating the latest base RTT, the average value of the RTTs in the group is calculated.
To sum up, the invention firstly calculates the bandwidth utilization rate of each data center link under the historical sending rate, then calculates the change information of each bandwidth utilization rate in a period of time, and finally obtains the position of the bottleneck of the whole link network according to the change information and the bandwidth utilization rate, namely, the sending rate can be adjusted in a targeted manner only by judging whether the bandwidth utilization rate of the data center link is close to the bottleneck bandwidth utilization rate of the data center link or the bandwidth utilization rate of the wide area network link is close to the bottleneck bandwidth utilization rate of the wide area network link, and obtaining the position of the bottleneck of the bandwidth utilization rate of the whole link network, so as to prevent the whole link network from being congested. The invention can adjust the sending rate before congestion occurs to avoid congestion, and simultaneously, the invention can also maximize the bandwidth utilization rate of the whole link network by adjusting the sending rate while avoiding congestion.
In addition, in the data center, the in-network data acquired from the programmable switch comprises queue length, dequeue timestamp and data volume processed by the switch, and the data volume is transmitted to the end side through the data packet to assist congestion control, so that a post feedback mechanism of a traditional congestion control algorithm is improved, the feedback rate and accuracy of congestion control are improved, and the sending rate is more matched with the existing network resources.
In the WAN link, the WAN link has the characteristic of dynamic change, and the base RTT and the bottleneck bandwidth are dynamically changed, so that the bottleneck point and the base RTT of the WAN can be predicted by the aid of the network information of the existing data center on the end side, accurate link detection and rate control can be realized under the scene of communication between the data centers, congestion is avoided, and the link resource utilization rate is improved.
By utilizing the reserved TCP option field in the protocol stack, a new TCP option is defined in the protocol stack at the end side, and data is added to the new TCP option in the switch in the network, thereby realizing the information transmission at the end side and in the network. The method has the advantages that the data in the network are transmitted efficiently, the protocol stack and the switch are not required to be modified greatly, and feasibility and safety are guaranteed.
Exemplary devices
The embodiment also provides a device of a sending rate control method based on cross-data center network communication, which comprises the following components:
the bandwidth utilization rate calculation module is used for calculating the bandwidth utilization rate of each data center link;
a bandwidth utilization rate change information calculation module, configured to calculate utilization rate change information corresponding to the bandwidth utilization rate of each data center link, where the utilization rate change information is used to represent an increase or decrease of the bandwidth utilization rate;
a bottleneck position calculating module, configured to obtain a link position corresponding to a bottleneck bandwidth utilization ratio of an entire link network according to a bandwidth utilization ratio of each data center link and utilization ratio change information of each data center link, where the entire link network includes the data center link serving as a sending end, the data center link serving as a receiving end, and a wide area network link, and the wide area network link is used to implement cross communication between the data center links;
a sending rate calculation module, configured to control a sending rate of the data center link as a sending end according to a link position corresponding to the bottleneck bandwidth utilization rate
Based on the above embodiments, the present invention further provides a terminal device, and a schematic block diagram thereof may be as shown in fig. 4. The terminal equipment comprises a processor, a memory, a network interface, a display screen and a temperature sensor which are connected through a system bus. Wherein the processor of the terminal device is configured to provide computing and control capabilities. The memory of the terminal equipment comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the terminal device is used for connecting and communicating with an external terminal through a network. The computer program is executed by a processor to implement a sending rate control method based on communication across a data center network. The display screen of the terminal equipment can be a liquid crystal display screen or an electronic ink display screen, and the temperature sensor of the terminal equipment is arranged in the terminal equipment in advance and used for detecting the operating temperature of the internal equipment.
It will be understood by those skilled in the art that the block diagram of fig. 4 is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation to the terminal device to which the solution of the present invention is applied, and a specific terminal device may include more or less components than those shown in the figure, or may combine some components, or have different arrangements of components.
In one embodiment, a terminal device is provided, where the terminal device includes a memory, a processor, and a transmission rate control program based on communication across a data center network stored in the memory and executable on the processor, and when the processor executes the transmission rate control program based on communication across the data center network, the following operation instructions are implemented:
calculating the bandwidth utilization rate of each data center link;
calculating utilization rate change information corresponding to the bandwidth utilization rate of each data center link, wherein the utilization rate change information is used for representing the increase amplitude or the decrease amplitude of the bandwidth utilization rate;
obtaining link positions corresponding to bottleneck bandwidth utilization rates of a whole link network according to the bandwidth utilization rates of the data center links and utilization rate change information of the data center links, wherein the whole link network comprises the data center links serving as a sending end, the data center links serving as a receiving end and a wide area network link, and the wide area network link is used for enabling the data center links to realize cross communication;
and controlling the sending rate of the data center link as a sending end according to the link position corresponding to the bottleneck bandwidth utilization rate.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, databases, or other media used in embodiments provided herein may include non-volatile and/or volatile memory. Non-volatile memory can include read-only memory (ROM), Programmable ROM (PROM), Electrically Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus Direct RAM (RDRAM), direct bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
In summary, the present invention discloses a sending rate control method and device based on cross-data center network communication, the method includes: calculating the bandwidth utilization rate of each data center link; calculating utilization rate change information corresponding to the bandwidth utilization rate of each data center link, wherein the utilization rate change information is used for representing the increase amplitude or the decrease amplitude of the bandwidth utilization rate; obtaining link positions corresponding to bottleneck bandwidth utilization rates of a whole link network according to the bandwidth utilization rates of the data center links and utilization rate change information of the data center links, wherein the whole link network comprises the data center links serving as a sending end, the data center links serving as a receiving end and a wide area network link, and the wide area network link is used for enabling the data center links to realize cross communication; and controlling the sending rate of the data center link as a sending end according to the link position corresponding to the bottleneck bandwidth utilization rate. The invention can adjust the sending rate before congestion occurs to avoid congestion, and simultaneously, the invention can also maximize the bandwidth utilization rate of the whole link network by adjusting the sending rate while avoiding congestion.
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 (14)
1. A sending rate control method based on cross-data center network communication is characterized by comprising the following steps:
calculating the bandwidth utilization rate of each data center link;
calculating utilization rate change information corresponding to the bandwidth utilization rate of each data center link, wherein the utilization rate change information is used for representing the increase amplitude or the decrease amplitude of the bandwidth utilization rate;
obtaining link positions corresponding to bottleneck bandwidth utilization rates of a whole link network according to the bandwidth utilization rates of the data center links and utilization rate change information of the data center links, wherein the whole link network comprises the data center links serving as a sending end, the data center links serving as a receiving end and a wide area network link, and the wide area network link is used for enabling the data center links to realize cross communication;
and controlling the sending rate of the data center link as a sending end according to the link position corresponding to the bottleneck bandwidth utilization rate.
2. The sending rate control method according to claim 1, wherein the obtaining a link position corresponding to a bottleneck bandwidth utilization of an entire link network according to a bandwidth utilization of each of the data center links and a utilization change information of each of the data center links, the entire link network includes the data center link as a sending end, the data center link as a receiving end, and a wide area network link, and the wide area network link is configured to implement cross communication between the data center links, includes:
obtaining an increase rate corresponding to the bandwidth utilization rate in the utilization rate change information according to the utilization rate change information;
and when the bandwidth utilization rate of each data center link is smaller than a set value and the growth rate corresponding to the bandwidth utilization rate is smaller than a preset value, obtaining that the link position corresponding to the bottleneck bandwidth utilization rate of the whole link network is positioned on the wide area network link.
3. The sending rate control method based on communication across data center networks according to claim 2, wherein the controlling the sending rate of the data center link as the sending end according to the link position corresponding to the bottleneck bandwidth utilization ratio comprises:
obtaining the link position corresponding to the bottleneck bandwidth utilization rate of the whole link network on the wide area network link according to the link position corresponding to the bottleneck bandwidth utilization rate;
acquiring historical sending speed v1 and historical sending speed v2 of the data center link as a sender, wherein the historical sending speed v1 is different from the historical sending speed v 2;
calculating a forwarding rate V1 aiming at a historical sending rate V1 in each data center link, wherein the forwarding rate is the data volume forwarded by the data center link in unit time;
calculating a forwarding rate V2 aiming at the historical sending rate V2 in each data center link;
and controlling the sending rate of the data center link as a sender at the next moment according to the historical sending rate V1, the historical sending rate V2, the forwarding rate V1 and the forwarding rate V2.
4. The sending rate control method based on communication across data center networks according to claim 3, wherein the controlling the sending rate of the data center link as the sender at the next moment according to the historical sending rate V1, the historical sending rate V2, the forwarding rate V1 and the forwarding rate V2 comprises:
obtaining a maximum forwarding rate V1max in the forwarding rates V1 according to the forwarding rate V1;
obtaining a maximum forwarding rate V2max in the forwarding rates V2 according to the forwarding rate V2;
and controlling the sending rate of the data center link as a sending end at the next moment according to the maximum forwarding rate V1max, the historical sending rate V1, the maximum forwarding rate V2max and the historical sending rate V2.
5. The sending rate control method according to claim 1, wherein the obtaining a link position corresponding to a bottleneck bandwidth utilization of an entire link network according to a bandwidth utilization of each of the data center links and a utilization change information of each of the data center links, the entire link network includes the data center link as a sending end, the data center link as a receiving end, and a wide area network link, and the wide area network link is configured to implement cross communication between the data center links, includes:
and when the bandwidth utilization rate of each data center link is greater than a set value, obtaining that the link position corresponding to the bottleneck bandwidth utilization rate of the whole link network is positioned on the data center link.
6. The sending rate control method based on communication across data center networks according to claim 5, wherein the controlling the sending rate of the data center link as the sending end according to the link position corresponding to the bottleneck bandwidth utilization ratio comprises:
obtaining the link position corresponding to the bottleneck bandwidth utilization rate of the whole link network on the data center link according to the link position corresponding to the bottleneck bandwidth utilization rate;
acquiring bottleneck bandwidth utilization rate of each data center link;
acquiring a historical sending rate v' of the data center link as a sending end;
calculating the bandwidth utilization rate f corresponding to each data center link and the historical sending rate v';
and controlling the sending rate of the data center link as a sending end at the next moment according to the bottleneck bandwidth utilization rate, the bandwidth utilization rate f and the historical sending rate v' of each data center link.
7. The sending rate control method according to claim 6, wherein the controlling the sending rate of the data center link as the sending end at the next time according to the bottleneck bandwidth utilization rate, the bandwidth utilization rate f and the historical sending rate v' of each data center link comprises:
calculating a difference value between the bottleneck bandwidth utilization rate and the bandwidth utilization rate f according to the bottleneck bandwidth utilization rate and the bandwidth utilization rate f;
and controlling the sending rate of the data center link as the sending end at the next moment according to the difference and the historical sending rate v'.
8. The method of claim 1, wherein calculating the bandwidth utilization of each data center link comprises:
calculating a forwarding rate corresponding to each data center link, wherein the forwarding rate is a data volume forwarded by the data center link in unit time;
calculating a queue utilization rate corresponding to each data center link, wherein the queue is located in the data center link, and the queue utilization rate is used for representing the occupied degree of the queue;
and obtaining the bandwidth utilization rate of each data center link according to the forwarding rate and the queue utilization rate of each data center link.
9. The sending rate control method based on communication across data center networks according to claim 8, wherein the obtaining the bandwidth utilization rate of each data center link according to the forwarding rate and the queue utilization rate of each data center link comprises:
and adding the forwarding rate and the queue utilization rate to obtain the bandwidth utilization rate of each data center link.
10. The method according to claim 8, wherein the calculating a forwarding rate corresponding to each of the data center links, the forwarding rate being a data amount forwarded by the data center link in a unit time, comprises:
acquiring timestamp information and switch data processing capacity corresponding to the timestamp information through a confirmation packet, wherein the confirmation packet is confirmation information sent to a data center link of a receiving end after the data center link serving as the receiving end receives data sent by the data center link serving as a sending end, and the switch is located in the data center link;
and obtaining the forwarding rate corresponding to each data center link according to the timestamp information and the data processing amount.
11. The method of claim 8, wherein the calculating a queue utilization corresponding to each of the data center links, the queue being located in the data center link, the queue utilization being used to characterize the degree to which the queue is occupied comprises:
acquiring the queuing length of the queue through a confirmation packet, wherein the confirmation packet is confirmation information sent to the data center link of a receiving end after the data center link of the receiving end receives data sent by the data center link of a sending end, and the switch is positioned in the data center link;
obtaining the round trip delay corresponding to the whole link network;
acquiring a bandwidth corresponding to the switch;
calculating the bandwidth times the round trip delay to obtain a product result;
and dividing the queuing length by the product result to obtain the queue utilization rate corresponding to each data center link.
12. An apparatus for a transmission rate control method based on cross-data center network communication, the apparatus comprising:
the bandwidth utilization rate calculation module is used for calculating the bandwidth utilization rate of each data center link;
a bandwidth utilization rate change information calculation module, configured to calculate utilization rate change information corresponding to the bandwidth utilization rate of each data center link, where the utilization rate change information is used to represent an increase or decrease of the bandwidth utilization rate;
a bottleneck position calculating module, configured to obtain a link position corresponding to a bottleneck bandwidth utilization ratio of an entire link network according to a bandwidth utilization ratio of each data center link and utilization ratio change information of each data center link, where the entire link network includes the data center link serving as a sending end, the data center link serving as a receiving end, and a wide area network link, and the wide area network link is used to implement cross communication between the data center links;
and the sending rate calculation module is used for controlling the sending rate of the data center link as a sending end according to the link position corresponding to the bottleneck bandwidth utilization rate.
13. A terminal device, characterized in that the terminal device comprises a memory, a processor and a transmission rate control program based on communication across a data center network, which is stored in the memory and can be run on the processor, and when the processor executes the transmission rate control program based on communication across a data center network, the steps of the transmission rate control method based on communication across a data center network according to any one of claims 1 to 11 are implemented.
14. A computer-readable storage medium, wherein the computer-readable storage medium has stored thereon a transmission rate control program based on communication across a data center network, and when the transmission rate control program based on communication across a data center network is executed by a processor, the steps of the transmission rate control method based on communication across a data center network according to any one of claims 1 to 11 are implemented.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111617446.5A CN114401230B (en) | 2021-12-27 | 2021-12-27 | Transmission rate control method and device based on cross-data center network communication |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111617446.5A CN114401230B (en) | 2021-12-27 | 2021-12-27 | Transmission rate control method and device based on cross-data center network communication |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114401230A true CN114401230A (en) | 2022-04-26 |
| CN114401230B CN114401230B (en) | 2023-10-20 |
Family
ID=81228969
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111617446.5A Active CN114401230B (en) | 2021-12-27 | 2021-12-27 | Transmission rate control method and device based on cross-data center network communication |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114401230B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114866474A (en) * | 2022-04-29 | 2022-08-05 | 鹏城实验室 | End network cooperative traffic scheduling method, device, system and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080037420A1 (en) * | 2003-10-08 | 2008-02-14 | Bob Tang | Immediate ready implementation of virtually congestion free guaranteed service capable network: external internet nextgentcp (square waveform) TCP friendly san |
| CN103825775A (en) * | 2013-12-31 | 2014-05-28 | 广东工业大学 | Multi-hop wireless network available bandwidth real-time detection method for adaptively detecting packet length |
| CN107171969A (en) * | 2016-03-07 | 2017-09-15 | 华为技术有限公司 | A kind of data transmission method, apparatus and system |
| US20180034740A1 (en) * | 2015-02-11 | 2018-02-01 | Telefonaktiebolaget Lm Ericsson (Publ) | Ethernet congestion control and prevention |
-
2021
- 2021-12-27 CN CN202111617446.5A patent/CN114401230B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080037420A1 (en) * | 2003-10-08 | 2008-02-14 | Bob Tang | Immediate ready implementation of virtually congestion free guaranteed service capable network: external internet nextgentcp (square waveform) TCP friendly san |
| CN103825775A (en) * | 2013-12-31 | 2014-05-28 | 广东工业大学 | Multi-hop wireless network available bandwidth real-time detection method for adaptively detecting packet length |
| US20180034740A1 (en) * | 2015-02-11 | 2018-02-01 | Telefonaktiebolaget Lm Ericsson (Publ) | Ethernet congestion control and prevention |
| CN107171969A (en) * | 2016-03-07 | 2017-09-15 | 华为技术有限公司 | A kind of data transmission method, apparatus and system |
Non-Patent Citations (1)
| Title |
|---|
| 黄俊琳;雷凯;汪漪;: "基于BBR的NDN拥塞控制算法", 重庆邮电大学学报(自然科学版), no. 01, pages 115 - 122 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114866474A (en) * | 2022-04-29 | 2022-08-05 | 鹏城实验室 | End network cooperative traffic scheduling method, device, system and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114401230B (en) | 2023-10-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9210063B2 (en) | Bandwidth policing apparatus and packet relay apparatus | |
| Lakshman et al. | Window-based error recovery and flow control with a slow acknowledgement channel: a study of TCP/IP performance | |
| JP4738594B2 (en) | Data flow control method and apparatus | |
| Dukkipati et al. | Processor sharing flows in the internet | |
| Snoeren | Adaptive inverse multiplexing for wide-area wireless networks | |
| US20180359185A1 (en) | System and method for a tcp mapper | |
| WO2021148020A1 (en) | Service class adjustment method, apparatus, device and storage medium | |
| Shihada et al. | A novel implementation of TCP Vegas for optical burst switched networks | |
| US8630204B1 (en) | System and method for a TCP mapper | |
| CN112737940B (en) | Data transmission method and device | |
| CN114401230B (en) | Transmission rate control method and device based on cross-data center network communication | |
| US20150358238A1 (en) | System and method for a tcp mapper | |
| Xiong et al. | Lambdastream–a data transport protocol for streaming network-intensive applications over photonic networks | |
| Sup et al. | Explicit non-congestion notification: A new AQM approach for TCP networks | |
| US10063489B2 (en) | Buffer bloat control | |
| CN114389999B (en) | Network transmission method, device, equipment and storage medium based on data packet | |
| Irawan et al. | Performance evaluation of queue algorithms for video-on-demand application | |
| Farahmand et al. | A closed-loop rate-based contention control for optical burst switched networks | |
| Lu et al. | EQF: An explicit queue-length feedback for TCP congestion control in datacenter networks | |
| CN112019443A (en) | Multi-path data transmission method and device | |
| Raniwala et al. | Evaluation of a stateful transport protocol for multi-channel wireless mesh networks | |
| Zampognaro | Enabling CoDel AQM with TCP Cubic connections over satellite links | |
| Kato et al. | How coarse-grained clock impacts on performance of ndn rate-based congestion control with explicit rate reporting | |
| CN110650491B (en) | Forward error correction FEC parameter analysis method for vehicle-mounted ad hoc network communication | |
| Peng et al. | Delay performance analysis of token-based fast TCP in window limit systems |
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 |