WO2008025194A1

WO2008025194A1 - A method and device for data flow controlling

Info

Publication number: WO2008025194A1
Application number: PCT/CN2007/000830
Authority: WO
Inventors: Weimin Qiu
Original assignee: Huawei Technologies Co., Ltd.
Priority date: 2006-08-25
Filing date: 2007-03-15
Publication date: 2008-03-06
Also published as: CN1921450A; CN100459586C

Abstract

A method for data flow control comprises: determining whether the difference between the current time and the time at which the last data packet of the current queue was transmitted is larger than the predetermined average time interval, if so, transmitting the first data packet in the queue and processing the next queue, otherwise, processing the next queue directly, said queues stores the data packets having the same destination address. The device for data flow control comprises a queue-storing module, a scheduling-and-transmitting module and a comparing module.

Description

Data flow control method and device

The present application claims priority to Chinese Patent Application No. 200610111848.7, entitled "Data Flow Control Device and Flow Control Method", filed on August 25, 2006, the entire contents of which is incorporated herein by reference. in.

Technical field

The invention relates to a network data transmission technology, in particular to a data flow control method and device. Background technique

With the development of modern electronic technology, various processor performances have rapidly increased. In the field of telecommunications network core networks, with the development of VOIP (voice over IP) technology, traditional switching technologies based on TDM (Time Division Multiplex) switching are moving to IP-based packet switching. Technological advancement. From the data transmitted by the switching network, the original connection-based data stream is gradually evolved into an IP-based connectionless data stream. Due to the high degree of standardization, low R&D costs, and short development cycle, the all-IP network system will become the development trend of the future network.

In the prior art, the traffic classification technology and the priority scheduling technology are generally used to control the traffic of the IP data packet (indicated by the flow control). Among them, the traffic classification technology usually divides the data flow into a Premium Service (PS) according to the type of service > Assured Service (AS) and Best Effort (BE). The PS implements delay-sensitive services, and its goal is to minimize the waiting time of packets in the queue of the network node, and to have characteristics such as jitter, low packet loss rate, and bandwidth guarantee, such as: IP telephony; AS implements bandwidth-sensitive services, and allocates them. A certain amount of bandwidth, but not strictly guaranteed by the user, is preferentially forwarded by a forwarding priority whose packet identifier is relatively high, for example: a video conferencing application; BE is a traditional service provided by the Internet, and does not provide any user with any QOS (quality of service) guarantee, for example: FTP (File Transfer Protocol), E-mail. When the flow control is performed according to the priority scheduling technology, different services are scheduled according to the priority, and the data with the highest priority is sent first, and then the service with the lower priority is sent. However, because IP packets have the following characteristics: Variable length, it is difficult to calculate accurate traffic; It is bursty, there is no fixed traffic, and it is difficult to predict. Therefore, although the traffic of IP packets can be controlled by the classification technology of the cartridges, it is difficult to meet the requirements of high real-time requirements, such as: voice, real-time image transmission, etc., when the traffic is large and exceeds the system performance, it is difficult to ensure The requirements of QOS are mainly reflected in the following aspects: 1. Because the packet length is uncertain, the calculation of the traffic is inaccurate or not calculated, so the delay of the data stream cannot be controlled within the allowable range;

2. Because the IP protocol has only a simple priority flow control mechanism, in the case of a relatively high priority data stream, many lower priority data cannot be transmitted, resulting in some packets with low real-time performance, such as : Network management information, short messages, etc. are discarded, in the upper layer retransmission mechanism (such as TCP

Under the influence of Transmission Control Protocol (Transmission Control Protocol), a large number of data packets are retransmitted, which leads to network congestion;

3. It is impossible to perform accurate bandwidth control on the end user, and it is not limited to the excessive traffic used by individual users, nor the minimum guaranteed bandwidth of a single user. Therefore, it is not conducive to the operation and management of operators. Summary of the invention

The invention provides a data flow control method and device to improve flow control accuracy and reduce network congestion.

A data flow control method provided by the embodiment of the present invention includes determining whether a difference between a current time and a transmission time of a previous data packet in a current queue is greater than a preset average time interval, where the queue is used to store the same destination address. The packet, if yes, sends the first packet in the queue and then processes the next queue; otherwise, the next queue is processed directly.

The data flow control device provided by the embodiment of the present invention includes: a queue storage module, configured to store a queue of data packets with the same destination address; and a scheduling and forwarding module and a comparison module, where the scheduling and forwarding module is used for The queues stored in the queue storage module are polled, and the comparison module is called to calculate the difference between the current time and the sending time of the last data packet in the current queue; when the difference is greater than a preset average time interval, the sending is performed. The first packet in the queue, then polls the next queue; otherwise, polls the next queue directly.

According to the technical solution provided by the foregoing embodiments of the present invention, the embodiment of the present invention strictly controls the sending of data packets according to preset time conditions, thereby realizing accurate control of the bandwidth of the terminal user and ensuring the accuracy of the flow control; It avoids the loss of real-time data packets and is re-transmitted to cause data packets to be sent, which avoids the access layer users competing for bandwidth and improving the operation quality.

The technical solutions of the present invention are further described in detail below with reference to the accompanying drawings and embodiments. DRAWINGS FIG. 1 is a flowchart of Embodiment 1 of a data flow control method according to the present invention.

FIG. 2 is a schematic structural diagram of Embodiment 1 of a data flow control apparatus according to the present invention.

FIG. 3 is a flowchart of Embodiment 2 of a data flow control method according to the present invention.

FIG. 4 is a schematic structural diagram of Embodiment 2 of a data flow control apparatus according to the present invention.

FIG. 5 is a flowchart of Embodiment 3 of the flow control method according to the present invention.

FIG. 6 is a schematic structural diagram of Embodiment 3 of a data flow control apparatus according to the present invention.

Specifically, the queues of the data packets sent to the user are sequentially processed, and if the current queue satisfies the average time interval condition, the first data packet in the queue is sent to the user, and then the next queue is processed; If the current queue does not meet the average time interval condition, the next queue is processed directly without sending a packet to the user.

1 is a flowchart of Embodiment 1 of a data flow control method according to the present invention, which is executed: Step 101: Determine whether a difference between a current time and a transmission time of a previous data packet in a current queue is greater than a pre-acquired average time interval. If yes, go to step 102; otherwise, go to step 103. The queue is used to store the data packet with the same destination address. The destination address can be the destination port number, the destination IP address, the destination MAC address, or any combination thereof. The average time interval can be set according to actual needs. For example, a fixed average time interval is preset, or it can be set according to the user's subscription traffic corresponding to the destination address.

Step 102: Send the first data packet in the queue. Then, go to step 104.

In step 103, it is determined whether there are other queues. If yes, go to step 104; otherwise, go back to step 101.

Step 104: Poll the next queue and go to step 101. Therefore, the data packet transmission is strictly controlled, the accuracy of the bandwidth control of the terminal user is realized, the accuracy of the flow control is ensured, and the burst traffic is avoided, the impact of the burst traffic on the core network is reduced, and the core is reduced. Network investment.

One of ordinary skill in the art can understand that all or part of the steps in the method of implementing the above embodiments can be completed by a program to instruct related hardware, and the program can be stored in a computer. The storage medium, such as a ROM RAM, a magnetic disk, an optical disk, or the like, is read from a storage medium. FIG. 2 is a schematic structural diagram of Embodiment 1 of the data flow control apparatus of the present invention which can be used to implement the method of FIG.

In this embodiment, the data flow control device includes: a scheduling forwarding module 1, a comparison module 2, and a queue storage module 3, wherein the queue storage module 3 is configured to store a queue of data packets with the same destination address; The queue storing the data packet is polled, and the comparison module 2 is called to calculate the difference between the current time and the transmission time of the previous data packet in the current queue. When the difference is greater than the preset average time interval, the first packet in the queue is sent, and then the next queue is polled; otherwise, the next queue is polled directly.

FIG. 3 is a flowchart of Embodiment 2 of a data flow control method according to the present invention. In this embodiment, an average packet length of data packets in each queue is dynamically calculated, thereby determining an average time interval of data packets leaving each queue. Specifically include:

Step 301: Calculate an average packet length of the data packet in the queue.

In this embodiment, the number of samples of the data packet preset by the user may be obtained from the description information of the queue, the data packet of the sample number is debounced, and then the packet length of the debounced data packet is averaged, thereby obtaining The average packet length of the packet.

Step 302: Determine, according to the average packet length of the data packet and the subscription traffic information of the user that is the same as the destination address of the data packet, the average time interval for sending the adjacent data packet by the queue.

First, the subscription traffic information of the user whose address is the same as the destination address of the data packet is obtained. When calculating the average time interval for sending the adjacent data packet, the average packet length and the subscription traffic may be negotiated to obtain the queue transmission phase. The average time interval between adjacent packets.

Step 303: Determine whether the difference between the current time and the sending time of the last data packet in the current queue is greater than the calculated average time interval. If yes, go to step 304; otherwise, go to step 305.

Step 304, sending the first data packet in the queue. Then, step 306 is performed.

In step 305, it is determined whether there are other queues. If yes, go to step 306; otherwise, go back to step 302.

Step 104: Poll the next queue and go to step 101.

The embodiment of the invention limits the time condition for sending a data packet to the user according to the subscription traffic of the user, thereby strictly controlling the transmission of the data packet, ensuring the accuracy of the flow control, and effectively ensuring the QOS; Calculate the average time interval for sending data packets to users according to the user's contracted traffic and the traffic received by the data packet, avoiding burst traffic, reducing the impact of burst traffic on the core network, and reducing the core network investment; Core network congestion caused by retransmissions after loss of real-time data packets is avoided.

A person skilled in the art can understand that all or part of the steps of implementing the above embodiments can be completed by a program to instruct related hardware, and the program can be stored in a computer readable storage medium, the storage. Media, such as: ROM/RAM, disk, CD, etc.

Figure 4 is a block diagram showing the structure of the second embodiment of the data flow control device of the present invention which can be used to implement the method of Figure 3, in the embodiment shown in Figure 2! ^, a computing processing module 4 connected to the dispatching and forwarding module 1 is added, which is used for calculating the average packet length of the data packet in the queue, and obtaining the subscription traffic information of the user whose address is the same as the destination address of the data packet. And calculating the average time interval for sending adjacent data packets in the queue by evaluating the average packet length and the subscription traffic.

The number of samples corresponding to a destination address, the debounce parameters eight and B, and the parameters N, A, and B can be set in advance, and can be modified at any time. For example: You can choose the default value of A and B as 1 or 2. The sample number is stored in the user information corresponding to the destination address, and the destination address may be a destination port number, a destination IP address, a destination MAC address, or any combination thereof. A queue for each packet for storing the same destination address is established in advance based on the destination address. When the data packet is received, the data flow control method provided by the present invention can be used to flow control the transmission of the data packet. FIG. 5 is a flowchart of Embodiment 3 of the data flow control method of the present invention, which includes the following steps:

Step 501: Receive a data packet, and detect parameter information of the data packet, including a destination address and a packet length information of the data packet, for example, a destination IP address is 10.10.10.101, and a packet length information is 64K.

Step 502: The data packet is stored in a queue with a destination IP address of 10.10.10.101. Meanwhile, the description information is established according to the parameter information of the data packet, where the description information includes a packet length of the data packet and a next data packet. Describe the address of the information, and store the description information of the data packet at the destination IP address of 10.10.10.101. At the same time, obtain the user information with the IP address of 10.10.10.101, and establish the destination IP address according to the description information and the user information. The description information of the queue of 10.10.10.101, the description information of the queue includes the subscription traffic of the user, the number of samples N, the average packet length of the data packet of the sample number N, and the longest packet length of the data packet of the sample number N, The shortest packet length of the packet of the sample number N, the transmission time of the previous packet, and the address information of the first packet in the queue. as follows Table 1 shows an example of the content of the description information of the data packet of a certain destination address. Table 2 below shows an example of the content of the description information of the queue of a certain destination address.

Table 1

Table 2

Tc (contracted traffic)

N (number of samples) Len-avg (average packet length of the first N packets)

Len_max (the longest packet length in the first N packets) Len-min (the shortest packet length in the first N packets)

Tprev (the moment when the last packet was sent)

Pkt -desc (address of the first packet in the queue) In Table 1, since the next packet that is the same as the destination address of the packet has not been received when a packet is received, the packet is The Pkt_nary in the description information is temporarily set to 0, and when the next data packet is received, the Pkt_next; is updated according to the address of the description information of the next data packet.

In Table 2, Tc is pre-applied by the user, stored in the user information, and can be obtained from the user information, and its value can be dynamically adjusted. If Tc is 0, the user is an invalid user. N can be set by the data flow control device administrator when setting up the queue, and can be dynamically adjusted. For the convenience of calculation, the value of N is preferably configured to be n to the power of n, n is a positive integer, for example: 2, 4, 8, 32 , .... Len_avg is the average packet length after de-jittering the first N packets in the queue.

Step 503: Obtain the number of samples of the data packet preset by the user from the description information of the queue to dejoke the parameters A and B, and debounce the first N data packets in the queue, that is, remove the pre-N according to the preset parameters. A longest packet length and B shortest packet length in each data packet (A and B are positive Number, and ^> eight +8), and then average the packet length of the de-jittered N - (A + B) packets, the average value can be recorded as Len - avg, and the Len - avg is written In the description of the queue.

Step 504: Obtain the subscription traffic Tc of the user from the description information of the queue, and use the formula 1/(8*Len_avg/Tc) to negotiate the packet length average Len_avg with the subscription traffic Tc, and obtain the current transmission in the queue. The average time interval T_int of adjacent packets.

Step 505: Determine whether the difference between the current time and the Tprev in the description information of the queue is greater than the calculated current T_int of the queue. If yes, go to step 506; otherwise, go to step 507.

Step 506: Send the first data packet in the queue to its destination address. Then, step 508 is performed.

In step 507, it is determined whether there are other queues. If yes, go to step 509; otherwise, go back to step 504.

Step 508: Update Len_max, Len_min, and Pkt_desc in the description information of the queue according to the description information of the data packet in the queue, and update the description information of the queue by sending the current time of the first data packet. In the Tprev, at the same time, recalculate the average of the packet lengths of the first (N - 1) packets in the queue by the formula (( N-1 ) *Len_avg+Pkt_len-Len_max- Len- min ) / ( N-2 ) Use this average to update Len-avg in the description of the queue.

Step 509, polling the next queue, and performing step 504 to step 508 for the next queue. The packet avoids the bandwidth of the access layer users and improves the quality of the operation. In addition, the present invention can perform flow control based on the destination port number, the destination IP address, or the destination MAC address or any combination thereof, and has high flexibility.

In order to avoid unnecessary subsequent operations and reduce the working efficiency of the data flow control device, before performing step 504, it may first check whether the Tc in the current queue is zero. If yes, the data packet in the queue is not sent, polling The next queue, step 509 is performed; otherwise, step 504 is performed.

FIG. 6 is a schematic structural diagram of Embodiment 3 of the data flow control apparatus of the present invention, which can be used to implement the method of FIG. 5. The embodiment is further provided with a data receiving mode based on the embodiment shown in FIG. Block 5, the user information storage module 6, the monitoring management module 7 respectively connected to the data receiving module 5, the user information storage module 6 and the queue storage module 3, and the data information storage module respectively connected to the monitoring management module 7 and the computing processing module 4 8. The data receiving module 5 is configured to receive the data packet, and the monitoring management module 7 is configured according to the parameter information of the data packet (including the destination address of the data packet (which may be a destination port number, a destination IP address, or a destination MAC address, or any combination thereof) and data. The packet length of the packet is stored in the queue with the same destination address. In addition, the data receiving module 5 can also provide a human-computer interaction interface through which the user can input commands for setting up or deleting the queue and related parameters, for example. The sample number N and the debounce parameters A and B; the parameter information of the data packet received by the monitoring management module 7 according to the data receiving module 5 and the user information stored in the user information storage module 6, including: port number, IP address or MAC address Or any combination thereof, and the number of samples N and de-jitter parameters A and 8, to establish description information of the data packet, including the destination address information of the data packet and the packet length information of the data packet; and, according to the parameter information and user information of the data packet Describe and update the description information of the queue, including: the user's contracted traffic, the number of samples, the The average packet length of the data packet of this number, the longest packet length of the data packet of the sample number, the shortest packet length of the data packet of the sample number, the transmission time of the previous data packet, and the address of the first data packet in the queue The information, and the received data packet are stored in a queue corresponding to the destination address thereof; and the description information of the established data packet and the description information of the queue are saved in the data information storage unit 8.

The calculation processing module 4 is configured to debounce the data packets in each queue, calculate the average packet length, and obtain the subscription traffic information of the user whose address is the same as the destination address of the data packet, and pass the average packet length and the subscription traffic. The quotient is used to calculate the average time interval for sending adjacent packets in the queue. The calculation processing module 4 may first obtain the number of samples of the data packet in the queue from the description information of the queue stored in the data information storage module 8 when calculating the average packet length of the data packet in the queue, and then the number of samples in the queue. The data packet is de-jittered; and the packet length of the de-jittered data packet is averaged to obtain the average packet length of the data packet in the queue.

The scheduling and forwarding module 1 is configured to poll the queue of the stored data packet by the average time interval calculated by the calculation processing module 4, and when the difference between the current time and the sending time of the last data packet in the current queue is greater than the average time interval, The first packet in the queue is sent, then the next queue is polled, otherwise, the next queue is polled directly.

In the above data flow control device, the data information storage module 8, the queue storage module 3, and / Or the user information storage module 6 can be integrally provided. In addition, the data information storage module 8, the queue storage module 3, and/or the user information storage module 6 may also be integrally provided with the monitoring management module 7; the monitoring management module may also be integrally provided with the data receiving module 5.

The foregoing embodiments provided by the present invention can be applied not only to the access layer for flow control based on the port number, the IP address, and/or the MAC address, but also to the core layer for flow control based on the port number.

With the embodiment provided by the present invention, the transmission of the data packet can be strictly controlled to ensure the accuracy and QOS of the flow control;

The burst traffic of the IP data packet is avoided, the impact of the burst traffic on the core network is reduced, and the core network investment is reduced; at the same time, the core network congestion caused by the retransmission of the data packet after the real-time non-strong packet loss is avoided;

The access layer users are prevented from competing for bandwidth and improving the operation quality;

The flow control can be performed based on the destination port number, the destination IP address, or the destination MAC address, or any combination thereof, with high flexibility.

It is to be understood that the above embodiments are merely illustrative of the technical solutions of the present invention and are not to be construed as limiting. Although the present invention has been described in detail with reference to the preferred embodiments thereof, those skilled in the art will understand that the invention may be modified or equivalently substituted without departing from the invention. The spirit and scope of the programme.

Claims

Rights request

A data flow control method, comprising:

Determining whether the difference between the current time and the sending time of the last data packet in the current queue is greater than the average time interval of sending the adjacent #1 data packet by the queue obtained in advance, and the queue is used to store the data packet with the same destination address;

If yes, send the first packet in the queue and then process the next queue; otherwise, process the next queue directly.

2. Method according to claim 1, characterized in that the average time interval is obtained as follows:

Calculate the average packet length of the packets in the queue;

And determining an average time interval for the queue to send adjacent data packets according to the average packet length and the subscription traffic in the user information corresponding to the queue.

3. The method according to claim 2, wherein the step of calculating an average packet length of the data packets in the queue comprises:

Obtaining the number of samples of the data packet in the queue;

De-jittering the data packet of the sample number;

The packet length of the de-jittered data packet is averaged to obtain the average packet length of the data packet in the queue.

The method according to claim 3, wherein the method further comprises: storing the received data packet in a pre-established queue corresponding to its destination address; establishing a description information of the data packet, The description information of the data packet includes packet length information of the data packet and address information of description information of the next data packet;

Determining the description information of the queue according to the description information of the data packet and the user information, where the description information of the queue includes the subscription traffic of the user, the number of samples, the average packet length of the data packet of the sample number, and the data packet of the sample number. The longest packet length, the shortest packet length of the packet of the sample number, the transmission time of the previous packet, and the address information of the first packet in the queue.

5. The method according to claim 4, wherein the method further comprises: updating a next one of the description information of the last data packet in the queue when a new data packet is stored in the queue The address information of the description information of the packet.

The method according to claim 4, wherein the method further comprises the steps of: OP070108

PC layer / 0 0 0 8 3 0

WO 2008/025194 PCT/CN2007/000830

- 11 - After sending the first data packet, updating the description information of the queue to which the data packet belongs.

The method according to claim 4, wherein the step of acquiring the number of samples of the data packet in the queue is specifically:

Obtaining, according to the description information of the queue, the number of samples of the data packet in the queue.

8. The method according to claim 1, wherein the average time interval is obtained according to a value of a preset average time interval.

The method according to any one of claims 1 to 8, characterized in that, in the determining whether the difference between the current time and the sending time of the last data packet in the current queue is greater than the previously sent queue sending neighbor The steps of the average time interval of the packet also include the steps:

Query whether the subscription traffic in the corresponding user information in the current queue is zero;

If yes, performing the step of directly processing the next queue;

Otherwise, the step of determining whether the difference between the current time and the transmission time of the last data packet in the current queue is greater than a pre-obtained average time interval of sending the adjacent data packet by the queue.

The method according to any one of claims 1 to 8, wherein the destination address is a destination port number, a destination IP address or a destination MAC address or any combination thereof.

A data flow control device, comprising: a queue storage module, configured to store a queue of data packets with the same destination address; and the method further includes: a scheduling forwarding module and a comparison module, where the scheduling forwarding module is used for The queues stored in the queue storage module are polled, and the comparison module is called to calculate the difference between the current time and the sending time of the last data packet in the current queue; when the difference is greater than a preset average time interval, the sending is performed. The first packet in the queue, then polls the next queue; otherwise, polls the next queue directly.

The device according to claim 11, wherein the device further comprises: a calculation processing module, connected to the scheduling and forwarding module, configured to calculate an average packet length of the data packet in the queue, and determine The average time interval in which the adjacent packets are sent in the queue.

The device according to claim 12, wherein the device further comprises: a user information storage module, configured to store the user information;

a data information storage module, configured to be connected to the calculation processing module, configured to store description information of the data packet in the queue and a description of the queue required by the calculation processing module to calculate an average packet length of the data packet in the queue information; OP070108

PCT/C 2007 / 0 0 0 8 3 0

WO 2008/025194 PCT/CN2007/000830

—12—

a data receiving module, configured to receive a data packet;

The monitoring management module is respectively connected to the data receiving module, the queue storage module, and the user information storage module, and configured to store the data packet received by the data receiving module in a queue corresponding to the destination address thereof, and Determining or updating the description information of the data packet and the description information of the queue to which it belongs according to the information in the data packet and the user information.