阅读说明：本技术 一种网络拥塞检测方法及装置 (Network congestion detection method and device ) 是由 杨晓云 于 2020-06-04 设计创作，主要内容包括：本发明公开了一种网络拥塞检测方法及装置。该方法包括如下步骤：实时向网络接收端发送网络数据包；实时接收网络接收端发送的反馈数据包,获得每个反馈数据包携带的各数据包组的发送时刻信息和接收时刻信息；其中,数据包组为网络接收端根据预设的时长周期对实时接收到的网络数据包进行分组得到的；根据各数据包组的发送时刻信息和接收时刻信息,采用双指数平滑滤波法,确定各数据包组对应的均衡平滑延迟；根据各数据包组的发送时刻信息、接收时刻信息,以及各数据包组对应的均衡平滑延迟,确定网络拥塞趋势值；根据网络拥塞趋势值确定当前的网络拥塞情况。本发明能够较好地满足视频会议系统对网络拥塞检测的要求。(The invention discloses a network congestion detection method and device. The method comprises the following steps: transmitting a network data packet to a network receiving end in real time; receiving feedback data packets sent by a network receiving end in real time, and acquiring sending time information and receiving time information of each data packet group carried by each feedback data packet; the data packet group is obtained by grouping the network data packets received in real time by the network receiving end according to a preset time period; determining the balanced smooth delay corresponding to each data packet group by adopting a bi-exponential smooth filtering method according to the sending time information and the receiving time information of each data packet group; determining a network congestion tendency value according to the sending time information and the receiving time information of each data packet group and the balanced smooth delay corresponding to each data packet group; and determining the current network congestion situation according to the network congestion tendency value. The invention can better meet the requirement of the video conference system on network congestion detection.)

1. A network congestion detection method is characterized by comprising the following steps:

transmitting a network data packet to a network receiving end in real time;

receiving feedback data packets sent by a network receiving end in real time, and acquiring sending time information and receiving time information of each data packet group carried by each feedback data packet; the data packet group is obtained by grouping the network data packets received in real time by the network receiving end according to a preset time period;

determining the balanced smooth delay corresponding to each data packet group by adopting a bi-exponential smooth filtering method according to the sending time information and the receiving time information of each data packet group;

determining a network congestion tendency value according to the sending time information and the receiving time information of each data packet group and the balanced smooth delay corresponding to each data packet group;

and determining the current network congestion situation according to the network congestion tendency value.

2. The method according to claim 1, wherein obtaining the sending time information and the receiving time information of each data packet group carried by each feedback data packet comprises:

and obtaining the sending time of the last data packet, the receiving time of the last data packet and the receiving time of the first data packet of each data packet group in each data packet group carried by each feedback data packet.

3. The method of claim 2, wherein determining the equalization smoothing delay corresponding to each packet group by using a bi-exponential smoothing filtering method according to the sending time information and the receiving time information of each packet group comprises:

determining the sending time interval delta corresponding to the data packet group of each time period according to the sending time of the last data packet of the data packet group of each time period and the data packet group of the last time period_{timestamp，i}(ii) a Therein, delta_{timestamp，i}＝T_i-T_i-1(ii) a i represents the packet group number of the ith duration period; t is_iThe sending time of the last data packet of the data packet group of the ith duration period; t is_i-1The time of sending the last data packet of the data packet group of the (i-1) th time period;

determining the receiving time corresponding to the data packet group of each time period according to the receiving time of the last data packet of the data packet group of each time period and the data packet group of the last time periodSpaced apart by a distance of delta_arrival，i(ii) a Therein, delta_arrival，i＝t_i-t_i-1(ii) a i represents the packet group number of the ith duration period; t is t_iThe receiving time of the last data packet of the data packet group of the ith duration period; t is t_i-1The receiving time of the last data packet of the data packet group of the ith-1 th time period;

according to the sending time interval delta corresponding to the data packet group of each time length period_{timestamp，i}And a receiving time interval delta_arrival，iDetermining the delay of the packet transmission increase for each duration period_i(ii) a Wherein, delay_i＝delta_arrival，i-delta_{timestamp，i}；

A delay of the transmission of the data packet according to the respective time period and the respective preceding time period₀To delay_iDetermining the overlap delay acc of the packet groups of each duration period_delayi(ii) a Wherein, acc_delayi＝∑delay₀+delay₁+…+delay_i；

Delay acc of packet group according to each duration period_delayiDetermining the balanced smooth delay smo corresponding to each data packet group by adopting a double-exponential smooth filtering method_delayi；

Wherein:

smo_delayi＝alpha*acc_delayi+(1-alpha)*(smo_delayi-1+b_delayi-1)

b_delayi＝alpha*(smo_delayi-smo_delayi-1)+(1-alpha)*b_delayi-1

wherein, smo_delayiThe equalization smooth delay corresponding to the data packet group of the ith duration period is represented; alpha represents a preset forgetting factor; smo_delayi-1The equalization smooth delay corresponding to the data packet group of the (i-1) th duration period is represented; b_delayiA correction term corresponding to the data packet group of the ith duration period; b_delayi-1And the correction terms correspond to the data packet groups of the (i-1) th time period.

4. The method of claim 3, wherein determining the network congestion tendency value according to the sending time information and the receiving time information of each data packet group and the corresponding equalization and smoothing delay of each data packet group comprises:

delay acc of packet group according to each duration period_delayiDetermining the cumulative delay mean x of the packet groups for each duration period_i(ii) a Wherein:

equalizing smooth delay smo of packet group according to each time length period_delayiDetermining the average value y of the equalized smooth delay of the packet group of each duration period_i(ii) a Wherein:

according to the receiving time t of the last data packet of the current ith data packet group_iAnd the first packet reception time first_{arrival_i}Determining the transmission duration trans of the current ith data packet group_iWherein: trans_i＝t_i-first_{arrival_i}；

Cumulative delay average x of packet groups according to each duration period_iEqualizing the smoothed delay average y_iEqualizing the smoothing delay smo_delayiAnd the transmission duration trans of the current ith data packet group_iDetermining the network congestion tendency value trenline of the current ith data packet group_i(ii) a Wherein the content of the first and second substances,where k represents the packet group of the kth duration period.

5. The method of claim 4, wherein determining the current network congestion situation according to the network congestion tendency value comprises:

determining the network congestion tendency value trenline of the current ith data packet group_iWhether the current value is greater than a preset trend threshold value;

network congestion tendency value trenline of the current ith data packet group_iWhen the current network congestion is greater than a preset trend threshold value, determining the current network congestion;

network congestion tendency value trenline of the current ith data packet group_iAnd when the current network is smaller than or equal to the preset trend threshold, determining that the current network is unobstructed.

6. A network congestion detection apparatus, comprising:

the transmitting unit is used for transmitting the network data packet to a network receiving end in real time;

the receiving unit is used for receiving the feedback data packets sent by the network receiving end in real time and acquiring the sending time information and the receiving time information of each data packet group carried by each feedback data packet; the data packet group is obtained by grouping the network data packets received in real time by the network receiving end according to a preset time period;

the equalization smooth delay calculation unit is used for determining the equalization smooth delay corresponding to each data packet group by adopting a bi-exponential smoothing filtering method according to the sending time information and the receiving time information of each data packet group;

the network congestion tendency value calculating unit is used for determining a network congestion tendency value according to the sending time information and the receiving time information of each data packet group and the balanced smooth delay corresponding to each data packet group;

and the network congestion detection unit is used for determining the current network congestion situation according to the network congestion tendency value.

7. A computer-readable storage medium having stored thereon instructions which, when run on a computer, cause the computer to perform a network congestion detection method according to any one of claims 1 to 5.

8. A computer program product comprising instructions which, when run on a computer, cause the computer to perform a method of network congestion detection as claimed in any one of claims 1 to 5.

Technical Field

The invention relates to a network congestion detection method and a corresponding network congestion detection device, belonging to the technical field of video conferences.

Background

With the development of multimedia as a basic trend of network communication systems, the demand of multimedia network communication mainly based on services such as audio and video conferences becomes increasingly strong. Compared with other data services, the video conference has high bandwidth requirements on the network, but the bandwidth requirements of video transmission data on the network cannot be met due to the fact that the current network often has the problems of bandwidth limitation such as too low bandwidth, drastic change and the like, so that the video transmission data often has the phenomena of packet loss, picture mosaic, blockage and the like, the video playing quality is affected, and the service quality cannot be guaranteed.

Wherein the degree of network congestion control can seriously affect the video transmission quality. The main reason for network congestion is that the network is not able to provide enough resources to meet the user's needs, including buffer space, link bandwidth capacity and the processing power of the intermediate nodes. Because of the lack of "admission control" capability due to internet mechanisms, usage cannot be limited when network resources are scarce, and users can only be served by decreasing quality of service, i.e., "best effort" service.

In a video conference system, because the amount of data to be transmitted by a network is large, network congestion is easily caused, so that the video conference system can show the phenomena of increased data packet delay, increased packet loss probability, reduced performance of an upper application system and the like. Therefore, whether the network is congested or not needs to be detected in real time, so that a worker can dynamically adjust the code rate according to the network congestion detection result, and the network resource node is prevented from being exhausted due to overlarge incoming data flow.

Disclosure of Invention

The invention provides a network congestion detection method.

Another technical problem to be solved by the present invention is to provide a network congestion detection apparatus.

In order to achieve the purpose, the invention adopts the following technical scheme:

according to a first aspect of the embodiments of the present invention, there is provided a network congestion detection method, including the following steps:

transmitting a network data packet to a network receiving end in real time;

receiving feedback data packets sent by a network receiving end in real time, and acquiring sending time information and receiving time information of each data packet group carried by each feedback data packet; the data packet group is obtained by grouping the network data packets received in real time by the network receiving end according to a preset time period;

determining the balanced smooth delay corresponding to each data packet group by adopting a bi-exponential smooth filtering method according to the sending time information and the receiving time information of each data packet group;

determining a network congestion tendency value according to the sending time information and the receiving time information of each data packet group and the balanced smooth delay corresponding to each data packet group;

and determining the current network congestion situation according to the network congestion tendency value.

Preferably, the obtaining of the sending time information and the receiving time information of each data packet group carried by each feedback data packet includes:

and obtaining the sending time of the last data packet, the receiving time of the last data packet and the receiving time of the first data packet of each data packet group in each data packet group carried by each feedback data packet.

Preferably, determining the equalization smoothing delay corresponding to each data packet group by using a bi-exponential smoothing filtering method according to the sending time information and the receiving time information of each data packet group, includes:

determining the sending time interval delta corresponding to the data packet group of each time period according to the sending time of the last data packet of the data packet group of each time period and the data packet group of the last time period_{timestamp，i}(ii) a Therein, delta_{timestamp，i}＝T_i-T_i-1(ii) a i represents the packet group number of the ith duration period; t is_iThe sending time of the last data packet of the data packet group of the ith duration period; t is_i-1The time of sending the last data packet of the data packet group of the (i-1) th time period;

determining the receiving time interval delta corresponding to the data packet group of each time length period according to the receiving time of the last data packet of the data packet group of each time length period and the data packet group of the last time length period_arrival，i(ii) a It is composed ofOf medium, delta_arrival，i＝t_i-t_i-1(ii) a i represents the packet group number of the ith duration period; t is t_iThe receiving time of the last data packet of the data packet group of the ith duration period; t is t_i-1The receiving time of the last data packet of the data packet group of the ith-1 th time period;

according to the sending time interval delta corresponding to the data packet group of each time length period_{timestamp，i}And a receiving time interval delta_arrival，iDetermining the delay of the packet transmission increase for each duration period_i(ii) a Wherein, delay_i＝delta_arrival，i-delta_{timestamp，i}；

A delay of the transmission of the data packet according to the respective time period and the respective preceding time period₀To delay_iDetermining the overlap delay acc of the packet groups of each duration period_delayi(ii) a Wherein, acc_delayi＝∑delay₀+delay₁+…+delay_i；

Delay acc of packet group according to each duration period_delayiDetermining the balanced smooth delay smo corresponding to each data packet group by adopting a double-exponential smooth filtering method_delayi；

Wherein:

smo_delayi＝alpha*acc_delayi+(1-alpha)*(smo_delayi-1+b_delayi-1)

b_delayi＝alpha*(smo_delayi-smo_delayi-1)+(1-alpha)*b_delayi-1

wherein, smo_delayiThe equalization smooth delay corresponding to the data packet group of the ith duration period is represented; alpha represents a preset forgetting factor; smo_delayi-1The equalization smooth delay corresponding to the data packet group of the (i-1) th duration period is represented; b_delayiA correction term corresponding to the data packet group of the ith duration period; b_delayi-1And the correction terms correspond to the data packet groups of the (i-1) th time period.

Preferably, the determining the network congestion tendency value according to the sending time information and the receiving time information of each data packet group and the balanced smooth delay corresponding to each data packet group includes:

delay acc of packet group according to each duration period_delayiDetermining the cumulative delay mean x of the packet groups for each duration period_i(ii) a Wherein:

equalizing smooth delay smo of packet group according to each time length period_delayiDetermining the average value y of the equalized smooth delay of the packet group of each duration period_i(ii) a Wherein:

according to the receiving time t of the last data packet of the current ith data packet group_iAnd the first packet reception time first_{arrival_i}Determining the transmission duration trans of the current ith data packet group_iWherein: trans_i＝t_i-first_{arrival_i}；

Cumulative delay average x of packet groups according to each duration period_iEqualizing the smoothed delay average y_iEqualizing the smoothing delay smo_delayiAnd the transmission duration trans of the current ith data packet group_iDetermining the network congestion tendency value trenline of the current ith data packet group_i(ii) a Wherein the content of the first and second substances,where k represents the packet group of the kth duration period.

Preferably, determining the current network congestion situation according to the network congestion tendency value includes:

determining the network congestion tendency value trenline of the current ith data packet group_iWhether the current value is greater than a preset trend threshold value;

network of the current ith data packet groupNetwork congestion tendency value trenline_iWhen the current network congestion is greater than a preset trend threshold value, determining the current network congestion;

network congestion tendency value trenline of the current ith data packet group_iAnd when the current network is smaller than or equal to the preset trend threshold, determining that the current network is unobstructed.

According to a second aspect of the embodiments of the present invention, there is provided a network congestion detection apparatus, including:

the transmitting unit is used for transmitting the network data packet to a network receiving end in real time;

the receiving unit is used for receiving the feedback data packets sent by the network receiving end in real time and acquiring the sending time information and the receiving time information of each data packet group carried by each feedback data packet; the data packet group is obtained by grouping the network data packets received in real time by the network receiving end according to a preset time period;

the equalization smooth delay calculation unit is used for determining the equalization smooth delay corresponding to each data packet group by adopting a bi-exponential smoothing filtering method according to the sending time information and the receiving time information of each data packet group;

the network congestion tendency value calculating unit is used for determining a network congestion tendency value according to the sending time information and the receiving time information of each data packet group and the balanced smooth delay corresponding to each data packet group;

and the network congestion detection unit is used for determining the current network congestion situation according to the network congestion tendency value.

According to a third aspect of embodiments of the present invention, there is provided a computer-readable storage medium having stored thereon instructions, which, when run on a computer, cause the computer to execute the network congestion detection method of the first aspect described above.

According to a fourth aspect of embodiments of the present invention, there is provided a computer program product containing instructions which, when run on a computer, cause the computer to perform the network congestion detection method of the first aspect described above.

The invention provides a network congestion detection method and a network congestion detection device, which can detect whether a network is congested or not in real time so that a worker can dynamically adjust a code rate according to a network congestion detection result, and therefore the network can be prevented from exhausting network resource nodes due to overlarge data flow. The invention adopts a double-exponential smoothing filtering method to determine the balanced smooth delay corresponding to each data packet group, namely, secondary exponential smoothing is carried out on the basis of primary exponential smoothing filtering, the development direction and the development trend of a curve are found out by utilizing the law of hysteresis deviation, and then a linear trend prediction model is adopted to determine the network congestion trend value according to the sending time information and the receiving time information of each data packet group and the balanced smooth delay corresponding to each data packet group, so that the invention has better network condition prediction effect and can reflect the current real-time network condition more timely.

Drawings

Fig. 1 is a first flowchart of a network congestion detection method according to an embodiment of the present invention;

fig. 2 is a second flowchart of a network congestion detection method according to an embodiment of the present invention;

fig. 3 is a schematic diagram of a network sending end (Sender) sending a plurality of network packets to a network receiving end (Receiver) in the embodiment of the present invention;

FIG. 4 is a graph of experimental data provided by an embodiment of the present invention;

fig. 5 is a schematic structural diagram of a network congestion detection apparatus according to an embodiment of the present invention.

Detailed Description

The technical contents of the present invention will be further described in detail with reference to the accompanying drawings and specific embodiments.

In the process of implementing the present invention, the inventor finds that, currently, in the calculation of the network congestion tendency, a method of one-time exponential smoothing filtering can be generally adopted, while a one-time exponential smoothing filtering method is generally applied to a scene in which linear data and a time sequence have no obvious trend change, and meanwhile, one-time exponential smoothing has obvious hysteresis, so that the estimated network congestion tendency value cannot timely reflect the current real-time network condition. Based on the fact that a network model in reality usually has a straight line trend along with the change of a time series, obvious lag deviation can be predicted by using a one-time exponential smoothing filtering method.

In order to overcome the above drawbacks, as shown in fig. 1, an embodiment of the present invention provides a network congestion detection method, including the following steps:

step 101, transmitting a network data packet to a network receiving end in real time.

And 102, receiving the feedback data packets sent by the network receiving end in real time, and obtaining the sending time information and the receiving time information of each data packet group carried by each feedback data packet.

The data packet group is obtained by grouping the network data packets received in real time by the network receiving end according to a preset time period.

And 103, determining the balanced smooth delay corresponding to each data packet group by adopting a bi-exponential smooth filtering method according to the sending time information and the receiving time information of each data packet group.

And step 104, determining a network congestion tendency value according to the sending time information and the receiving time information of each data packet group and the balanced smooth delay corresponding to each data packet group.

And 105, determining the current network congestion situation according to the network congestion tendency value.

In order that those skilled in the art will better understand the present invention, embodiments of the present invention will be described below with reference to specific examples. As shown in fig. 2, an embodiment of the present invention provides a network congestion detection method, including:

step 201, sending a network data packet to a network receiving end in real time.

It should be noted that, the network data packet sent here may be a network sending end, and the network sending end may be a computer, a network server, a smart phone, a tablet computer, or the like. The network receiving end can be a computer, a smart phone, a tablet computer and the like. The network data packet may be a video data packet, an audio data packet, or an audio/video data packet, but is not limited thereto.

The network transmitting end transmits the data to the network receiving endWhen a network data packet is received, the network data packet may carry a transmission time (denoted as T in the embodiment of the present invention) at which each network data packet is transmitted_i)。

Thus, after the network receiving end receives the network data packet, the step 202 can be continued.

Step 202, receiving a feedback data packet sent by a network receiving end in real time.

Specifically, the network receiving end may send the feedback data packet to the network sending end in the form of an RTCP RR (Real-time Transport control protocol, Receiver Reports) message. When receiving the network data packets in real time, the network receiving end may group the network data packets received in real time by using a preset time period (for example, one period every 5 ms) to form a plurality of data packet groups.

And 203, obtaining the sending time of the last data packet, the receiving time of the last data packet and the receiving time of the first data packet of each data packet group in each data packet group carried by each feedback data packet.

For example, as shown in fig. 3, a network transmitter (Sender) transmits a plurality of network packets (seq ═ 100 to seq ═ 107) to a network Receiver (Receiver), and the network Receiver groups the plurality of network packets into two packet groups (G1 and G2, only two packet groups are taken as an example here) according to a preset time period, so that, for example, the transmission time of the first packet of each packet group (G1 firststimestamp, G2 firststimestamp), the transmission time of the last packet (G1 laststimestamp ═ T1, G2 laststimestamp ═ T2), the reception time of the first packet (G2 firststimestarrival), the reception time of the last packet (T1, T2), and the number of packets included in the packet group (G1_ size, G2_ size) can all be obtained.

Step 204, determining the sending time interval delta corresponding to the data packet group of each time period according to the data packet group of each time period and the sending time of the last data packet of the data packet group of the last time period_{timestamp，i}(ii) a Therein, delta_{timestamp，i}＝T_i-T_i-1(ii) a i represents the packet group number of the ith duration period; t is_iThe sending time of the last data packet of the data packet group of the ith duration period; t is_i-1The time of the last data packet transmission of the data packet group of the (i-1) th duration period.

Step 205, determining a receiving time interval delta corresponding to the data packet group of each time period according to the receiving time of the last data packet of the data packet group of each time period and the data packet group of the last time period_arrival，i(ii) a Therein, delta_arrival，i＝t_i-t_i-1(ii) a i represents the packet group number of the ith duration period; t is t_iThe receiving time of the last data packet of the data packet group of the ith duration period; t is t_i-1The time of receiving the last data packet of the data packet group of the (i-1) th time period.

Step 206, according to the sending time interval delta corresponding to the data packet group of each time length period_{timestamp，i}And a receiving time interval delta_arrival，iDetermining the delay of the packet transmission increase for each duration period_i(ii) a Wherein, delay_i＝delta_arrival，i-delta_{timestamp，i}。

Step 207, according to each duration period and the delay increased by the transmission of the data packet group of each preceding duration period₀To delay_iDetermining the overlap delay acc of the packet groups of each duration period_delayi(ii) a Wherein, acc_delayi＝∑delay₀+delay₁+…+delay_i。

Step 208, according to the superposition delay acc of the data packet groups of each duration period_delayiDetermining the balanced smooth delay smo corresponding to each data packet group by adopting a double-exponential smooth filtering method_delayi。

Wherein:

smo_delayi＝alpha*acc_delayi+(1-alpha)*(smo_delayi-1+b_delayi-1)

b_delayi＝alpha*(smo_delayi-smo_delayi-1)+(1-alpha)*b_delayi-1

wherein, smo_delayiThe equalization smooth delay corresponding to the data packet group of the ith duration period is represented; alpha represents a preset forgetting factor; smo_delayi-1The equalization smooth delay corresponding to the data packet group of the (i-1) th duration period is represented; b_delayiA correction term corresponding to the data packet group of the ith duration period; b_delayi-1And the correction terms correspond to the data packet groups of the (i-1) th time period.

Wherein, smo_delayiThe information of the queuing delay of the data packet through the network link is reflected, and the growth condition of the network queuing queue is reflected. The estimation method and accuracy of the value are key factors for evaluating the final trend value in the congestion detection method. Therefore, the correction term b is introduced here_delayiTherefore, a double-exponential smoothing filtering method is adopted, namely, secondary exponential smoothing is carried out on the basis of primary exponential smoothing filtering, the development direction and the development trend of the curve are found out by utilizing the law of hysteresis deviation, and then a linear trend prediction model is established, so that the network condition prediction effect is better.

Step 209, delay acc of data packet group according to each time length period_delayiDetermining the cumulative delay mean x of the packet groups for each duration period_i(ii) a Wherein:

step 210, balancing and smoothing delay smo according to data packet group of each duration period_delayiDetermining the average value y of the equalized smooth delay of the packet group of each duration period_i(ii) a Wherein:

step 211, receiving the time t according to the last data packet of the current ith data packet group_iAnd the first packet reception time first_{arrival_i}Determining the transmission duration trans of the current ith data packet group_iWherein:trans_i＝t_i-first_{arrival_i}。

step 212, according to the average value x of the accumulated delay of the data packet group of each time period_iEqualizing the smoothed delay average y_iEqualizing the smoothing delay smo_delayiAnd the transmission duration trans of the current ith data packet group_iDetermining the network congestion tendency value trenline of the current ith data packet group_i(ii) a Wherein the content of the first and second substances,

where k represents the packet group of the kth duration period.

Step 213, determining the network congestion tendency value trenline of the current ith data packet group_iIf it is greater than a preset trend threshold.

Network congestion tendency value trenline of current ith data packet group_iIf the value is greater than the preset trend threshold, step 214 is executed. Otherwise, the network congestion tendency value trenline of the current ith data packet group_iAnd when the value is less than or equal to the preset trend threshold value, step 215 is executed.

Step 214, determining the current network congestion.

Step 215, determining that the current network is unobstructed.

It is worth noting that the delay time for transmitting data at a stable network speed is the data size divided by the rate, and if the data block size is large and exceeds the upper delay limit at a constant network speed, which means that it occupies the transmission time of other subsequent data blocks, and so on, the network generates delay and congestion. Network congestion tendency value trenline_iIf the value is larger, the network delay is more and more serious, and if the value is smaller, the delay is gradually reduced. For example, the above-described trend threshold may be set to 0.

Thus, in step 214, when it is determined that the network is congested, relevant staff may be notified to dynamically adjust the code rate of the network sending end, so as to make the subsequent network smooth.

In order to verify the effect of the embodiment of the present invention, as shown in fig. 4, the embodiment of the present invention provides experimental data to verify the code rate estimated by the double-exponential smoothing filtering algorithm in the present scheme. The curve Lan is the actual sending code rate in the network, the curve Hong is the code rate estimated by adopting the double-exponential smoothing filter algorithm after improvement, and the curve Lv is the original code rate estimated by adopting the one-time exponential smoothing filter algorithm. As can be seen from fig. 4, the improved algorithm in the embodiment of the present invention has a better effect of tracking the real-time network condition, and greatly improves the problems of hysteresis and inaccuracy of the original algorithm.

In addition, as shown in fig. 5, an embodiment of the present invention further provides a network congestion detection apparatus, including:

a sending unit 31, configured to send a network data packet to a network receiving end in real time.

A receiving unit 32, configured to receive feedback data packets sent by a network receiving end in real time, and obtain sending time information and receiving time information of each data packet group carried by each feedback data packet; the data packet group is obtained by grouping the network data packets received in real time by the network receiving end according to a preset time period.

And an equalization smooth delay calculation unit 33, configured to determine an equalization smooth delay corresponding to each data packet group by using a bi-exponential smoothing filtering method according to the sending time information and the receiving time information of each data packet group.

And the network congestion tendency value calculating unit 34 is configured to determine a network congestion tendency value according to the sending time information and the receiving time information of each data packet group and the balanced smooth delay corresponding to each data packet group.

And the network congestion detection unit 35 is configured to determine a current network congestion situation according to the network congestion tendency value.

For a specific implementation manner of the network congestion detection apparatus, reference may be made to the method embodiments corresponding to fig. 1 to fig. 4, which are not described herein again.

In addition, an embodiment of the present invention further provides a computer-readable storage medium, where instructions are stored on the computer-readable storage medium, and when the instructions are run on a computer, the computer is caused to execute the network congestion detection method according to the first aspect.

In addition, an embodiment of the present invention further provides a computer program product containing instructions, which when run on a computer, causes the computer to execute the network congestion detection method according to the first aspect.

The invention provides a network congestion detection method and a network congestion detection device, which can detect whether a network is congested or not in real time so that a worker can dynamically adjust a code rate according to a network congestion detection result, and thus, the network resource node is prevented from being exhausted due to overlarge data flow. The invention adopts a double-exponential smoothing filtering method to determine the balanced smooth delay corresponding to each data packet group, namely, secondary exponential smoothing is carried out on the basis of primary exponential smoothing filtering, the development direction and the development trend of a curve are found out by utilizing the law of hysteresis deviation, and then a linear trend prediction model is adopted to determine the network congestion trend value according to the sending time information and the receiving time information of each data packet group and the balanced smooth delay corresponding to each data packet group, so that the invention has better network condition prediction effect and can reflect the current real-time network condition more timely. Therefore, the method can better meet the requirement of the video conference system on network congestion control in actual application, and improves the call quality of the video conference system.

The above describes the network congestion detection method and apparatus provided by the present invention in detail. It will be apparent to those skilled in the art that various modifications can be made without departing from the spirit of the invention.