CN115314407A - Network flow based online game QoE detection method - Google Patents
Network flow based online game QoE detection method Download PDFInfo
- Publication number
- CN115314407A CN115314407A CN202210927727.9A CN202210927727A CN115314407A CN 115314407 A CN115314407 A CN 115314407A CN 202210927727 A CN202210927727 A CN 202210927727A CN 115314407 A CN115314407 A CN 115314407A
- Authority
- CN
- China
- Prior art keywords
- qoe
- packet loss
- data
- game
- delay
- 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.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 55
- 238000000034 method Methods 0.000 claims abstract description 36
- 238000012549 training Methods 0.000 claims abstract description 24
- 238000013145 classification model Methods 0.000 claims abstract description 20
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 8
- 238000002474 experimental method Methods 0.000 claims abstract description 8
- 238000010801 machine learning Methods 0.000 claims abstract description 7
- 238000012545 processing Methods 0.000 claims description 17
- 238000000605 extraction Methods 0.000 claims description 11
- 230000001934 delay Effects 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 6
- 238000012790 confirmation Methods 0.000 claims description 4
- 238000007637 random forest analysis Methods 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 238000002372 labelling Methods 0.000 claims description 3
- 238000007619 statistical method Methods 0.000 claims description 3
- 238000013480 data collection Methods 0.000 claims 1
- 238000012544 monitoring process Methods 0.000 abstract description 3
- 238000011160 research Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 3
- 238000005315 distribution function Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012163 sequencing technique Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 230000003111 delayed effect Effects 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0852—Delays
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/02—Capturing of monitoring data
- H04L43/028—Capturing of monitoring data by filtering
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/04—Processing captured monitoring data, e.g. for logfile generation
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0823—Errors, e.g. transmission errors
- H04L43/0829—Packet loss
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
- H04L43/0876—Network utilisation, e.g. volume of load or congestion level
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Environmental & Geological Engineering (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a network flow based QoE (Quality of Experience) detection method for a network game. Firstly, an experimental environment with controllable delay and packet loss is built, a subjective QoE assessment experiment is carried out, data statistics and analysis are carried out according to experimental results, and a QoS-QoE detection model with the delay and the packet loss rate as independent variables is obtained. Then, collecting flow sample data with different packet loss rates in an experimental environment, extracting flow characteristics, optimizing to obtain a packet loss rate training sample set, and training by using a machine learning algorithm to obtain a packet loss rate classification model. And finally, acquiring game flow by using data acquisition equipment, obtaining a packet loss grade through a packet loss rate classification model, obtaining delay data through flow statistics, and substituting a packet loss rate classification result and a delay detection result into a QoS-QoE detection model to realize the detection of the QoE of the user game. The method and the device can detect the QoE of the user game based on the network flow, and can be used for network management and user QoE monitoring.
Description
Technical Field
The invention relates to a QoE (quality of experience) detection method for online games based on network traffic, belonging to the technical field of network security.
Background
The QoE (Quality of Experience) is a concept centered on a user, is very closely associated with the subjective feeling of an end user, and can provide a decision maker, such as a service provider and a network operator, with a wider and more comprehensive understanding about how the service Quality affects the user Experience. In recent years, research on QoE of users in using network services, and building and optimizing internet business services with QoE as a center are important topics of consistent attention in the industry and academia.
For network operators, in practice, the operators have great technical difficulty in perceiving the user QoE. First, as mentioned above, the QoE is a subjective experience of the user, and the operator can only observe data on the network side, and if there is no cooperation between the end user and the operator, the operator cannot know the QoE. Although objective QoS parameters such as delay, packet loss, bandwidth, frame rate, etc. can be used to evaluate the QoE of the user, these parameters are still located at the client (i.e. belong to application layer parameters), and are still difficult for the operator to obtain. For operators, it is a good solution to predict the QoE of users across layers by detecting network traffic.
Currently, researchers have proposed a series of research methods for detecting game QoE, which are mainly classified into research methods for network parameters and video coding parameters according to different types of parameters to be researched, but these methods still have some limitations.
(1) Research method based on network parameters
The research method based on network parameters sets network parameters such as delay, packet loss and jitter between a game server and a user client by using a network simulation device, and then directly analyzes the parameters, however, network parameters required by such a method cannot be obtained by a network provider.
(2) Research method based on video coding parameters
The research method based on the video coding parameters mainly explores the influence of the video coding parameters such as bit rate, frame rate and resolution on the QoE of the game, and therefore the QoE of the game is detected according to the parameters. However, the parameters required by such methods are all present at the client and there is no way for the network provider to obtain them.
Disclosure of Invention
In order to solve the problems, the invention discloses a network flow-based QoE (Quality of Experience) detection method for a network game. Then, collecting flow sample data with different packet loss rates in an experimental environment, extracting flow characteristics, optimizing to obtain a packet loss rate training sample set, and training by using a machine learning algorithm to obtain a packet loss rate classification model. And finally, acquiring game flow by using data acquisition equipment, obtaining a packet loss grade through a packet loss rate classification model, obtaining delay data through flow statistics, and substituting a packet loss rate classification result and a delay detection result into a QoS-QoE detection model to realize the detection of the QoE of the user game. The method and the device can detect the QoE of the user game based on the network flow, and can be used for network management and user QoE monitoring.
In order to realize the purpose of the invention, the technical steps of the scheme are as follows: a network traffic-based online game QoE detection method comprises the following steps:
step (1) carrying out a subjective QoE assessment experiment;
step (2) performing data statistical analysis according to the experimental result obtained in the step (1) to obtain a QoS-QoE detection model;
step (3) carrying out game flow acquisition through data acquisition equipment;
step (4) carrying out data processing, feature extraction and feature optimization on the game flow obtained in the step (3);
step (5) obtaining a large amount of game flow data by using the step (3), processing the data by the step (4) to obtain a packet loss rate training sample set, and training by using a machine learning algorithm to obtain a packet loss rate classification model;
step (6) collecting new game flow, performing data processing and feature extraction through the step (4), and inputting the new game flow into the packet loss rate classification model in the step (5) to identify the packet loss rate grade;
step (7) the game flow statistical data information collected in the step (6) is detected to realize delay;
and (8) substituting the packet loss rate identification result obtained in the step (6) and the delay detection result obtained in the step (7) into the QoS-QoE detection model obtained in the step (2) to detect the QoE of the user game.
Further, in the step (1), the experimental method for subjective QoE assessment is as follows:
(1.1) building an experimental environment for simulating the real network condition;
and (1.2) finding a plurality of users to play games under various network conditions (different delays and packet loss rates), and acquiring the QoE scores of the players after each game is finished.
Further, in the step (2), a specific process of obtaining the QoS-QoE detection model is as follows:
(2.1) calculate average QoE values for all players under each network condition.
And (2.2) fitting the QoE scoring data by utilizing nonlinear fitting to obtain a model for QoE detection according to delay and packet loss rate.
Further, in the step (3), the method for acquiring the game flow rate is as follows:
(3.1) capturing game flow data under different network conditions by using tcpdump operated on an OpenWrt softrouter on the basis of the experimental environment built in the step (1);
and (3.2) after each game is finished, storing the flow data of each game.
Further, in the step (4), the data processing and feature extraction specifically include the following sub-steps:
(4.1) filtering game flow from the flow obtained in the step (3) and segmenting data;
(4.2) extracting the characteristics of the flow data, and labeling all data samples;
and (4.3) inputting the sample data set into a random forest classifier to iteratively optimize the features.
Further, in the step (5), training the packet loss rate classification model specifically includes the following sub-steps:
(5.1) capturing a large amount of game flow data with different packet loss rates through the step (3), and performing data processing, feature extraction and optimization through the step (4) to obtain a packet loss rate training sample set;
(5.2) training the packet loss rate training sample set by using a machine learning algorithm to obtain a packet loss rate classification model;
further, in the step (6), the packet loss rate identification specifically includes the following sub-steps:
(6.1) collecting new game flow;
(6.2) carrying out data processing on the acquired flow through the step (4) and extracting characteristics;
and (6.3) inputting the obtained flow characteristic result into the packet loss rate classification model in the step (5) and identifying the packet loss rate grade.
Further, in the step (7), the detection of the delay specifically includes the following sub-steps:
(7.1) dividing the game flow filtered in the step (6) into data packets and confirmation packets according to the payload length distribution of all the packets, and counting the transmission direction and the time stamp of each packet;
(7.2) calculating the difference value of the time stamps of the data packet and the acknowledgement packet as the alternative unilateral delay;
(7.3) for each alternative single-sided delay, counting the matching logarithm of the data packet and the confirmation packet which meet the alternative single-sided delay;
(7.4) taking the unilateral delay corresponding to the maximum matching logarithm as the optimal unilateral delay, and calculating the ratio of the maximum matching logarithm to all data packets to be used as the judgment of the accuracy of the found optimal matching time;
(7.5) executing the steps (7.2) to (7.4) from the acquisition point to the server side and from the acquisition point to the client side respectively;
and (7.6) adding the optimal unilateral delays in the two directions to obtain a delay result, and taking the average value of the accuracy indexes in the two directions as a delay detection accuracy index.
Further, in the step (8), the packet loss rate level identification result obtained in the step (6) and the delay detection result obtained in the step (7) are substituted into the QoS-QoE detection model obtained in the step (2), so as to realize the detection of the user game QoE.
Compared with the prior art, the technical scheme of the invention has the following advantages:
(1) The invention provides a method for detecting QoE of a network game based on network flow, which can obtain delay and packet loss by monitoring the network flow transmitted between a user and a game server, and further substitutes the network flow into a constructed QoS-QoE relation model, thereby achieving the purpose of detecting the QoE of the user, leading an access service provider to improve the network condition in time according to the detection condition of the QoE of the user, ensuring the user experience quality and having stronger practicability.
(2) The invention provides a method for calculating the end-to-end delay of a user through game data obtained by passive measurement, which can directly and accurately obtain the end-to-end delay between the user and a server through the flow acquired by an access point.
Drawings
FIG. 1 is an overall architecture diagram of the present invention;
FIG. 2 is a confusion matrix diagram of the packet loss rate classification model according to the present invention;
FIG. 3 is a flow chart of single-edge detection for delay detection according to the present invention;
FIG. 4 (a) is a graph of probability density function in a delayed detection procedure;
fig. 4 (b) is a diagram of a cumulative distribution function in the delay detection flow.
Detailed Description
The technical solutions provided by the present invention will be described in detail below with reference to specific examples, and it should be understood that the following specific embodiments are only illustrative of the present invention and are not intended to limit the scope of the present invention.
The specific embodiment is as follows: the invention provides a QoE (quality of experience) detection method for online games based on network traffic, the general architecture of which is shown in figure 1, and the method comprises the following steps:
step (1) carrying out a subjective QoE assessment experiment;
step (2) performing data statistical analysis according to the experimental result obtained in the step (1) to obtain a QoS-QoE detection model;
step (3) carrying out game flow acquisition through data acquisition equipment;
step (4) carrying out data processing, feature extraction and feature optimization on the game flow obtained in the step (3);
step (5) obtaining a large amount of game flow data by using the step (3), processing the data by the step (4) to obtain a packet loss rate training sample set, and training by using a machine learning algorithm to obtain a packet loss rate classification model;
and (6) acquiring new game flow, performing data processing and feature extraction through the step (4), and inputting the new game flow into the packet loss rate classification model in the step (5) to identify the packet loss rate grade.
And (7) detecting delay for the game flow statistical data information collected in the step (6).
And (8) substituting the packet loss rate identification result obtained in the step (6) and the delay detection result obtained in the step (7) into the QoS-QoE detection model obtained in the step (2) to detect the QoE of the user game.
In one embodiment of the present invention, in the step (1), the subjective QoE assessment experiment specifically includes the following steps:
(1.1) establishing an experimental environment for simulating the real network condition, comprising the following steps: the soft route with the OpenWrt firmware is used as a gateway, a Wan interface of the gateway is connected with a campus network, and a Lan interface of the gateway is connected with a PC and a TP-LINK router; setting a routing rule by using a netem network simulation module operated in OpenWrt firmware, and applying additional delay and packet loss rate to a network port connected with a TP-LINK router and a PC;
(1.2) finding a plurality of users with game experience to participate in a subjective QoE experiment aiming at the current popular network game hero alliance, and enabling the users to play games under the conditions of independently applying different delays, independently applying different packet loss rates, jointly applying the delays and the packet loss rates and under the normal condition, wherein the normal condition represents the condition that neither the delay nor the packet loss rate is applied. And then obtaining the game QoE scores of all players for each network condition, wherein the score range is 1 to 5.
In an embodiment of the present invention, in the step (2), the specific steps of obtaining the QoS-QoE detection model are as follows:
(2.1) calculating the average value of the QoE scores of all the players under each network condition, and table 1 shows the average QoE score result of each network condition (the basic delay is about 18ms, and the basic packet loss rate is 0%).
TABLE 1 average QoE score for each network condition
And (2.2) mapping the MOS value of the QoE to delay and packet loss rate by utilizing regression analysis according to the grade data of the QoE, designing a binary nonlinear function, and fitting parameters to obtain the following QoS-QoE detection model.
QoE＝(a+b*Delay+c*Delay 2 +d*PLR 2 +e*Delay*PLR)/(1+e f*Delay+g*PLR )
In the formula, qoE is measured in MOS (mean opinion score), delay (Delay) is in milliseconds, and Packet Loss Rate (PLR) is in percentage. Wherein the parameters a =10.28, b = -0.05543, c = -0.0000912, d = -0.000164, e = -0.0001083, f = -0.006765, g = -0.02816.
The root mean square error sum (RMSE) of the performance evaluation indexes obtained after the fitting of the formula is 0.07883, the value of the number is close to 0, the fitting coefficient (R-Square) is 0.9941, and the value of the number is close to 1; the QoS-QoE detection model has good performance effect, the detection result is basically consistent with the actual QoE, and the method has strong practicability.
In one embodiment of the present invention, in the step (3), the specific steps of obtaining the game flow are as follows:
(3.1) capturing game flow data under different delays and different packet loss rates by using tcpdump operated on the OpenWrt softrouter on the basis of the experimental environment established in the step (1);
and (3.2) after each game is finished, storing the flow data of each game into a pcap format.
In an embodiment of the present invention, in the step (4), the extracting and optimizing the flow characteristics specifically includes the following processes:
(4.1) filtering the game data transmitted between the client and the game server from the game flows with different packet loss rates obtained in the step (3) according to the server IP and the client IP, eliminating the flows from other application programs, and then segmenting the data by taking 10s as a sample;
(4.2) extracting the characteristics of the traffic data, such as the number of packets, the average length of the packets, the maximum value and the minimum value of the packet length, the average time and the standard length between streams and the like, and labeling all data samples;
(4.3) inputting the sample data set with the label into a random forest classifier to perform multiple rounds of training, and performing feature importance evaluation, namely averaging the contribution of each feature in each tree in the random forest, sequencing, eliminating the feature with a lower importance coefficient according to the generated sequencing result of feature-import, and iterating to obtain better features, wherein table 2 shows 3 finally used flow features related to packet loss rate classification.
TABLE 2 flow characteristics information
| Feature(s) | Description information |
| FD_Pck | Number of packets transmitted by server to client |
| BD_Pck | Number of packets transmitted by client to server |
| F_avg_PL | Average packet length transmitted by server to client |
In an embodiment of the present invention, in the step (5), training the packet loss rate classification model specifically includes the following processes:
(5.1) capturing a large amount of game flow data with different packet loss rates through the step (3), and then performing data processing, feature extraction and optimization through the step (4) to obtain a packet loss rate training sample set, wherein the packet loss rate training sample set comprises game data under normal conditions and three packet loss rates (10%, 20% and 30%) applied, and the total number of the game flow data is 26000 samples;
and (5.2) training the training data set obtained in the step (5) by using an XGboost (eXtreme Gradient Boosting) algorithm to obtain a packet loss rate classification model. The confusion matrix of the model accuracy is shown in fig. 2, the accuracy of the performance index is 0.97, the recall rate is 0.97, and the F1 score is 0.97.
In an embodiment of the present invention, in the step (6), the packet loss rate identification specifically includes the following processes:
(6.1) collecting new game flow by using flow collection equipment;
(6.2) carrying out data processing and extracting the characteristics of game flow through the step (4), and obtaining flow characteristics with partial time length of 10s as shown in a table 3;
TABLE 3 partial flow characteristics
| FD_Pck | BD_Pck | F_avg_PL |
| 506 | 506 | 115 |
| 552 | 554 | 146 |
| 485 | 502 | 183 |
| 487 | 500 | 165 |
| 413 | 475 | 237 |
| 368 | 412 | 198 |
(6.3) inputting the game flow characteristics into the packet loss rate classification model in the step (5) to realize the detection of the packet loss rate, wherein the detection results aiming at the partial flow characteristics in the table 3 are shown in the table 4;
table 4 partial flow characteristic test results
| FD_Pck | BD_Pck | F_avg_PL | Packet loss rate |
| 506 | 506 | 115 | 0％ |
| 552 | 554 | 146 | 0％ |
| 485 | 502 | 183 | 10％ |
| 487 | 500 | 165 | 10％ |
| 413 | 475 | 237 | 30％ |
| 368 | 412 | 198 | 30％ |
In an embodiment of the present invention, in step (7), the delay detection is implemented, which specifically includes the following processes:
and (7.1) counting the length distribution of the payloads of all the packets for the game traffic filtered in the step (6). As shown in fig. 4 (a) and (b), a probability density function graph and an accumulative distribution function graph are drawn according to all payload lengths, and a threshold for dividing a data packet and an acknowledgement packet is found, since the data packet and the acknowledgement packet are in a one-to-one correspondence relationship and the number of the data packet and the acknowledgement packet should be consistent, the threshold should be at the payload length corresponding to the accumulative probability of 0.5 in the accumulative distribution function graph, and the threshold should be 23 in this example. Then, the packet type is marked: if the length of the payload of the packet is greater than the threshold value, recording the packet as a data packet, and if the length of the payload of the packet is less than or equal to the threshold value, recording the packet as an acknowledgement packet; simultaneously recording the transmission direction and the timestamp of each packet according to the source IP and the destination IP;
(7.2) traversing each packet from the head, looking for acknowledgement packets with timestamps within the next 1 second for the first 5 packets marked as packets, and taking the difference between the timestamps of all acknowledgement packets and the timestamp of the packet as the alternative one-sided delay, as shown in fig. 3;
(7.3) for each calculated alternative single-side delay, determining the matching logarithm of the acknowledgement packet and the data packet satisfying the single-side delay behind the data packet corresponding to the single-side delay in the data stream, wherein the determination condition is as follows: if the time difference between the acknowledgment packet and the data packet is within 0.02 seconds of the single-sided delay, then the log of the match is incremented by 1 until all packets have been calculated.
And (7.4) taking the unilateral delay corresponding to the maximum matching logarithm as the optimal unilateral delay. And calculating the ratio of the maximum matching logarithm to the number of all data packets, namely the ratio of the number of packets which accord with the optimal unilateral delay in all packets, and using the ratio as an accuracy performance index of the obtained optimal unilateral delay, wherein the higher the ratio is, the more accurate the optimal unilateral delay obtained by the method is.
(7.5) respectively executing the steps (7.2) - (7.4) from two directions, recording the optimal unilateral delay from the acquisition point to the server side as time1, recording the corresponding matching logarithm as cntmachmax 1, recording the number of all data packets in the direction as cntall1, and recording the accuracy index 1 as ratio1; the optimal unilateral delay from the acquisition point to the client is recorded as time2, the corresponding matching logarithm is cntmatchmax2, the number of all data packets in the direction is recorded as cntall2, and the accuracy index 2 is recorded as ratio2;
(7.6) adding time1 and time2 which are the best unilateral delays in the two directions to obtain tmRTT which is the detection result of the delay, and taking the average value of the accuracy indexes of ratio1 and ratio2 in the two directions to obtain ratio which is the detection accuracy index of the delay. As shown in table 5, the detection calculation process of applying a delay of 100ms (about 18ms for the base delay) for a game flow of 10s is performed.
TABLE 5 calculation procedure for delay detection
In an embodiment of the present invention, in step (8), the packet loss rate level identification result obtained in step (6) and the delay detection result obtained in step (7) are substituted into the QoS-QoE detection model obtained in step (2), so as to detect the QoE of the user game, where a part of the detection results are shown in table 6.
Table 6 results of QoE test for partial games
| Detecting a delay | Detecting packet loss rate | Detecting QoE scores |
| 118.72ms | 0％ | 3.44 |
| 69.38 |
10％ | 3.73 |
| 18.29 |
30％ | 3.07 |
The technical means disclosed in the invention scheme are not limited to the technical means disclosed in the above embodiments, but also include the technical scheme formed by any combination of the above technical features. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and such improvements and modifications are also considered to be within the scope of the present invention.
Claims (9)
1. A QoE detection method for online games based on network traffic is characterized by comprising the following steps:
step (1) carrying out a subjective QoE assessment experiment;
step (2) performing data statistical analysis according to the experimental result obtained in the step (1) to obtain a QoS-QoE detection model;
step (3) carrying out game flow acquisition through data acquisition equipment;
step (4) carrying out data processing, feature extraction and feature optimization on the game flow obtained in the step (3);
step (5) obtaining a large amount of game flow data by using the step (3), processing the data by the step (4) to obtain a packet loss rate training sample set, and training by using a machine learning algorithm to obtain a packet loss rate classification model;
step (6) collecting new game flow, performing data processing and feature extraction through the step (4), and inputting the new game flow into the packet loss rate classification model in the step (5) to identify the packet loss rate grade;
step (7) the game flow statistical data information collected in the step (6) is detected to realize delay;
and (8) substituting the packet loss rate identification result obtained in the step (6) and the delay detection result obtained in the step (7) into the QoS-QoE detection model obtained in the step (2) to detect the QoE of the user game.
2. The method for detecting QoE of online games based on network traffic as claimed in claim 1, wherein in step (1), the method for performing the subjective QoE assessment experiment is as follows:
(1.1) building an experimental environment simulating the real network condition;
and (1.2) finding a plurality of users to play games under various network conditions (different packet loss rates and delays), and acquiring the QoE scores of the players after each game is finished.
3. The method for detecting QoE of online game based on network traffic as claimed in claim 1, wherein in the step (2), the specific process of obtaining the QoS-QoE detection model is as follows:
(2.1) calculating an average QoE value for all players under each network condition;
and (2.2) fitting the QoE scoring data by utilizing nonlinear fitting to obtain a model for QoE detection according to packet loss rate and delay.
4. The method for detecting QoE of online games based on network traffic as claimed in claim 1, wherein in step (3), the method for obtaining the traffic of the game is as follows:
(3.1) capturing game flow data under different network conditions by using tcpdump operated on an OpenWrt softrouter on the basis of the experimental environment built in the step (1);
and (3.2) after each game is finished, storing the flow data of each game.
5. The method for detecting QoE of online game based on network traffic as claimed in claim 1, wherein in the step (4), the data processing and feature extraction specifically comprise the following sub-steps:
(4.1) filtering game flow from the flow obtained in the step (3), and segmenting data;
(4.2) extracting the characteristics of the flow data, and labeling all data samples;
and (4.3) inputting the sample data set into a random forest classifier to iteratively optimize the features.
6. The method for detecting QoE of online games based on network traffic as claimed in claim 1, wherein in the step (5), the training of the packet loss rate classification model specifically includes the following sub-steps:
(5.1) capturing a large amount of game flow data with different packet loss rates through the step (3), and performing data processing, feature extraction and optimization through the step (4) to obtain a packet loss rate training sample set;
and (5.2) training the packet loss rate training sample set by using a machine learning algorithm to obtain a packet loss rate classification model.
7. The method for detecting QoE of online game based on network traffic as claimed in claim 1, wherein in the step (6), the packet loss rate identification specifically comprises the following sub-steps:
(6.1) collecting new game flow by using data collection equipment;
(6.2) carrying out data processing on the acquired flow through the step (4) and extracting characteristics;
and (6.3) inputting the obtained flow characteristic result into the packet loss rate classification model in the step (5) and identifying the packet loss rate grade.
8. The method for detecting QoE of network game based on network traffic as claimed in claim 1, wherein the step (7) of detecting the delay specifically comprises the following sub-steps:
(7.1) dividing the game flow filtered in the step (6) into data packets and confirmation packets according to the payload length distribution of all the packets, and counting the transmission direction and the time stamp of each packet;
(7.2) calculating the difference value of the time stamps of the data packet and the acknowledgement packet as an alternative unilateral delay;
(7.3) for each alternative single-sided delay, counting the matching logarithm of the data packet and the confirmation packet which meet the alternative single-sided delay;
(7.4) taking the unilateral delay corresponding to the maximum matching logarithm as the optimal unilateral delay, and calculating the ratio of the maximum matching logarithm to all data packets to be used as a judgment for the accuracy index of the found optimal unilateral delay;
(7.5) executing the steps (7.2) to (7.4) from the acquisition point to the server side and from the acquisition point to the client side respectively;
and (7.6) adding the optimal unilateral delays in the two directions to obtain a delay result, and taking the average value of the accuracy indexes in the two directions as a delay detection accuracy index.
9. The method for detecting QoE of online games based on network traffic as claimed in claim 1, wherein in step (8), the packet loss rate level identification result obtained in step (6) and the delay detection result obtained in step (7) are substituted into the QoS-QoE detection model obtained in step (2), so as to achieve QoE detection of the user game.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210927727.9A CN115314407A (en) | 2022-08-03 | 2022-08-03 | Network flow based online game QoE detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210927727.9A CN115314407A (en) | 2022-08-03 | 2022-08-03 | Network flow based online game QoE detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115314407A true CN115314407A (en) | 2022-11-08 |
Family
ID=83858989
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210927727.9A Pending CN115314407A (en) | 2022-08-03 | 2022-08-03 | Network flow based online game QoE detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115314407A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116996397A (en) * | 2023-09-27 | 2023-11-03 | 之江实验室 | Network packet loss optimization method and device, storage medium and electronic equipment |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004236263A (en) * | 2003-02-03 | 2004-08-19 | Nippon Telegr & Teleph Corp <Ntt> | Packet communication quality measurement method and delay calculation apparatus |
| CN101359944A (en) * | 2006-10-31 | 2009-02-04 | 株式会社Ntt都科摩 | Method, system, base station and user equipment determining cyclic delay and diversity delay |
| US20100273558A1 (en) * | 2007-10-03 | 2010-10-28 | Telefonaktiebolaget Lm Ericsson (Publ) | Method of Determining Quality of Service for On-Line Gaming in a Network |
| US20130148525A1 (en) * | 2010-05-14 | 2013-06-13 | Telefonica, S.A. | Method for calculating perception of the user experience of the quality of monitored integrated telecommunications operator services |
| US20140201330A1 (en) * | 2011-04-05 | 2014-07-17 | Telefonica, S.A. | Method and device for quality measuring of streaming media services |
| CN104113788A (en) * | 2014-07-09 | 2014-10-22 | 北京邮电大学 | QoE training and assessment method and system of TCP video stream service |
| CN109873797A (en) * | 2018-02-14 | 2019-06-11 | 南京邮电大学 | Conversational video business QoE-QoS parameter mapping method based on statistical analysis |
| CN110913414A (en) * | 2019-12-19 | 2020-03-24 | 中国卫通集团股份有限公司 | QoS guarantee system of 5G satellite convergence network |
| CN113055251A (en) * | 2021-04-17 | 2021-06-29 | 东南大学 | High-speed network-oriented real-time sensing method for flow packet loss state |
-
2022
- 2022-08-03 CN CN202210927727.9A patent/CN115314407A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004236263A (en) * | 2003-02-03 | 2004-08-19 | Nippon Telegr & Teleph Corp <Ntt> | Packet communication quality measurement method and delay calculation apparatus |
| CN101359944A (en) * | 2006-10-31 | 2009-02-04 | 株式会社Ntt都科摩 | Method, system, base station and user equipment determining cyclic delay and diversity delay |
| US20100273558A1 (en) * | 2007-10-03 | 2010-10-28 | Telefonaktiebolaget Lm Ericsson (Publ) | Method of Determining Quality of Service for On-Line Gaming in a Network |
| US20130148525A1 (en) * | 2010-05-14 | 2013-06-13 | Telefonica, S.A. | Method for calculating perception of the user experience of the quality of monitored integrated telecommunications operator services |
| US20140201330A1 (en) * | 2011-04-05 | 2014-07-17 | Telefonica, S.A. | Method and device for quality measuring of streaming media services |
| CN104113788A (en) * | 2014-07-09 | 2014-10-22 | 北京邮电大学 | QoE training and assessment method and system of TCP video stream service |
| CN109873797A (en) * | 2018-02-14 | 2019-06-11 | 南京邮电大学 | Conversational video business QoE-QoS parameter mapping method based on statistical analysis |
| CN110913414A (en) * | 2019-12-19 | 2020-03-24 | 中国卫通集团股份有限公司 | QoS guarantee system of 5G satellite convergence network |
| CN113055251A (en) * | 2021-04-17 | 2021-06-29 | 东南大学 | High-speed network-oriented real-time sensing method for flow packet loss state |
Non-Patent Citations (1)
| Title |
|---|
| 吴玉峰;李建;阮雅端;陈启美;: "基于BP神经网络QoS到QoE映射模型", 电子测量技术, no. 01, pages 84 - 87 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116996397A (en) * | 2023-09-27 | 2023-11-03 | 之江实验室 | Network packet loss optimization method and device, storage medium and electronic equipment |
| CN116996397B (en) * | 2023-09-27 | 2024-01-09 | 之江实验室 | Network packet loss optimization method and device, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Bujlow et al. | A method for classification of network traffic based on C5. 0 Machine Learning Algorithm | |
| CN103840988B (en) | A kind of network flow programming method method based on RBF neural | |
| CN109714343B (en) | Method and device for judging network traffic abnormity | |
| JP4553315B2 (en) | Congestion path classification method, management apparatus, and program for classifying congestion path from packet delay | |
| EP1727375A1 (en) | Assessment of perceived quality of a packetized video stream | |
| CN106992974B (en) | Live video information monitoring method, device and equipment | |
| CN102904773B (en) | Method and device for measuring network service quality | |
| CN107483267B (en) | EIGRP route fault identification method | |
| CN115314407A (en) | Network flow based online game QoE detection method | |
| CN110048962A (en) | A kind of method of net flow assorted, system and equipment | |
| CN106330611A (en) | Anonymous protocol classification method based on statistical feature classification | |
| CN114866485B (en) | Network traffic classification method and classification system based on aggregation entropy | |
| Manzoor et al. | How HTTP/2 is changing web traffic and how to detect it | |
| Oršolić et al. | In-network qoe and kpi monitoring of mobile youtube traffic: Insights for encrypted ios flows | |
| CN104657747A (en) | Online game stream classifying method based on statistical characteristics | |
| CN113645182B (en) | Denial of service attack random forest detection method based on secondary feature screening | |
| CN111600878A (en) | Low-rate denial of service attack detection method based on MAF-ADM | |
| CN111478921A (en) | Method, device and equipment for detecting communication of hidden channel | |
| JP2008042879A (en) | Congestion path classification method to classify congestion path based on packet delay fluctuation, management apparatus and program | |
| Jaber et al. | Using host profiling to refine statistical application identification | |
| Giorgi et al. | Detection of anomalous behaviors in networks from traffic measurements | |
| CN112235254A (en) | Rapid identification method for Tor network bridge in high-speed backbone network | |
| CN108646688B (en) | A kind of process parameter optimizing analysis method based on recurrence learning | |
| Gupta et al. | Revisiting coexistence of poissonity and self-similarity in internet traffic | |
| CN115174961A (en) | Multi-platform video flow early identification method facing high-speed network |
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 |