CN110691007A

CN110691007A - Method for accurately measuring QUIC connection packet loss rate

Info

Publication number: CN110691007A
Application number: CN201910885435.1A
Authority: CN
Inventors: 吴桦; 王路; 程光
Original assignee: Southeast University
Current assignee: Southeast University
Priority date: 2019-09-19
Filing date: 2019-09-19
Publication date: 2020-01-14

Abstract

The invention discloses a method for accurately measuring the packet loss rate of QUIC connection. The method comprises the steps of firstly collecting data message files on network equipment, then filtering UDP messages from the data message files and analyzing the heads of the UDP messages to distinguish QUIC protocol messages, then analyzing the heads of the QUIC data messages and reforming PKN message serial numbers, and finally counting and calculating the packet loss rate of QUIC connection according to the recorded message serial numbers. The invention has universality, and all applications using QUIC protocol for data transmission are applicable to the method for calculating the packet loss rate.

Description

Method for accurately measuring QUIC connection packet loss rate

Technical Field

The invention belongs to the technical field of network measurement, and particularly relates to a method for accurately measuring the QUIC connection packet loss rate.

Background

In the Internet architecture, the transport layer is one of the key layers and is responsible for providing communication services between two processes in the host. The two most common protocols in the transport layer are the transmission control protocol tcp (transmission control protocol) and the user Datagram protocol udp (user data protocol). The TCP protocol is connection-oriented and provides stable and reliable communication services from end to end, but the time overhead for establishing and maintaining a connection is high. The UDP protocol is a connectionless transport layer protocol that provides transaction-oriented simple unreliable messaging service, does not require the establishment and maintenance of connections, and therefore has little time overhead. Compared with TCP, UDP has higher transmission efficiency and lower overhead, but cannot ensure the reliability of data transmission.

QUIC (Quick UDP-Internet Connection) is a new type of UDP-based secure multiplex transmission protocol. QUIC is built with HTTP2.0 as the main application protocol and implements some new mechanisms and features, making it an attractive modern general-purpose transport protocol. Compared with the current mainstream TCP protocol, the QUIC adds some new features, mainly in allowing multiple transmissions on the same transport layer connection, enforcing encryption to improve security, and smaller handshake delay and re-establishment delay. QUIC provides multiplexing and flow control equivalent to the HTTP2.0 protocol, security equivalent to TLS, and connection semantics, reliability, and congestion control equivalent to the TCP protocol. With the continuous development of the internet, data security and privacy protection are more and more emphasized by manufacturers and users, and QUICs have been widely used on the internet.

ISP providers need to monitor the quality of communication services to improve the network environment of users and provide better services to users. The measurement of the packet loss rate is a key ring for measuring the network service quality. The packet loss rate is a ratio of the number of lost data packets to the number of transmitted data packets, and reflects the stability of the data transmission capability of the network, and a lower ratio represents a higher stability.

The measurement of the packet loss rate is divided into two methods, namely active measurement and passive measurement. The active measurement requires a tester to actively send a test message and count results, which are different from the data transmission distribution of real application, and because the protocol of the test message is different from the protocol used by user application, the obtained results are often inconsistent with the actual data of the user. The passive measurement is to collect the data message of the user at the intermediate node, and to analyze the data message to obtain the packet loss rate of the application, and the packet loss rate obtained by the passive measurement method can reflect the real packet loss rate of the user.

In the existing solution, the packet loss rate of the transmission path is estimated by passively collected traffic data, and the UDP protocol cannot count the packet loss rate because of no sequence number. QUIC is established on UDP protocol, and has no sequence number of TCP protocol, so the existing method for estimating the packet loss rate by passively collecting TCP data cannot be applied.

The invention utilizes the PKN serial Number of the Packet Number field of the QUIC protocol, allows data to be collected at any node in a path passed by the QUIC connection and counts Packet loss rate information, and has wider application range. On the other hand, only the packet loss rate of a single flow can be counted based on the TCP serial number information, and the QUIC connection is oriented to multiple flows, so that the packet loss rates of multiple flows in the QUIC connection can be counted, the packet loss rate of one flow can be counted independently, and the achievable function is stronger.

Disclosure of Invention

The purpose of the invention is as follows: aiming at the problems, the invention discloses a method for accurately measuring the QUIC connection packet loss rate. The method comprises the steps of firstly collecting data on equipment, then filtering a UDP message from the equipment, analyzing the head of the UDP message to distinguish a QUIC protocol message, then analyzing the head of a QUIC data packet, reforming a PKN message serial number, and finally calculating the packet loss rate of QUIC connection according to the recorded message serial number. The invention has universality, and all applications using QUIC protocol for data transmission are applicable to the method for calculating the packet loss rate.

The technical scheme is as follows: in order to realize the purpose of the invention, the technical scheme adopted by the invention is as follows: a method for accurately measuring the packet loss rate of QUIC connection specifically comprises the following steps:

(1) acquiring QUIC (quantum data integration) interaction flow data of a client and a server on equipment such as a mobile phone, a PC (personal computer), a router and the like which can be accessed to a network, and storing the data as a message data file;

(2) filtering out a UDP message from the acquired message data file, analyzing the head of the UDP message, and filtering out a QUIC protocol message;

(3) on the basis of the step (2), analyzing the filtered QUIC data message header, extracting a PKN value, (Packet Number, message serial Number), and then recording the values of the client-side and server-side QUIC message serial Number PKN in an array;

(4) and (4) calculating the packet loss rate of the QUIC connection according to the record array in the step (3) and the maximum PKN statistic of the client and the server.

Further, in step (1), the method for acquiring QUIC data on the device is as follows:

(1.1) accessing the acquisition equipment to a network, and running flow data capturing software on the equipment;

and (1.2) collecting flow data and storing the flow data as a message file.

Further, in the step (2), the UDP packet is filtered from the collected data traffic, the header of the UDP packet is analyzed, and then the method of filtering out the QUIC protocol packet is as follows:

(2.1) setting an IP _ addresses array for storing a specified IP address pair;

the IP address pair is a pair of IPs set by the observer based on the service address and the user address of the application observed in advance, and is set to [10.100.34.28,203.210.8.140] in an example of YouTube video service using QUIC transmission.

(2.2) setting an array Receive [ ], and initializing the array initial value to 0, wherein the value can be 0 or 1.

The value of the Receive [ i ] is 0 to represent that the message which is sent by the server and has PKN of i +1 is not sent or the message is not received by the client, and the value of the Receive [ i ] is 1 to represent that the message which is sent by the server and has PKN of i +1 is sent and the message is received by the client;

(2.3) pointing to a first message of the data message file;

(2.4) reading a current message, judging whether the IP address pair of the client and the server recorded in the message is consistent with that stored in the IP _ addresses array or not, wherein the message is a UDP message, if so, entering (2.5), and otherwise, entering (2.6);

(2.5) judging the UDP message, if a QUIC protocol message is encapsulated in the UDP, recording the QUIC protocol message, otherwise, entering (2.6);

(2.6) if the subsequent message is not read, reading the next message and entering (2.4); otherwise, the message is completely processed.

Further, in the step (3), on the basis of the step (2), the filtered QUIC data message header is analyzed, and the concrete method of recording the client and server QUIC message sequence number PKN values in the array is as follows:

(3.1) reading the first recorded QUIC data message;

(3.2) the first byte of the data message is a Public Flags field, and a Flag parameter contained in the header of the QUIC message is read through the Public Flags field;

(3.3) judging the offset and the length value of the PKN from the Flag parameter on the basis of the step (3.2), and extracting the value of the PKN;

(3.4) for the current QUIC data message, according to the value k of the PKN extracted in (3.3), if the message is sent to the client side from the server side, making the value of Receive [ k-1] be 1, and updating the value of Receive _ max to be the value k;

(3.5) judging whether a next recorded QUIC data message still exists, if so, continuing to read the next QUIC data message, and returning to the step (3.2); otherwise, the recorded QUIC data message is completely processed.

Further, in step (4), the method for calculating the packet loss rate of the QUIC connection according to the record array in step (3) and the maximum PKN statistic value received by the client is as follows:

(4.1) traversing the Receive array from subscript 0, and recording the number of values 1 in the Receive array as Receive _ Num;

(4.2) calculating packet loss rate P as (1-Receive _ Num/Receive _ max) × 100% from the results of (3.3) and (4.1).

Has the advantages that: compared with the prior art, the technical scheme of the invention has the following beneficial technical effects:

(1) the existing method for calculating the path packet loss rate through passive measurement mostly aims at the message flow of a TCP (transmission control protocol), the calculation method is not suitable for a new QUIC (fast QuIC) protocol any more, and the problem of calculating the packet loss rate through the data of the QUIC protocol is solved.

(2) The invention has universality, and the application of using the QUIC protocol to carry out data transmission is all applicable to calculating the packet loss rate by the method.

Drawings

FIG. 1 is a schematic diagram of a system configuration of the method of the present invention;

FIG. 2 is a process of encapsulating application layer data into UDP messages via QUIC protocol and UDP protocol;

FIG. 3 is a general flow diagram for filtering and processing all messages belonging to the same QUIC connection from a data file.

FIG. 4 is a flow chart of parsing the QUIC message header and reforming the PKN sequence number.

FIG. 5 is a flowchart of calculating the packet loss rate of QUIC connection according to the record array and the maximum PKN statistics of the client and the server.

Detailed Description

The technical solution of the present invention is further described below with reference to the accompanying drawings and examples.

The invention provides a method for accurately measuring the packet loss rate of QUIC connection, which specifically comprises the following steps:

In one embodiment of the method of the present invention, in step (1), the method for collecting QUIC data on a device is as follows:

and (1.2) collecting flow data and storing the flow data as a message file.

In an embodiment of the method of the present invention, in step (2), the method of filtering out the UDP packet from the acquired data traffic, parsing the header of the UDP packet, and then filtering out the QUIC protocol packet is as follows:

(2.1) setting an IP _ addresses array for storing a specified IP address pair;

(2.3) pointing to a first message of the data message file;

In an example of the method of the present invention, in the step (3), on the basis of the step (2), the filtered QUIC data message header is analyzed, and the concrete method of recording the client and server QUIC message sequence number PKN values in the array is as follows:

(3.1) reading the first recorded QUIC data message;

the first byte of the QUIC message header is a Public Flags field, wherein flag bits such as version number, set identifier, CID (connection number) length, PKN length and the like are recorded, and the initial position and the length of a PKN value can be determined according to the flag bit information, so that the PKN value is extracted.

In the method, in the embodiment of the QUIC data message collected near the user endpoint, the obtained message information and the PKN value after filtering the QUIC message are shown in the following table:

QUIC message sequence number	Message receiver	PKN field length (unit: Bytes)	PKN
				1	Sever	1	1
2	Sever	1	2
				3	Sever	1	3
4	Client	1	1
				5	Client	1	2
6	Client	1	3
				7	Sever	1	4
8	Sever	1	5
				9	Client	1	4
10	Sever	1	6
				11	Sever	1	7
12	Client	1	5
				13	Sever	1	8
14	Client	1	6
				15	Client	1	7
16	Client	1	8
				17	Sever	1	9
18	Client	1	9
				19	Client	1	10
20	Client	1	11

In the embodiment of the acquired QUIC data message, the process of sequentially processing the filtered QUIC message is as follows:

in an example of the method of the present invention, in step (4), the method for calculating the packet loss rate of the QUIC connection according to the record array in step (3) and the maximum PKN statistic value received by the client is as follows:

(4.2) calculating packet loss rate P ═ 100% of (1-Receive _ Num/Receive _ max) from the results of (3.3) and (4.1);

in the method, in the example of the acquired QUIC data message, the step and the result of calculating the QUIC packet loss rate are as follows:

receive _ Num is 11; recieve _ Max is 11; therefore, the packet loss rate P is (1-11/11) × 100% ═ 0%.

The above examples are only preferred embodiments of the present invention, it should be noted that: it will be apparent to those skilled in the art that various modifications and equivalents can be made without departing from the spirit of the invention, and it is intended that all such modifications and equivalents fall within the scope of the invention as defined in the claims.

Claims

1. A method for accurately measuring the packet loss rate of QUIC connection is characterized by comprising the following steps:

(1) acquiring QUIC (quantum data integration) interaction flow data of a client and a server on equipment which can access a network, and storing the data as a message data file;

(3) on the basis of the step (2), analyzing the filtered QUIC data message head, extracting a PKN value, and recording the values of client and server QUIC message sequence numbers PKN in an array;

(4) and (4) calculating the packet loss rate of the QUIC connection according to the PKN values recorded by the recording array in the step (3) and the maximum PKN statistic of the client and the server.

2. The method for accurately measuring the packet loss rate of QUIC connection according to claim 1, wherein the data collection method in step (1) is as follows:

(1.1) accessing acquisition equipment to a network, and running flow data acquisition software on the equipment;

and (1.2) collecting flow data and storing the flow data as a message file.

3. The method for accurately measuring the QUIC connection packet loss rate according to claim 1 or 2, wherein in the step (2), the UDP messages are filtered from the collected message data file, the UDP message headers are analyzed, and then the QUIC protocol messages are filtered out by the following method:

(2.1) setting an IP _ addresses array for storing an IP address pair of a specified service address and a user address;

(2.2) setting an array Receive, and initializing an array initial value to be 0;

(2.3) pointing to a first message of the data message file;

4. The method for accurately measuring the QUIC connection packet loss rate according to claim 3, wherein in the step (3), on the basis of the step (2), the filtered QUIC data message header is analyzed, the PKN value is extracted, and the values of the client-side and server-side QUIC message sequence numbers PKN are recorded in an array, and the specific method is as follows:

(3.1) reading the first recorded QUIC data message;

5. The method for accurately measuring the packet loss rate of the QUIC connection according to claim 4, wherein in the step (4), the packet loss rate of the QUIC connection is calculated according to the record array in the step (3) and the maximum PKN statistic of the client and the server as follows:

(4.1) traversing the Receive array, counting the number of values 1 in the Receive array, and recording as Receive _ Num;

(4.2) calculating the packet loss rate P ═ 100% of (1-Receive _ Num/Receive _ max).