RU2562772C1 - Method of measuring information (data) transmission speed with broadband internet access - Google Patents

Abstract

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to means of measuring information transmission speed with broadband Internet access. Disclosed is a method of measuring information transmission speed with broadband Internet access on a Client-Server network architecture, wherein when measuring information transmission speed on a forward and a return channel, which link the measuring means and service user equipment, a measurement sequence L is transmitted, said sequence consisting of N packets selected from a set M in a random manner in accordance with a given distribution law; current time values are recorded for the beginning and end of transmitting each packet at the output and input of the measuring means and the input and output of the user equipment; after transmitting the measurement sequence of packets, the recorded current time values are transmitted to the measuring means, where the information transmission speed is determined as the ratio of the obtained volume of information to the reception time thereof.

EFFECT: providing given accuracy of measuring information transmission speed and providing repeatability of measurement results.

9 cl, 1 dwg

Description

Technical field

The invention relates to data transmission networks (technical field H04L 12/00 in international classification); to maintenance and management technologies (H04L 12/24); to monitoring and testing devices (H04L 12/26); to packet switching systems (H04L 12/56).

BACKGROUND OF THE INVENTION

Currently, there are many different ways to determine the transmission speed when accessing the Internet. However, there is not one among them that can withstand metrological verification and find legal suitability for resolving conflicts between the telecom operator and the consumer. The proposed metrologically sound method for measuring the speed of broadband Internet access is designed to fill this gap.

The claimed invention is aimed at achieving the following technical result: ensuring the specified accuracy of measuring the data transmission speed, ensuring the repeatability of measurement results, ensuring the modeling of the nature of telecommunication traffic, ensuring uniformity of measurements by the Internet provider and user, ensuring the highest quality of telecommunication services to the consumer, ensuring reducing the impact of the network topology on accuracy measurements, providing the opportunity to study Dean iki changing the quality of the telecommunications service, providing simultaneous measurement of transmission quality (bit-Error Probability).

The following abbreviations and English terms may be used in the description and / or drawings:

Similar measurement methods

One of the most popular applications for assessing Internet access speed is Speedtest software [1]. In the Play market app store, it has over 10 million downloads for devices running the Android operating system. The functionality of the application allows you to evaluate the downstream and downstream data transfer rates, as well as the delay of packets using the analogue of the ping program, which operates with 64 byte packets for checking connectivity in data networks based on the TCP / IP protocol stack.

Speedtest [2] estimates the speed of information transfer using the HTTP protocol, the transmission of which is allowed by most network devices. In this case, Speedtest operates with three subroutines:

- HTTP ping - measures the transmission delay of a small packet (containing only the text "test") between the client equipment and the selected server;

- Download and Upload - evaluate the speed of digital streams between the server and the client (downstream) and between the client and the server (upstream).

The information transfer rate between the server and the client is estimated as follows:

- The client computer downloads binary files of small volume (a few ten kilobytes) from the selected server. The parameters of these downloads are measured to estimate the transmission speed in this direction. The measurement time is 10 seconds.

- Binary data is generated randomly to exclude the effect of caching on the result of the evaluation.

- To increase the correctness of the measurement result, file transfer is carried out in several (from two to four) streams.

- The results of the transmission of packets containing fragments of binary files are recorded up to 30 times per second.

- The results are aggregated in 20 assemblies (each - 5% of the total sample).

- 10% of assemblies (2 assemblies) with the highest speed and 30% (6 assemblies) with the lowest are discarded and are not taken into account when calculating the result.

- The result of calculating the speed for the remaining assemblies is averaged and presented as the final speed of information transfer.

The following random factors influence the measurement result: the amount of related overhead information, the method of data buffering, the CPU load level. These factors make it necessary to drop 10% of the results at the highest and lowest speeds.

Additionally, 20% of the results with the lowest speed are discarded in order to exclude the values measured from the start of the test to the highest speed.

The information transfer rate between the client and server is estimated as follows:

- The client computer sends small-sized binary files (several tens of kilobytes) to the server. The parameters of these downloads are measured to estimate the transmission speed in this direction.

- Binary data is randomly generated.

- A special code is launched on the client terminal, which transfers files using the POST method of the HTTP protocol.

- To increase the correctness of the measurement result, file transfer is carried out in several (from two to four) streams.

- The result of the assessment is the average value of half the results at the highest speed.

In a similar way, a mechanism for evaluating the speed of information transfer for analogues of Speedtest software is organized [3-13]. In most well-known methods, measurements are made: the speed of receiving and transmitting information, as well as the signal delay at the application level of the interaction model of open systems.

Fundamental disadvantages of known methods for measuring the speed of information transfer.

1. Current software versions of Speedtest analogues are available on data networks of large operators. Therefore, speed estimation is usually performed only within the operator’s data network or to the interface between operators.

2. In general, the results of evaluating Speedtest analogs may differ from each other and from the results of other methods implemented on servers outside the operator's data network and / or using a different methodology for processing measurement results.

3. The following factors may influence the results of measuring the speed of information transfer:

- load on the server and user terminal;

- hardware configuration and user terminal performance;

- the composition of the software (software) installed and / or working on the user terminal at the time of measurement;

- the state of the access network and user line during the measurement period;

- the use of security tools (firewalls, firewalls, antiviruses) on the network and the subscriber's terminal at the time of measurement.

4. The time for measuring speed towards the subscriber is limited to 10 seconds. To obtain a “reliable” result, it is recommended to make several assessment attempts at different times. The period of the day is not specified.

5. The evaluation uses an arbitrary measurement signal.

6. Processing of measurement results includes discarding up to 40% of the obtained values, which excludes the possibility of determining and normalizing the error of the method.

Prototype of the invention

The closest way to measure the speed of information transmission is proposed in US patent 8,228,815, Equipment and methods for calculating the characteristics of data transmission over a network connection based on one-way measurements [12]. Here, a calculation method is provided, including: transmitting a packet hook containing at least one first TCP packet of a first type at the head end of the packet hook; at least one second TCP packet of the first type at the tail end of the packet hook, and a plurality of second type TCP packets located between the head and tail of the packet hook. In this method, each TCP packet in a hitch is addressed to a free port of a neighboring router via a network connection in which TCP packets of the first type initiate a router response, but packets of the second type do not. The received responses from each router along the network path at the head end and at the tail end of the packet concatenation are used to calculate at least one characteristic of the data flow through the network connection. The method comprises the steps of: determining a delay time between responses received from each router; determining an interval between packets for each router based on the corresponding delay time of the router; Calculation of the network connection throughput depending on the packet link length and the maximum packet interval for each router.

This method is not free from the disadvantages noted above, and does not allow to reliably measure the speed of information transfer as the main indicator of the quality of telecommunication services.

SUMMARY OF THE INVENTION

The method is implemented with the following initial positions:

1. Architecture is used for measurements: a measuring instrument is user equipment connected by communication channels. As an example of a measuring instrument, a telecommunication server equipped with specialized software can be considered, and a client computer as a user equipment. In terms of telecommunications, this architecture is called Client - Server. Further in the description of the invention we will use this example. Measurement is carried out by transferring data streams in the forward and reverse directions between the server and the client.

2. The server is a hardware-software complex and is a means of measurement with metrological characteristics (time stamp error +/- 1 μs, frame length error +/- 1 byte).

3. Data packets for measuring the speed and quality of transmission when accessing the Internet are formed in the TCP / IP stack and use the following Ethernet, IP and TCP protocols, as well as the application layer protocol (browser protocol) - HTTP. The standard HTTP protocol port, 80, is used to transfer control and test data. The format of the data packet is shown in the table:

4. Measuring equipment - servers are located at Internet traffic exchange points. The server for processing measurement results is located in the Center for Quality Control of Telecommunication Services.

5. The measuring signal is a reference amount of information in the form of a segment of a digital random sequence of binary pulses formed from "white noise". This sequence is divided into packets of length 1514, 1280, 1024, 512, 256, 128 or 64 bytes, and the probability of each of which corresponds to a predetermined distribution law. Data on the total reference amount of information in the measuring signal is entered into each packet.

Preliminary Actions:

1. Before measurement, the Client and Server synchronize the internal clock using the NTP protocol.

2. The client initiates a connection to the Server via HTTP.

3. On the Server, a web page is opened, which is transmitted to the Client, and a script is run with commands for taking measurements on the part of the Client.

4. Server and Client exchange packets with a data field in the format:

where Magic Number (16 bits) is a unique protocol identifier (COFE); Sequence Number (16 bits) - serial number of the packet; Quantity of information (32 bits) - the total amount of information in the measuring signal; NTP Time Stamp (64 bit) - time by world clock; Data - data up to 1386 bytes (the corresponding part of the measuring signal is a random digital sequence).

5. The client, as a result of the script, creates a table in which for each packet are entered: packet number (N), packet length (L), packet sending time (T _S ) and packet reception time (T _E ). The amount of information in the measuring signal is also entered in the table. The table is transferred to the server.

6. The client opens a web page that is transmitted to the Server and runs a script with commands for taking measurements from the Server. Having made similar measurements, the Server forms its own Table, in which the same measurement data are entered. Tables can be stored in the Quality Control Center on a separate server, which accumulates the database of measurement results.

7. The server performs data processing to determine the information transfer rate (Internet access speed) for the downstream and upstream streams V _C and V _S as the ratio of the total amount of information to the time of reception of the measuring signal, as well as the fidelity of transmission as the ratio of correctly received characters to their total quantity in the measuring signal.

Detailing the measurement method. The method of measuring information transfer rate (SPI) is to perform the following operations:

A. to measure the STI on the direct and then on the return channels connecting the measuring instrument and the computer of the service user (consumer of information), a measuring sequence L consisting of N packets selected from a set of M packets randomly in accordance with a given distribution law is transmitted ;

B. at the following four points of the communication network, Figure 1: - output of the measuring instrument (point 1), - input of the equipment (computer) of the user (point 2), - output of the equipment (computer) of the user (point 3), - input of the measuring instrument ( point 4), - the current time values corresponding to the implementation of given events and the total reference amount of information Q _m are recorded.

Specification of operation B. Registered:

- the beginning of the transfer of each packet from the measuring instrument to the user equipment (at point 1),

- the end of the transmission of each packet from the measuring instrument to the user equipment (at point 1),

- the beginning of the reception of each packet by the user equipment (at point 2),

- the end of the reception of each packet by the user equipment (at point 2),

- the beginning of the transfer of each packet from the user equipment to the measuring instrument (at point 3),

- the end of the transmission of each packet from the user equipment to the measuring instrument (at point 3),

- the beginning of the reception of each packet by the measuring instrument (at point 4),

- the end of the reception of each packet by the measuring instrument (at point 4).

C. The recorded values of the current time are transmitted after the transmission of the measuring sequence. This ensures a minimum value of delays in the transmission of the measuring sequence and minimizes the systematic component of the error in the measurement of time intervals.

D. after passing the entire measuring sequence through point 4 (see Fig. 1), for each of the measuring sequence packets, the following types of recorded information are available:

1) number i = {1, 2, ..., N} of the packet,

2) the amount I _{Wi of} information resulting from the analog-to-digital conversion of the “white noise” implementation and contained in the packet with number i;

3) the value of the current time T _B1i corresponding to the beginning (index B) of the transmission of the packet with number i, registered by the clock of the computer located at point 1;

4) the value of the current time T _E1t , corresponding to the end (index E) of the transmission of the packet with number i, registered by the clock of the computer located at point 1;

5) the value of the current time T _B2i corresponding to the beginning of receiving a packet with number i, registered by the clock of the computer located at point 2;

6) the value of the current time T _E2t , corresponding to the end of receiving a packet with number i, registered by the clock of the computer located at point 2;

7) the value of the current time T _B3i , corresponding to the beginning of the transmission of the packet with number i, registered by the clock of the computer located at point 3;

8) the value of the current time T _E3t , corresponding to the end of the transmission of the packet with number i, registered by the clock of the computer located at point 1;

9) the value of the current time T _B4i corresponding to the beginning of receiving a packet with number i, registered by the clock of the computer located at point 4;

10) the value of the current time T _E4t , corresponding to the end of receiving a packet with number i, registered by the clock of the computer located at point 4;

11) the total amount I _{W of} information resulting from the analog-to-digital conversion of the “white noise” implementation and contained in packets from number 1 to number N at point 4.

The proposed method for measuring the speed of broadband Internet access has metrological consistency, sufficient measurement accuracy, provides repeatability of results, consistency with the protocols and interfaces used, and natural conjugation with the quality assessment of this Internet provider.

Literature

Claims (9)

1. A method for measuring the information transfer rate (SPI) for broadband Internet access using the Client-Server network architecture, which provides for the following operations:
A. for measurements of STI on the direct and then return channels connecting the measuring instrument and the user equipment of the service, transmit a measurement sequence L consisting of N packets selected from a plurality of M packets at random in accordance with a given distribution law;
B. register the values of the current time: at the output of the measuring instrument, the beginning and end of transmission of each packet from the measuring instrument to the user equipment is recorded, at the input of the user equipment, the beginning and end of receiving each packet is recorded by the user equipment, at the output of the user equipment, the beginning and end of transmission of each a packet from user equipment to a measuring instrument; at the input of a measuring instrument, the beginning and end of receiving each packet are recorded;
B. after transmitting the measuring sequence, the recorded values of the current time are transmitted to the measuring means;
D. after passing the entire measuring sequence through the input of the measuring instrument for each of the packets of the measuring sequence, the following types of registered information are saved: number i = {1, 2, ..., N} of the packet; the amount I _{Wi of} information resulting from the analog-to-digital conversion of the “white noise” implementation and contained in the packet with number i; the value of the current time T _B1i corresponding to the beginning (index B) of the transmission of the packet with number i, registered by the clock of the measuring instrument; the value of the current time T _E1t , corresponding to the end (index E) of the transmission of the packet with number i, registered by the clock of the measuring instrument; the value of the current time T _B2i corresponding to the beginning of receiving a packet with number i, recorded by the user equipment clock; the value of the current time T _E2t , corresponding to the end of receiving a packet with number i, registered by the user equipment clock; the value of the current time T _B3i corresponding to the beginning of the transmission of the packet with the number i, registered by the hours of the user equipment; the value of the current time T _E3t , corresponding to the end of the transmission of the packet with number i, registered by the hours of the user equipment; the value of the current time T _B4i corresponding to the beginning of receiving a packet with number i, registered by the clock of the measuring instrument; the value of the current time T _E4t , corresponding to the end of receiving a packet with number i, registered by the clock of the measuring instrument; the total amount I _{W of} information resulting from the analog-to-digital conversion of the implementation of “white noise” and contained in packets with numbers from 1 to N at the input of the measuring instrument;
D. Based on the available data, the measuring tool determines the information transfer rate as the ratio of the amount of information received to the time it was received. 2. A method for measuring the information transfer rate (SPI) for broadband Internet access according to claim 1, characterized in that for ensuring the repeatability of the measurement results, a test measurement signal not compressible by network protocols is used as a segment of a random sequence of binary symbols generated by “white noise” . 3. The method of measuring the information transfer rate (STI) for broadband Internet access according to claim 1, characterized in that in order to simulate the nature of telecommunication traffic, a random sequence of binary symbols is divided into packets of different lengths, and the probability of a packet of a certain length corresponding to a predetermined law distribution. 4. The method of measuring the speed of information transfer (STI) for broadband Internet access according to claim 1, characterized in that to ensure the legal consistency of the method, all users and all Internet providers test using the same measuring signal with a predetermined amount of information . 5. The method of measuring the speed of information transfer (STI) for broadband Internet access according to claim 1, characterized in that the measurements are carried out at the busy hour (NNI) to ensure the highest quality telecommunication services to the consumer. 6. The method of measuring the information transfer rate (STI) for broadband Internet access according to claim 1, characterized in that to reduce the influence of the network topology on the measurement accuracy, the measuring instruments are placed at the point of highest Internet traffic concentration in the region. 7. The method of measuring the speed of information transfer (STI) for broadband Internet access according to claim 1, characterized in that in order to provide the possibility of studying the dynamics of changes in the quality of a telecommunication service, the processing and storage of measurement results is performed in a single center for monitoring the quality of telecommunication services. 8. The method of measuring the information transfer rate (STI) for broadband Internet access according to claim 1, characterized in that for the simultaneous measurement of the quality of information transmission, the fidelity of the transmission is controlled by bitwise comparison of the received signal with the reference measuring signal. 9. The method of measuring the information transfer speed (SPI) for broadband Internet access according to claim 2, characterized in that to ensure the specified measurement accuracy for the “white noise” generator and the ADC, parameters are defined that determine the error of the analog-to-digital conversion: frequency band, peak factor and resolution.