Fractal telecommunication traffic transmission quality increasing method

The invention relates to the field of flow control or congestion control over communication channels and can be used to provide quality of service (QoS) requirements for the transmission of fractal telecommunications IP traffic.

The ability of this technical solution to work with all types of traffic regardless of the transmission medium used allows it to be used in any packet-based IP data transmission systems.

The 21st century is a century of widespread informatization of all spheres of human activity. Information is the property of all mankind and the new productive force.

The main difference between next-generation networks (NGN) and traditional networks is that all information circulating in the network is divided into two components. This is signaling information that provides for the switching of subscribers and the provision of services, and directly user data containing the payload intended for the subscriber (voice, video, data). Paths of signaling messages and user load may not match. The main devices that organize the route, the sending order, the priority of IP packets and, therefore, ensure the fulfillment of the requirements for quality of service (QoS) are switches and routers.

Currently, the existing methods and technical solutions do not fully take into account the heterogeneity of the transmitted information (services provided), i.e. multicomponentity and packet traffic structure.

A feature of fractal traffic is the presence of stable clustering. For example, in traffic, which is subject to the law of Poisson distribution, clustering exists on a short-term scale, and on a long-term it is smoothed. As a result, if the traffic is Poisson, the queues appearing during the ripple will be cleared in the short term. But as studies of traffic routes show, the uneven behavior of the network itself is uneven, that is, a grouping of pulsations occurs. In such cases, the intensity of the packets cannot be considered subordinate to the Poisson law. The resulting long-term overloads can have a very negative impact on the performance of the data network.

This necessitates the use of technical solutions capable of taking into account the fractal nature of traffic in the operation of telecommunication networks when providing quality service.

The closest to the technical nature of the claimed method and selected as a prototype is the method RU №2614983, 2017.

The essence of this method is as follows.

For each active satellite terminal, change the rate of incoming packet data traffic and the fullness of the input buffer of the satellite terminal, determine the optimal value of the quantile level for the predicted values ​​of the rate of receipt of packet data into the buffer of the satellite terminal for one, two and three transmission cycles forward, form a dynamic bandwidth reservation request taking into account the QoS requirements, based on the fullness of the input buffer and requests formed on previous transmission cycles, as well as forecasting the arrival rate of packet traffic, based on the quantile level characterizing redundancy redundancy, send the generated requests to the central station of the network, the central station recalculates the amount of bandwidth requested by the terminal into the number of time slots,

At the same time, in each active satellite terminal, traffic is classified by service discipline and type of reservation (CRA, RBDC, VBDC).

For initial requests, the long-term statistical profile of the incoming aggregated traffic is determined, the short-term statistical profile of the aggregated traffic is determined, the measurement parameters of the sample mean and the sample dispersion of the bit rate of the incoming user traffic, namely the measurement period, the quantization step of the speed range, are measured on the selected period of the sample mean and selectively varying the rate of incoming user traffic.

The parameters of the predicted dynamics of the conditional average and the variance of the conditional average bit rate are determined, namely, the regression coefficients and the moving average, based on the short-term statistical profile of the traffic, the expected time of inertia of the reservation and distribution, the conditional average and the variance of the conditional average bit rate of the incoming satellite terminal are predicted traffic.

A request for bandwidth allocation with information about the predicted statistical traffic profile and reservation parameters is formed and sent to the central station, taking into account the statistical traffic profiles at the central station, forming a time-frequency transmission plan, determining the optimal value of the time slots for satellite channels of the network for one cycle transmissions send time-frequency plans to terminals.

The method prototype has the following disadvantages.

1. Do not use fractal analysis of the pulsating structure of telecommunications traffic.

2. Do not use simulation to predict the occurrence of overload.

The prototype does not take into account the fractality of telecommunications traffic, and it is therefore not able to provide quality service with strong pulsations.

Also in the prototype do not produce the determination of the queue size for the future through simulation, which does not allow the mechanisms to ensure the quality of service in advance to adjust the parameters of the serving device.

The problem that the proposed method solves is to improve the quality of telecommunications traffic service by determining the optimal parameters of the queues of the serving device, the traffic profile and the actual size of the queue for the future using mathematical and simulation modeling.

The operation of the invention is illustrated in the following graphics:

FIG. 1 is a functional diagram of a method for improving the quality of transmission of fractal telecommunications traffic.

To solve this problem, a method is proposed for improving the quality of transmission of fractal telecommunication traffic, namely, that:

from the interface of the operator 6 enter into the mathematical modeling unit the required values ​​of the quality of service indicators for queues with different priority;

they configure the router through the operator interface, determine priorities for different flows and set the current buffer size and service intensity for them, which are subsequently adjusted from the mathematical modeling block in automatic mode.

a sniffer 7 is connected to one of the outputs of router 5;

The sniffer 7, which is a computer with special software installed on it, collects data on incoming traffic 2 of router 5 (arrival time of the packet, priority class, packet size);

data on the input traffic from the sniffer 7 is transmitted to the fractality analyzer 9. The functionality of this unit is implemented by specialized software (Belov P.Yu., Belov K.Yu., Polous AI, Klimenko VV) Program for analyzing the fractality of real information traffic (PAFRIT). Certificate of state registration of computer programs No. 20134613976 dated 04.14.2014.). Traffic analysis is performed as follows.

Step 1. Generate tables of data on input packets. At this step, the software performs the removal of information about traffic with a sniffer 7.

Step 2. Aggregate traffic. At this step, the traffic is aggregated, which is necessary for the work of the methods for determining the Hurst index H and determining the form of distribution of the input stream.

Step 3. Determine the Hurst index H by various methods The Hurst index is determined by three methods: normalized span; dispersion index for samples; dispersion graph. To obtain the numerical values ​​obtained in the methods used, the dependences are approximated by the least squares method.

Step 4. Determine the resulting Hurst score as the average of the three methods.

The obtained value of the measure of fractality H is transmitted to the block of mathematical modeling 11 and the traffic generator 8;

In the mathematical modeling unit 11, the queue parameters are calculated for each type of traffic in accordance with the fractal telecommunication traffic service models (Nazarov A.N., Sychev KI. Models and methods for calculating the performance indicators of the node equipment and the structural network parameters of the communication networks generations - Krasnoyarsk: Publishing house LLC Polikom, 2010. - p. 254). The mathematical models used allow us to calculate the values ​​of the quality of service indicators through the values ​​of the Hurst indicator.

The Hurst index H in turn is calculated by the following known methods, followed by averaging the sum of the values ​​obtained:

1. Graph of the variance;

2. R / S statistics

<img file="RU2677373C1_D0001.tif" he="29" wi="129" img-format="jpg" img-content="undefined" />

In expression (1) <img file="RU2677373C1_D0002.tif" he="14" wi="33" img-format="jpg" img-content="undefined" /> and is a sample average for the period of time N. When plotting <img file="RU2677373C1_D0003.tif" he="14" wi="16" img-format="jpg" img-content="undefined" /> from N on a logarithmic scale, after least squares approximation, get a straight line with a slope of N.

3. The variance index for reports <img file="RU2677373C1_D0004.tif" he="15" wi="33" img-format="jpg" img-content="undefined" />.

To determine the Hurst index, the obtained dependences are approximated by the least squares method.

on the traffic generator 8, the current value of the Hurst index is transmitted from the fractality analyzer 9; the telecommunication traffic is generated with the distribution of the data volume corresponding to the incoming input;

when generating, use the function inverse of the Pareto distribution function and the random number sensor evenly distributed from 0 to 1;

The generated traffic is fed to the simulation simulation unit 10, where accelerated simulation of the operation of router 5 is performed. Accelerated simulation is performed by not performing modulation and coding operations, which during actual processing are carried out in modem 4 and transmitter 3;

the acceleration of the simulation of the router, due to the fact that the speed of the input and output interface is limited by the performance of the computer, carried out by maximizing the use of computer processor resources;

during the operation of the simulation model, the real values ​​of the quality of service indicators (delay, variation of delay, probability of failure in service) for the future are determined and transferred to the mathematical modeling unit 11, which is compared with the QoS requirements and, if necessary, the parameters of the mathematical model are adjusted;

in the block of mathematical modeling 11 determine the optimal parameters of the queues based on the fractal traffic servicing models and make their adjustment in the router 5.

Router 5, thanks to the settings received from the mathematical modeling unit, processes traffic taking into account the fulfillment of the quality of service requirements, and generates data flow at the input of the modem 4.

modem 4 is connected to transmitter 3 in which they form signals transmitted through propagation medium 1.

The “industrial applicability” of the method is due to the availability of the possibility to implement it on standard computers by installing additional software and connecting them via Ethernet to the node's router.

Comparison of the claimed method of transmitting fractal telecommunication traffic with the prototype shows that the claimed method differs significantly from the prototype.

General features of the proposed method and prototype:

1. Use a mathematical calculation of parameters characterizing the distribution of the data volume of the input stream.

2. Use standard methods for ensuring quality of service (redundancy, profiling, shaping).

3. Get the current statistical characteristics of the input telecommunications traffic.

Distinctive features of the proposed solution.

1. Use simulation modeling, which allows to predict the occurrence of overload in case of incorrect operation of the mathematical model used.

2. Use the Pareto distribution as a model for packet receipt and service.

3. As a distribution parameter, the Hurst index is used, which characterizes the measure of fractality, that is, describes the degree of pulsations.

Graphic materials used to illustrate the proposed solution:

FIG. 1 is a functional diagram of a method for improving the quality of transmission of fractal telecommunications traffic.

The developed method allows to distribute the throughput capacity of the control center in such a way as to ensure that the requirements of quality of service are met by the average delay time, in accordance with the self-similar structure of each priority flow.

FIG. 2, 3, and Table 2 present the results of modeling the operation of the router when servicing self-similar telecommunication traffic using the prototype method and the proposed method. The simulation conditions are presented in Table 1. A comparison of the graphs in Figs 2 and 3 allows us to conclude that the delay value takes a smaller value under the same initial conditions. Thus, the objective of the invention to improve the quality of service is achieved.

<img file="RU2677373C1_D0005.tif" he="174" wi="160" img-format="jpg" img-content="undefined" />

<img file="RU2677373C1_D0006.tif" he="43" wi="160" img-format="jpg" img-content="undefined" />