RU2678470C1 - Data blocks in the switched network multi-routing method - Google Patents

Abstract

FIELD: data processing.SUBSTANCE: invention relates to the field of telecommunications, in particular to the data routing methods, and can be used in existing and emerging networks with data blocks switching to provide traffic engineering and the single data stream segments parallel delivery. Data blocks in the switched network multi-routing method is implemented by means of the switched network node software, including the data blocks preparation module, the data blocks transmission module, the multi-routing module. Method comprises the following steps: executing the several shortest routes on the network connectivity graphs searching algorithm for the multi-routing tables formation at the network's boundary nodes; stream segments distribution along several routes in the route subsystem; setting the route in the data block header field in the form of the network addresses chain or the subscribers and nodes network addresses identifiers; data block switching implementation in accordance with the network addresses or identifiers chain in the data block header field on one of the switching tables at intermediate nodes; the data stream collection in the boundary nodes.EFFECT: technical result consists in reduction in the time of data streams in the network bringing.3 cl, 3 dwg

Description

The invention relates to telecommunications, and in particular, to data routing methods, and can be used in existing and emerging networks with switching data blocks (messages, packets, cells, segments of the data stream, etc.) to provide traffic engineering (Traffic Engineering, TE) and parallel delivery of segments of the same data stream along different routes in order to reduce the time to bring the streams to the network.

Currently, there are a number of well-known routing protocols - RIP (Routing Information Protocol), IGRP (Interior Gateway Routing Protocol), EIGRP (Extended IGRP), IS-IS (Intermediate System - to– Intermediate System), OSPF (Open Shortest Path First) , PNNI (Private Network - to - Network Interface), etc. The listed protocols, which are based on the solutions of the "shortest path" search problem, implement mainly a one-step, one-way routing strategy. In one-step routing, each router is responsible for choosing only one step of the route (to the next router), and the final route is the result of all routers.

A number of routing protocols out of the above provide load balancing, which assumes multi-path routing of transmitted data. There is, for example, a multi-path routing method using splitting a data traffic stream (patent RU No. 2636665, IPC H04W 40/02, H04L 12/803, publ. 11/27/2017), which consists in the fact that the intermediate system receives data traffic flows from the terminal sender systems, each data traffic stream is represented as protocol data units or bytes, in the intermediate system, each data traffic stream is split into two or more data traffic substreams, each data traffic substream is represented as protocol data blocks or and bytes of the corresponding data traffic stream, route protocol data units or bytes of each data traffic substream along two or more reliable routes in the intermediate system, recombine each data traffic stream consisting of two or more data traffic substreams in the intermediate system, and the streams are transmitted from the intermediate system data traffic to the destination terminal system. Two or more reliable routes in the communication network are calculated, using a combination of Dijkstra's algorithms, constructing a truncated tree of events wide and a method for combining simple chains taking into account the absorption effect, filling out routing tables for two or more reliable routes of the communication network in order to reduce their reliability, form a vector of values of splitting the data traffic stream into two or more substreams of data traffic, taking into account the calculated reliable routes of the communication network, the throughput of communication channels, download channel communication fishing.

The disadvantages of such routing methods are as follows. From the found path in the routing table, only the next IP address of the nearest router (one-step routing) is remembered, and the remaining intermediate addresses are discarded. Each router in IP networks decides on only one step of the route. Ultimately, all threads converge on a more “better” resource. In addition, the organization of parallel delivery of segments of the same stream on different routes is impossible.

Multistep routing methods are source routing. In accordance with this method, the node (router, switch, server, etc.) of the sending subscriber sets the complete route of the data block through all intermediate nodes. The most famous implementation of multi-step routing is the MPLS protocol (Multi-Protocol Label Switching - multi-protocol switching based on the use of labels). In addition, it is known, for example, a multi-step routing method from application No. 2016127806 dated 01/19/2015 (IPC H04L 12/28 (2006.01), H04L 29/06 (2006.01)), the last change in status: 12/20/2017, namely, that the method comprises the steps of receiving, at a boundary node in a domain, a data packet along an interdomain path, the packet header having a list of jumps on the route from the source attached to the data packet, the list of jumps on the route from the source indicating the identifier of the path segment of the intra-domain path without indications of individual transitions along about intradomain cutting path, and wherein the segment is a portion intradomain path interdomain path that passes through the active domain; identify, at the boundary node, a transition list associated with the path identifier, the transition list indicating a sequence of transitions along a segment of the intra-domain path; replace the path identifier in the packet header with the transition list; and relay the data packet to the next transition of the segment of the intra-domain path in accordance with the list of transitions.

As part of the MPLS technology, a traffic engineering procedure is envisaged, which assumes multi-path routing of transmitted data. Traffic Engineering refers to methods and mechanisms that allow achieving a balanced load of all network resources by rational selection of a set of routes for traffic flow through the network. An implementation of this mechanism is the MPLS TE protocol (RFC 2702, Requirement for Traffic Engineering Over MPLS). In order to ensure that packets are not promoted using routing tables, but using switching tables, MPLS is used to promote packets by labels.

MPLS TE is difficult to implement, since in practice it is necessary to build a transport network based on MPLS technology and deploy a stack of additional extension protocols, which makes this technology available only to telecom operators and large corporations, and statistical data is often used to build a load balancing plan nature and not reflecting possible unexpected traffic changes. In addition, the poor scalability of MPLS TE solutions due to the quadratic growth rate of the number of virtual tunnels in the fully connected topology of a single transport network can be attributed to disadvantages. It is known that the choice of paths for tunnels for each stream is usually carried out in turn by the "shortest path" method. In this case, the administrator determines the order of choice only on the basis of his intuition. A backup tunnel is formed at the very beginning and its quality depends on the administrator, and in practice has a lower quality, since it is considered as a temporary solution. In addition, it is possible that the standby tunnel can pass through the failed equipment of the main tunnel. In order to provide Quality of Service (QoS) parameters for different types of traffic, the service provider for each class of forwarding equivalence establishes a separate tunnel system in the network. This creates inconsistency in MPLS TE technology. On the one hand, the objective function TE is to increase the utilization rate of equipment and the network as a whole, and on the other hand, for traffic sensitive to delays, it is necessary to make a reservation so that the maximum utilization rate of tunnel resources is in the range of 0.2-0.3 - otherwise delays and their variations, packet loss will go beyond acceptable limits.

Thus, the specificity of the existing methods of multi-route delivery is that the paths are selected only in order to maintain the network load balance and are not used, due to the inherent capabilities, to improve the quality of data delivery, for example, for parallel delivery of a single stream segment to reduce the time for bringing everything flow.

The objective of the invention is to increase the efficiency of the distribution of flows in a network with switching data blocks, namely, increasing the utilization of equipment and network performance in general, increasing the stability of the functioning of a network with switching data blocks to the effects of computer attacks, reducing the time to bring data streams.

The technical result of the invention is to maintain the automatic establishment of route subsystems, consisting of an excess (more than one) of the number of explicitly specified routes between the interacting equipment (nodes) of the network and multi-step routing of the data blocks that make up the stream using several or all routes from the subsystem.

The problem is solved, and the technical result is achieved by the method of multi-routing data blocks in a switched network, implemented by means of the software of the switched network node, including the module for preparing data blocks of the sending subscriber node, the receiving subscriber boundary node and the intermediate node, the data block transmitting unit the sender, the boundary node of the recipient subscriber and the intermediate node, the multi-routing module installed on the boundary nodes, providing the construction of one or more graphs of network connectivity based on receiving node state information, the method comprising the steps of executing an algorithm for searching for several explicitly given shortest routes on network connectivity graphs to form multi-routing tables in the network boundary nodes, distributing stream segments along several routes in the route subsystem , setting the route in the header field of the data block in the form of a chain of network addresses or identifiers of network addresses of subscribers and nodes, communicating data block in accordance with the chain of network addresses or identifiers in the header field of the data block for one of the switching tables in the intermediate nodes, as well as the collection of the data stream in the boundary nodes, and the multi-routing table is carried out for each connectivity graph and consists of a list of route subsystems between the boundary node of the sending subscriber and other boundary nodes, a list of several shortest routes for each route subsystem with a complete list of network addresses of subscribers and nodes from the route, with the correspondence list of identifiers to network addresses is stored in each boundary node, and the switching table containing a list of network addresses and a list of their identifiers is stored for each connectivity graph in intermediate nodes. According to the invention:

- the number of switching tables and multi-routing tables is determined by the number of network connectivity graphs;

- introduce nodes, considered as an external element of the calculation of several explicitly given shortest routes on the graphs.

The technical result is achieved due to:

- rejection of “shortest path” routing and the transition to multi-routing along several shortest routes — the formation of route subsystems;

- setting explicit routes for the data block in the header of the data block using a chain of network addresses or network address identifiers;

- distribution of flows along several routes within the route subsystem;

- the formation of several switching tables in the intermediate nodes of the network with switching data blocks, each switching table has its own identifier.

The applicant has not found technical solutions having a combination of essential features that are not obvious leading to the specified technical result. That allows us to conclude that the invention meets the criterion of "inventive step". An inventive step is seen in the following:

- multi-routing along several shortest routes allows to increase the utilization rate of equipment and increase the network performance as a whole, as well as carry out parallel delivery of segments of a stream of data blocks to reduce the time to bring the entire stream;

- setting explicit routes for the data block in the header of the data block using network address identifiers can improve the stability of the network with the switching of data blocks under the influence of computer attacks;

- the formation of several switching tables in the intermediate nodes of the network with switching data blocks with its identifier allows you to create subnets with different characteristics on the resources of one network, including virtual networks.

The invention is illustrated by drawings:

Figure 1 shows the header format of the IP packet (data block) version 4;

Figure 2 shows the interaction of nodes in a switched network;

Figure 3 shows a histogram of the effectiveness of countering computer attacks.

The method of multi-routing data blocks in a switched network is carried out using the router software as follows.

формируют из данных пользователя поле «данных» блока данных и заголовок блока данных (например, заголовок IP-пакета, фиг.1). Выбирают по предопределенным правилам распределения потоков (например, по правилам параллельной доставки сегментов потока по разным маршрутам (фиг.2) или по правилам ремаршрутизации – правилам переключения на резервный маршрут) несколько маршрутов из

в граничный узел абонента получателя данных

, объединенные в маршрутную подсистему

. Вносятся в заголовок, как стандартные данные, такие как «адрес источника»

, «идентификатор блока данных», который используют для распознавания блоков данных, образовавшихся путем фрагментации потока данных на сегменты, и т.д., а в поле «адрес назначения» ставится сетевой адрес первого промежуточного узла на маршруте. В поле «параметры» заголовка блока данных вносится список идентификаторов сетевых адресов узлов, через которые проходит маршрут, а также, в соответствии с выбранным вариантом структуры сети из заранее определенных, идентификатор таблиц коммутации в промежуточных узлах. Поле «параметры» заголовка IP-пакета предназначено, в том числе, и для этих целей. Идентификаторы таблиц коммутации едины для каждого варианта структуры сети (фрагмента сети). В предлагаемом способе формируется несколько таблиц коммутации в промежуточных узлах. Ведение таких таблиц позволяет сформировать на ресурсах одной сети с коммутацией блоков данных нескольких подсетей с различными характеристиками. Например, некоторые виды трафика ни в коем случае не должны идти через радиорелейные линии - проблемы с синхронизацией. Или трафик широкополосного радиодоступа нельзя пускать через линии какого-то оператора связи по экономическим соображениям. В промежуточном узле производится коммутация пакетов. Получив блок данных, в промежуточном узле анализируется поле «адрес назначения», если адрес не соответствует сетевому адресу промежуточного узла, блок данных сбрасывается. В противном случае, анализируется поле «параметры» заголовка блока данных. В поле «параметры» выбирается идентификатор таблицы коммутации, а в таблице коммутации по идентификатору адреса следующего узла его сетевой адрес и прописывается в поле «адрес назначения». При несовпадении каких-либо идентификаторов блок данных сбрасывается. В поле «адрес источника» проставляется свой сетевой адрес. Каждый промежуточный узел осуществляет аналогичные действия. В граничном узле

получателя производятся стандартные операции обработки заголовка и поля «данных» блока данных для предоставления данных абоненту-получателю.1. By means of the data block preparation module in the boundary node of the subscriber of the data sender

form from the user data the “data” field of the data block and the header of the data block (for example, the header of the IP packet, Fig. 1). Select according to predefined rules for the distribution of flows (for example, according to the rules for the parallel delivery of stream segments along different routes (Fig. 2) or according to the rules of rerouting - the rules for switching to the backup route) several routes from

to the boundary node of the subscriber of the data recipient

combined into a route subsystem

. Entered in the header as standard data, such as “source address”

, “Data block identifier”, which is used to recognize data blocks formed by fragmenting the data stream into segments, etc., and the network address of the first intermediate node on the route is set in the “destination address” field. In the “parameters” field of the data block header, a list of identifiers of the network addresses of the nodes through which the route passes is entered, as well as, in accordance with the selected option of the network structure from predefined ones, the identifier of the switching tables in the intermediate nodes. The “parameters” field of the IP packet header is intended, among other things, for these purposes. The identifiers of the switching tables are the same for each variant of the network structure (network fragment). In the proposed method, several switching tables are formed in the intermediate nodes. Keeping such tables allows you to create several subnets with different characteristics on the resources of one network with switching data blocks. For example, some types of traffic should in no case go through radio relay lines - problems with synchronization. Or broadband radio access traffic should not be allowed through the lines of any telecom operator for economic reasons. In the intermediate node, packet switching is performed. Having received the data block, the “destination address” field is analyzed in the intermediate node; if the address does not match the network address of the intermediate node, the data block is reset. Otherwise, the “parameters” field of the data block header is analyzed. In the “parameters” field, the identifier of the switching table is selected, and in the switching table, by the identifier of the address of the next node, its network address is written in the “destination address” field. If any identifiers do not match, the data block is reset. In the field "source address" is set your network address. Each intermediate node performs similar actions. In the boundary node

the receiver performs standard operations of processing the header and the data field of the data block to provide data to the subscriber-recipient.

и

. Маршрутная подсистема представляет собой совокупность явно заданных маршрутов. Постановка задачи выработки маршрутов в маршрутной подсистеме следующая. Пусть сеть с коммутацией блоков данных задана графом

с двумя выделенными (граничными) непересекающимися узлами

, вес

каждого ребра

и положительные целые числа

и

. Требуется найти

явно заданных простых маршрутов (список, включающий

и все промежуточные узлы между ними) от

до

, веса которых не превосходят

. Простой маршрут – маршрут без петель. Каждому элементу списка сетевых адресов узлов, через которые проходит маршрут, присваивается идентификатор. Формируется таблица мультимаршрутизации (маршрутных подсистем

). Таблица мультимаршрутизации содержит поля «список адресов узлов маршрута» и список идентификаторов узлов маршрута». Таблица мультимаршрутизации служит для ведения списка маршрутов в маршрутной подсистеме для каждой пары

. Идентификаторы сетевых адресов узлов едины для всех узлов сети (фрагментов сети). С заданной периодичностью и/или при выходе параметров сетевого оборудования за установленные пределы, корректируется таблица мультимаршрутизации.2. The routing module, based on information about the state of the network, generates route subsystems for each pair of boundary nodes

and

. The route subsystem is a collection of explicitly defined routes. The statement of the problem of generating routes in the route subsystem is as follows. Let a network with data block switching be given by a graph

with two selected (boundary) disjoint nodes

, the weight

each rib

and positive integers

and

. Required to find

explicitly defined simple routes (list including

and all intermediate nodes between them) from

before

whose weights do not exceed

. A simple route is a route without loops. Each element of the list of network addresses of the nodes through which the route passes is assigned an identifier. A table of multi-routing (route subsystems is formed

) The multi-routing table contains the fields “list of addresses of route nodes” and a list of identifiers of route nodes ”. The multi-routing table serves to maintain a list of routes in the route subsystem for each pair

. The identifiers of network addresses of nodes are the same for all nodes of the network (network fragments). With a given periodicity and / or when the network equipment parameters go beyond the set limits, the multi-routing table is adjusted.

3. Using the data transmission module, the checksum of the header of the data block is checked, the address (link level) of the receiver is determined, and the data block is sent directly taking into account the sequence, fragmentation, filtering, etc.

The multi-routing method can be used to develop routing algorithms:

- centralized or decentralized;

- static or dynamic;

- single-level or hierarchical;

- intra-domain or cross-domain;

- unicast or group.

The multi-routing method can be used both to create network equipment of a network with switching data blocks, for example, an IP network with multi-routing, and to create an overlay routing network. An overlaid network (or overlay) in the general case is a logical network organized on top of the existing infrastructure and is an add-on over standard network protocols. In this case, the method solves several problems of networks with switching data blocks:

1. It is focused on improving the performance of bringing traffic:

- increasing the likelihood of bringing data blocks;

- reducing the time to bring the data stream by transmitting segments of this stream by different routes.

2. Focused on improving utilization:

- maximization of network equipment load;

- maximization of the overall network performance with switching data blocks.

The proposed multi-routing method was investigated for the possibility of distributing a stream entering the network according to the criterion of the average packet delivery time. In this case, the following mathematical model was used for research.

, где

 – число пакетов на узле

, адресованных узлу

. Обозначим через

суммарное число пакетов в начальный момент времени

:Let the network have several independent routes, each of which can transmit one packet per unit of time. The initial state of the network is set by the matrix

where

- the number of packets on the node

addressed to the node

. Denote by

total number of packets at the initial time

:

.

.

No new packets arrive in the network, time is discrete,

. We will call the assignment step for each route the distribution of flows.

,

, a specific package for delivery at a given time interval (

) Upon reaching the end of the route, the packet is dropped from the network. In this model there are no priority packets, but there are only counters of the number of packets

 on the node

 at time

addressed to the node

,

. Totality of these counters

 describes the state of the network at the time

.

момент полного освобождения сети. Тогда критерием среднего времени доставки пакетов будетDenote by

the moment of complete release of the network. Then the criterion of the average packet delivery time will be

Studies have shown that:

0,6).1. The total realized load (performance) in the network is higher than when using the MPLS TE mechanism (average utilization of network equipment

0.6).

2. The value of the realized generalized load on the network is higher than the realized load on any single route. The maximum value of the realized generalized load at any time is exactly equal to the total value of the realized load on each route, and its value exceeded the value of the realized generalized load by more than 30% relative to the MPLS TE mechanism. At the same time, this is not reflected in the average packet delivery time.

определяется по формулеParallelization of information transfer has been known for a long time and at the present time, parallelization of information transfer through various communication channels is actively used, including to reduce the time of information delivery. In contrast to this known mechanism, in the proposed method, streaming is parallelized along network routes, this solution uses the natural redundancy of the network and is, of course, much cheaper. Efficiency of parallelization of flows on several network routes

determined by the formula

,

,

 - время затраченное на последовательную доставку потока данных одним маршрутом,

 – время затраченное на доставку того же потока данных по нескольким маршрутам.

зависит от характеристик каждого элемента этого маршрута, а

 – от характеристик самого медленного маршрута при делении потока по маршрутам, то естьWhere

- the time spent on sequential delivery of the data stream in one route,

- The time taken to deliver the same data stream over multiple routes.

depends on the characteristics of each element of this route, and

- from the characteristics of the slowest route when dividing the stream by routes, i.e.

,

.

,

.

Obviously, the time required to bring large amounts of information will depend on the number of routes used and their performance characteristics, but the efficiency of thread parallelization of 30% is minimal in any ramified network.

The record operation in the “parameters” field of the data block header of the list of identifiers of network addresses of nodes through which the route passes and switching table identifiers allows you to resist computer attacks such as “denial of service” and “imposing false control information (address, route)”.

и

обеспечивается коэффициент связности выше требуемого. Воздействие компьютерной атаки на смежный узел с граничным узлом (критический узел) при штатной высокой эффективности выполнения процессов обмена данными снижает коэффициент связности в IP-сети до 0,9988 (что уже ниже технической нормы), а в сети с мультимаршрутизацией – до 0,999932. При воздействии на два узла смежных с граничным узлом, IP-сеть полностью деградирует. Сеть с мультимаршрутизацией сохранит 43% от возможных маршрутов доставки данных даже при воздействии на три узла смежных с граничным узлом.The histogram (figure 3) shows that for the selected network structure with its normal functioning between

and

a connectivity coefficient above the required value is provided. The impact of a computer attack on an adjacent node with a boundary node (critical node) with regular high-performance data exchange processes reduces the connectivity coefficient in an IP network to 0.9988 (which is already below the technical norm), and in a network with multi-routing to 0.999932 . When exposed to two nodes adjacent to the boundary node, the IP network is completely degraded. A network with multi-routing will save 43% of the possible data delivery routes, even if three nodes adjacent to the boundary node are exposed.

Thus, unlike the existing routing methods, the multi-routing method simultaneously distributes segments along several routes, thereby reducing the time it takes to stream data streams, and traffic engineering is also carried out - increasing the utilization rate of all network devices. The use of address identifiers in host address chains, in contrast to source routing, is designed to counter computer attacks.

An example of a specific implementation of the method.

Precondition: The sending subscriber (the source of the data block stream) intends to send this stream to the receiving subscriber. The task is to reduce the time to bring this stream to the subscriber-recipient.

Incoming information : stream (set) of data blocks for transfer, address of the subscriber-receiver, type of service of data blocks, etc.

The table shows the algorithm of the main actions.

Table

Step Description one The subscriber-sender, on the basis of his funds, prepares a stream of data blocks.
At the same time, he performs the following operations:
- forms a set of data blocks (stream) from the message;
- fills in the service part of the data block, namely its address and the address of the recipient, type of service, etc. For example, figure 1 shows the format of the IP packet (data block) version 4. In the field "type of service" is the identifier of the service "reduce the time of bringing the stream" 2 The sending subscriber directs the flow of data blocks to the preparation module of the data blocks of the boundary node 3 Multi-routing module (installed only on boundary nodes):
- Builds a graph of network connections, the vertices of the graph are nodes with modules for preparing data blocks interacting with subscribers (boundary nodes) and intermediate nodes with modules for preparing data blocks that interact only with boundary nodes and with each other. The edges of the graph are node interfaces;
- exchanges information about the state of nodes and interfaces for constructing a graph of network connections. For example, in the OSPF (Open Shortest Path First) routing protocol, such messages are called router links advertisement. four Multi-routing module:
- forms a “tree” of route subsystems. Each routing module considers the node on which it is installed (boundary node) to be the center of the network and looks for a set number of explicitly specified shortest routes to each other boundary node. This set of routes is called a route subsystem;
- each route found in this way is entered in the multi-routing table in the form of a chain of addresses of subscribers and nodes through which the data block will pass;
- maintains a multi-routing table, in which it enters data on each route subsystem: each route (addresses of sending subscribers, a chain of node addresses and addresses of receiving subscribers and a list of identifiers of these addresses). 5 When operating a network, the multi-routing module, based on the status information of a functioning network, makes adjustments to the multi-routing table. 6 Module for preparing data blocks of the boundary device of the sending subscriber. Having received a request from the sending subscriber for the delivery of a stream of data blocks:
- at the address of the subscriber-recipient selects the route subsystem;
- by the service identifier, “reduction of the flow completion time” selects the necessary number of routes in the route subsystem;
- splits the data stream into segments (the number of segments corresponds to the number of routes). Enters the identifier of the stream segment in the service field of the data block;
- enters the corresponding header field of each data block, for example, in the subfield of the IP Options field of the packet (Figure 1), of each stream segment, the corresponding chain of identifiers of network addresses of subscribers and route nodes;
- in the "Sender Address" field (for example, the "Source Address" field in the IP packet, Fig. 1) replaces the address of the subscriber-sender with his address;
- selects the address of the next node after himself in the address chain and enters this address in the "Destination Address" field (for example, the "Destination Address" field in the IP packet, Fig. 1);
- transmits each data block to the data block transmission module. 7 The data block transfer module performs standard data transfer operations: it checks the header checksum, determines the address (link level) of the recipient, and directly sends the data block taking into account the sequence, fragmentation, filtering, etc. 8 Intermediate node data block preparation module:
- Having received the data block, it checks the field “Recipient address”, if the data block is not addressed to it, it resets the data block;
- on the field "Sender address" and the identifier of the previous node checks the correctness of the chain of nodes. To do this, the module for preparing the data blocks of the intermediate node at the sender address, in the switching table, finds the identifier of this address and checks the identifier from the chain of identifiers of the node addresses with the previous identifier in the chain written in the corresponding field of the data block. If the identifier does not match, the data block is discarded;
- selects the identifier of the switching table in the corresponding field of the data block, and in the switching table, selects the network address from the address identifier of the next node and writes it in the “Recipient address” field;
- the “Sender address” field is entered your network address;
- transmits a data block to a data block transmission module.
Note: this step is repeated until the data block reaches the module for preparing the data blocks of the boundary node of the subscriber-recipient. 9 The data block transfer module performs the operations of step 7. 10 The module for the preparation of data blocks of the boundary node of the subscriber-recipient performs the steps of step 8, with the following changes:
- in the field “Recipient address” of each data block, affix the address of the recipient;
- in the field "Sender address" of each data block, affix the address of the subscriber-sender. eleven Module for preparing data blocks of the boundary node of the subscriber-recipient:
- clears other service fields of the data block that provide the implementation of the multi-routing method (brings the data block to the form of the data block represented by the sending subscriber);
- collects stream from segments of the stream;
- transmits a data block to a data block transmission module. 12 The data block transfer module performs the operations of step 7.

Thus, the use of the invention allows to increase the efficiency of the distribution of flows in a network with switching data blocks, namely, to increase the utilization rate of equipment and network performance in general, to increase the stability of the functioning of a network with switching data blocks to the effects of computer attacks, to reduce the time for bringing data streams.

Claims (3)

1. A method for multi-routing data blocks in a switched network, implemented by means of the software of a node of a switched network, including a module for preparing data blocks of a boundary node of a sending subscriber, a boundary node of a subscriber-receiver and an intermediate node, a module for transmitting data blocks of a boundary node of a subscriber-sender, a boundary node subscriber-recipient and an intermediate node, a multi-routing module installed on the boundary nodes, providing the construction of one or more graphs connectivity a network based on receiving node state information, the method comprising the steps of executing an algorithm for searching for several explicitly defined shortest routes on network connectivity graphs to form multi-routing tables at the network boundary nodes, distributing stream segments along several routes in the route subsystem, setting a route in the header field data block in the form of a chain of network addresses or identifiers of network addresses of subscribers and nodes, the implementation of the switching of the data block in accordance with the network chain addresses or identifiers in the header field of the data block for one of the switching tables in the intermediate nodes, as well as the collection of the data stream in the boundary nodes, and the multi-routing table is maintained for each connectivity graph and consists of a list of route subsystems between the boundary node of the sending subscriber and other boundary nodes , a list of several shortest routes for each route subsystem with a complete list of network addresses of subscribers and nodes from the route, a list of correspondence of identifiers to network addresses and is stored in each boundary node, and the switching table containing a list of network addresses and a list of their identifiers is stored for each connectivity graph in the intermediate nodes. 2. The method according to claim 1, characterized in that the number of switching tables and multi-routing tables is determined by the number of network connectivity graphs. 3. The method according to claim 1, characterized in that the nodes are introduced, considered as an external element of the calculation of several explicitly given shortest routes on the graphs.

Images