DE19923245A1 - Route selection method especially for ATM - Google Patents

Abstract

The method involves providing several routes (R(V)) for one information transmission (V). The route (Rv) for the information transmission is determined by an at least partly arbitrary selection of one of the routes. Preferably, the communication network (KN) which provides the transmission paths (U) is connection orientated. The routes are formed as optimum routes (Rmin), which have identical route costs (RK) determined in dependency of the link costs (L) allocated to the transmission paths.

Description

Communication networks are usually either packaged oriented or trained as line-oriented networks. Here, packet-oriented networks are more for transmission of information without real-time character such as B. data, Email or files suitable while line-based Networks well on the transmission of information with real time character such as B. designed language or moving images are. In the course of the convergence of line and package However, networked networks are used in packet-oriented networks taking also voice and moving picture information transmitted. Examples of packet-oriented networks are the Internet or ATM (= Asynchronous Transfer Mode), where the name ATM occasionally also as a synonym for B-ISDN (= Broadband In tegrated Services Digital Network) is used. At Bei ATM is the packet-oriented network technology in the further explained.

The packet is characteristic of packet-oriented networks oriented transmission of information. In ATM networks the information is, for example, packets of equal length - also called "ATM cells" - divided the a 5-byte cell header and one Have 48 bytes of information (payload). The individual cells are loaded by the cell heads agreed information flows - also "virtual connections" called - assigned. In contrast to, for example, one line-oriented TDMA method, in which time slots assigned different types of data in advance are, the In arriving at an ATM interface formation streams segmented into the 53-byte cells mentioned and then sequentially these cells in the order in which they were created. The TDMA at  Coming multiplexing is also called "static Multiplexing "and that used at ATM as" sta statistical multiplexing ". As a result of the flexibility statistical multiplexing can be the information currents at ATM have any data rates, while at static multiplexing the data rate of the individual informa tion currents - also called "connections" - because of the fixed Assignment of the time slots to the information streams sets is - e.g. B. to 64 kbit / s with ISDN.

As a result of this difference is in packet-oriented networks routing a requested connection depending on the route remaining capacity while in lei tion-oriented networks are basically independent of the load on the individual transmission paths. For example can ent on a route in a line-oriented network long the z. B. according to a TDM method 30 connections in dedicated time slots with 64 kbit / s capacity each can be performed, in any case, another Connection established if there are already 29 connections are established because the further connection because of their con constant data rate does not require a higher data rate than that 64 kbit / s remaining capacity available. Along one Route in a packet-oriented network with an assumed one The remaining capacity of 30 Mbit / s can only be connected that a data rate lower than 30 Mbit / s has been requested. Connections with a height However, their data rate is rejected. If alterna tive routes exist, they can be replaced along a route Alternative route set up with sufficient remaining capacity become. To determine an alternative route, however, is re-routing required.

Various routing methods are known with which Routes in networks can be determined. A possibility is the so-called "source routing", in which starting from a Initial switching node the complete route to a destination  Switching node is determined. For ATM networks is at for example on the part of the ATM Forum in the context of the so-called in the PNNI (= Private Network-Network Interface) specifics tion source routing required. Here the route from the Initial switching node determined and then at Establishing the connection, the calculated route to the operator knots along the route with the help of signaling averages. Another possibility is the so-called "hop-by- Hop routing "in which ent from each switching node the rest or the next section of the route Route is recalculated. This method comes for example assign them on the Internet or in ATM networks without source routing Application.

In order to exclude those routes that are use constant or interrupted transmission paths so-called flooding process proposed. In doing so, everyone Switching nodes the traffic at defined times values of the transmission paths connected to them and to all other switching nodes within a group passed on. This information sharing will be "flooding" called. Flooding can also be initiated become when the market values of the transmission routes sig change significantly - e.g. B. if the current utilization of a Transmission path with a total capacity of 150 Mbit / s more than 10 Mbit / s from the last passed load deviates. For example, in ATM networks under the PNNI specification proposed methods that a Rou ting algorithm in the switching nodes of the ATM network last measured traffic values of the directly to these provide connected transmission paths.

Routing in line-oriented, public is known Telephone networks. The routing is usually done in several steps because these networks are mostly large because of the The number of switching nodes is usually hierarchical are building. Connections are made in one of these networks  most step from an initial switching node to one lower hierarchy level to a switching node on the top hierarchical level, then in a second Step within the top hierarchical level to one that Switching node representing the connection destination and finally in a third step to the destination mediator routing nodes routed at a lower hierarchical level. Here, the first and third steps in general NEN by fixed routes or, if these z. B. un are broken by fixed alternative routes works, while the second step is often just the off choice of the transmission path between the two concerned Switching nodes of the highest hierarchical level requires because the switching nodes are at the top level which are usually largely completely meshed. The standardized signal for line-oriented telephone networks However, method 7 does not support Source Rou ting, d. H. the initial switching node can be one of it Do not forward the calculated route. Consequently, the Ver averaging nodes along the route also the did not put route, so this routing method in not hierarchically structured or only partially meshed Networks - e.g. B. the Internet - can only be used to a limited extent can. This can result in loops in the routes.

For routing in packet-oriented networks, this means patent DE 44 41 356 a dynamic routing method of enlisted in the blockage of transmission paths and from their frequency the status of the transmission way is determined. From destination traffic data can by the Using a routing management processor the probability Calculation of the availability of transmission paths offline be net. For such a calculation, for example, is suitable the forward-looking routing algorithm according to K.R. Krishnan, T.J. Ott in Forward-Looking Routing, A New State- Dependent Routing Scheme, Teletraffic Science for New Cost Effective Systems, Networks and Services, ITC-12 (1989). The  However, the method only considers connections of the same and constant bandwidth, as is the case for conventional telephone ver connections in circuit switching networks are typical, d. H. the bandwidth of a connection is z. B. 64 kbits / s. For packet-oriented networks such as B. ATM networks (Asynchronous Transfer Mode), however, is a constant bit rate of the off case, because connections can be made according to the verb wishes of the participants with different and time variable bandwidth. In addition to the ge wanted bandwidth, e.g. B. 1 Mbit / s, contain connection requests from participants often include information visibly the required connection quality.

The invention has for its object the routing for Improve communication networks. The task is performed by the Features of claim 1 solved.

The essential aspect of the invention is the selection a route through one of switching nodes and transmissions existing communication network, in which several Suitable routes are determined for the transmission of information and the route along which the information is transmitted tion takes place by at least partially any selection egg the appropriate routes are determined. The main advantage Part of the invention is to be seen in that the In formation transmissions in the communication network called traffic load on average evenly on the appropriate Routes is distributed. This reduces the likelihood of rejection probability, d. H. the probability that the informa not transmitted. In addition, the number of evaluating routes and consequently the processing time for the Evaluation of the routes advantageously reduced, since only the suitable routes.

According to an embodiment of the method according to the invention it is envisaged that the communication network connection theory  tert is - Claim 2. As a result, the invention used for example in ATM networks.

According to an embodiment of the method according to the invention it is envisaged that the suitable routes as optimal rou ten are formed, which are identical, depending on the Link costs assigned to transmission routes, mini Male route costs have - claim 3. Thus is an advantage the number of routes to be rated is further reduced. In addition, the communication network through the information transfer averages minimally burdened.

After a development of the method according to the invention for a period of time with unchanged link costs for several pieces of information requested for the same connection destination mation transmissions any of the routes suitable for these selected at most once more than the rest of the for these suitable routes - claim 4. This is a largely equal distribution of the traffic load on the appro neten routes causes, which makes the rejection probability is minimized.

According to an embodiment of the Ver driving is provided that with more information about communications to the same information delivery destination suitable routes are numbered in ascending order and cyclically be selected as the route - claim 5. This will a particularly even and fair distribution of ver loads on the appropriate routes.

According to a variant of the method according to the invention the route by randomly selecting one of the suitable routes determined - claim 6. The random selection is exemplary wise with the help of a random number generator particularly effi effect. In addition, there is no prior knowledge General selection decisions are required.  

According to a variant of the method according to the invention provided that the communication network is connectionless - Claim 7. As a result, the invention is advantageous for. B. in IP networks, such as B. the Internet applied.

According to an embodiment of the method according to the invention is selected in the switching node by random selection of a suitable transmission path for the transmission of information the route to a next switching node step by step determined - claim 8. Thus, even in connectionless Networks advantageously a statistical average division of the traffic load.

The method according to the invention is described below with the aid of several figures explained in more detail. It shows

Fig. 1 is a block diagram of a communication network with switching nodes and transmission paths,

Fig. 2 go out in a table all the routes, the switching nodes of the Vermitt K ₁ to the other switch node of the communication network shown in FIG. 1 and valuation rules to iden lung whose route cost,

Fig. 3a in a table the transmission costs assigned link costs, and

Fig. 3b in a table an evaluation of the routes in dependence on the link costs according to the Be valuation rules.

In Fig. 1, a communication network KN with four switching nodes K _i , 1 <= i <= 4 is shown. The switching node K ₁ is connected to the switching node K ₂ by a transmission path U ₁₂ and to the switching node K ₃ by a transmission path U ₁₃ . The switching node K ₄ is connected to the switching node K ₂ by a transmission path U ₂₄ and to the switching node K ₃ by a transmission path U ₃₄ ; A transmission path U ₁₄ is also provided between the switching nodes K ₁ and K ₄ , which is shown in dotted lines in the drawing. This indicates that transmission paths U - for example the transmission path U ₁₄ - may be temporarily overloaded and / or interrupted. Routing information RINF (K _i ) is assigned to each of the switching nodes K _i . An arrow fed to the switching node K ₁ also indicates that this switching node K ₁ receives a request VA for an information transfer V to an information transfer target VZ - for example the switching node K ₄ .

In FIG. 2, the switching node K is shown ₁ overall routing information to RINF (K ₁₎ by way of example. It contains, for example, the routes R _1j leading from the switching _node K ₁ to the switching nodes K _j , 2 <= j <= 4 and their route costs RK (R _1j ). The routes R _1j are defined here as one of possibly several different possibilities, from the switching node K ₁ including the switching nodes K _j , 2 <= j <= 4 and the transmission paths U to the switching destination VZ - in the example the switching node K ₄ - to arrive. In the example, including the optional transmission path U ₁₄ , three routes R _1j-k , 1 <= k <= 3 each lead from the switching _node K ₁ to the switching node K _j , starting from the switching node K _1, the route R _{12- 1} directly, the route R _12-2 via the switching nodes K ₃ and K ₄ and the route R _12-3 via the switching node K ₄ to the switching node K ₂ ; the route R _13-1 via the switching nodes K ₂ and K ₄ , the route R _13-2 directly and the route R _13-3 via the switching node K ₄ to the switching node K ₃ ; the route R _14-1 via the switching node K ₂ , the route R _14-2 via the switching node K ₃ and the route R _14-2 directly to the switching node K ₄ . The routing costs RK (R _1j-k ) of the routes R _1j-k each result from the sum of the link costs L of the respective routes used by the routes U. In this example, it is assumed for the sake of simplicity that all the routes U bi directional and the link costs L are independent of the direction of the information transmission V. In the case of a more differentiated view with unidirectional transmission paths U, two switching nodes K _i , K _{j would} possibly be connected by two transmission paths U _ij , U _ji , to which individual link costs L _ij , L _ji would be assigned depending on the direction of transmission . In this case, the route costs RK of a bidirectional route R would be, for example, the sum of the route costs RK for the unidirectional outward and the unidirectional return direction of the route R, possibly reduced by a correction term. The correction term would e.g. B. take into account that a certain proportion of the route costs RK is counted twice by the summation.

In Fig. 3a on are given for the transmission paths U link costs L. For example, the link costs L (U _ij ) = 1 are defined for the transmission paths U _ij , ij = 12, 13, 14, 24, 34. It should be noted here that the term "link costs" should not be interpreted literally in the sense of "costs". To form the link costs L, any values suitable for the transmission paths U can be used, such as, for. B. Traffic values or quality-of-service values. The choice of all link costs L equal to 1 has the advantage that, when using a Dijkstra algorithm, those routes R have optimal route costs RK that reach their information transmission destination VZ via as few switching nodes K as possible - this optimization metric is also known as "least hops" in the professional world. called. According to this optimization metric, the routes R are preferred which reach their information transmission destination VZ with the least delay times, since the total delay time of a route R is usually essentially determined by the sum of the delay times of the exchanges by the switching nodes K, provided the transmission paths U are terrestrial and are not guided by satellites.

In Fig. 3b, the route costs RK of the routes R _1j-k listed in Fig. 2 are listed as an example, which are determined in accordance with the calculation rules given in Fig. 2 for the determination of the route costs RK on the basis of the link costs L indicated in Fig. 3a have been. Without taking into account the optional transmission path U ₁₄ , the routes R _12-1 and R _{13-2 are} the optimal routes R _MIN with the lowest route costs RK of all routes R. However, they are not routes R (V) suitable for the transmission of information V. For the requested information transfer V to the switching node K ₄ , the routes R _14-1 and R _{14-2 are} suitable routes R (V) and at the same time the optimal routes R _MIN (V), since they compare all suitable routes R (V ) have minimal route costs RK. Taking into account the optional transmission path U ₁₄ , the routes R _12-1 , R _13-2 and R _{14-3 are} the optimal route R _MIN with the lowest route costs RK of all routes R. The routes R _12-1 and R _13-2 are however, no routes R (V) suitable for the transmission of information V. In this case, the routes R _14-1 , R _14-2 and R _{14-3 are} the suitable routes R (V) for the requested information transmission V to the switching node K ₄ . The route R _14-3 is the optimal route R _MIN (V) for the information _transmission V, since it has one hop less than the routes R _14-1 and R _14-2 .

For the exemplary embodiment it is assumed that the switching node K ₁ requests the establishment of an information transmission V to the information transmission target VZ by the request VA. This information transmission target VZ is designed as the switching node K ₄ , the information transmission V thus forms as information transmission V ₁₄ . The appropriate for this information transfer routes R V ₁₄ (V ₁₄₎ are consequently the routes R _14-1, R _14-2 and R _14-3 if necessary.

There are beautiful advantages when using the invention in connection-oriented networks with source routing. In these networks, the information transmission V takes place with the aid of connections established in advance along the selected route R _V , the entire route R _{V being} determined in the switching node K from which the information transmission V is requested - for example in the switching node K ₁ . In contrast to hop-by-hop-based routing, this enables a route R to be determined with minimal route costs RK. On the basis of the link costs L are also in a variant of the invention - z. B. with the help of a Dijkstra algorithm - only the suitable optimal routes R _MIN (V ₁₄ ) be true, ie taking into account the possibly overloaded and / or interrupted transmission path U _14, the route R _14-3 , otherwise the routes R _{14 -1} and R _14-2 , which further reduces the number of suitable routes R (V) to be evaluated by the switching node K ₁ . The route R _V is z. B. determined by overdue selection of one of the suitable routes R (V). Alternatively, the suitable routes R (V) z. B. ascending numbered and cyclically selected as the route R _V for each further requested information transmission V to the same information transmission destination VZ. Following this selection, the establishment of a connection is requested of the switching node K _1, wherein the route R _V is signaled Vermitt all switching nodes along the route K R _V. After establishment of the connection, the requested information is carried transmission V ₁₄ along selected for the transmission of information V ₁₄ Route _V R (V _14).

In line-oriented networks, a largely uniform distribution of requested information transmissions V over several suitable routes R (V), in particular optimal routes R _MIN (V), can thus be achieved, at least on statistical average. A major advantage of this largely uniform distribution can be seen in the fact that, on average, the probability of rejecting connections for several information transmissions that vary in their requested data rate is significantly reduced.

The probability of rejection is advantageously further reduced if a flooding method, for example the PNNI method, is additionally used in the network in order to avoid overloaded and / or interrupted transmission paths, e.g. B. the transmission path U ₁₄ , at least during the period of overload and / or interruption for routing. This is done e.g. B. in that the transmission path U ₁₄ for the duration of the overload and / or interruption compared to the link costs L of the not overloaded and / or interrupted transmission paths U are assigned significantly increased link costs L.

The invention can of course be used in any communication network zen KN, in particular connectionless communication networks KN such as B. the packet-oriented Internet can be used. in the The Internet, for example, runs along every single packet transmitted a packet-specific route R, d. H. the route an information transmission V effected with a packet an information transmission target VZ is independent of the Routes R of the predecessor and successor packets to the same VZ target; the switching nodes K, who are trained as Internet routers, for example average the same information for each package transmission target VZ only the next operator node K - also known as "hop" in the professional world. From each router according to the inventive method successive, to the same transmission of information destination VZ sent packets as a result of a randomly controlled Selection of a suitable transmission path U to a next Most leading to the information transfer target VZ averaging node K possibly on several transmission paths U ver Splits. It is advantageous to use a router closed transmission paths U on average evenly continuous Here, it can e.g. B. by different terms of the individual packages to change the original series follow the packages coming. In this case the  original order of the same information about destination VZ sent packets through a higher proto restoration of the collision. There are several ways to do this known, e.g. B. the Transport Control Protocol TCP.

Claims (8)

1. Method for selecting a route (R _V ) through a communication network (KN) consisting of switching nodes (K) and transmission paths (U), in which

• - Several routes (R (V)) are suitable for transmitting information (V), and
• - The route (R _V ) along which the information is transmitted (V) is determined by at least partially selecting any one of the suitable routes (R (V)).

2. The method according to claim 1, characterized, that the communication network (KN) is connection-oriented. 3. The method according to claim 2, characterized in that the suitable routes (R (V)) are designed as optimal routes (R _MIN ), the identical, depending on the transmission paths (U) assigned link costs (L) determined, minimal route costs (RK). 4. The method according to claim 3, characterized, during a period with unchanged link costs (L) for several to the same information transfer destination (VZ) required information transfers (V) each of these for this suitable routes (R (V)) selected at most once is considered the rest of the routes suitable for these (R (V)).   5. The method according to any one of claims 2 to 4, characterized in that in the case of a plurality of information transmissions (V) to the same information transmission destination (VZ), the routes (R (V)) suitable for this are numbered in ascending order and selected cyclically as the route (R _V ) . 6. The method according to any one of claims 2 to 4, characterized in that the route (R _V ) is determined by random selection of one of the suitable routes (R (V)). 7. The method according to claim 1, characterized, that the communication network (KN) is connectionless. 8. The method according to claim 7, characterized in that in the switching node (K) by random selection of egg nes suitable for the information transmission (V) transmission path (U) to a next switching node (K) the route (R _V ) is determined step by step .