DE102006007070B3 - Method of determining transmission paths or routes of a communication network transmitting data packets and having several nodes with bidirectional ports - Google Patents

Abstract

The method involves each node initiating a topology packet containing that node's address and outputting this to each port. A node that receives a topology packet checks it for its own address and if not present adds it and sends it to the remaining ports. The address of the node that initiated that packet is also determined together with a hop number proportional to the number of addresses entered. These are stored with an identification code in a table and sorted according to the initiating node and then hop number so that for a node address, the port with the lowest hop number to this node can be derived from the table. Independent claims also cover network nodes.

Description

The The invention relates to a method according to the preamble of the claim 1 and a network node for transmitting a data packet Communications network.

dates are now being inserted into packets and transmitted using data packets Transfer communication networks. Such communication networks are for example networks of synchronous Digital hierarchy, short SDH, of the synchronous optical network, short SONET, the optical transport network, briefly OTN, Internet protocol based networks, in short IP networks, so-called OSI networks, Ethernet networks, ATM networks as well as other data packet transmitting networks. The communication networks have several network nodes, each one unique Node address is assigned and their ports bidirectional with each other are connected. The connections can be used as lines such as Two- or multi-wire cables, coaxial cables or optical fibers accomplished be. The communication network can be at least partially meshed be. The network nodes can Network elements, routers / layer 3 routers, switches / layer 2 switches, bridges, Hubs, gateways or similar be. Similarly, a network node may be part of a router or switch exhibit. For example, by means of a in a network node integrated router formed a subnet in the communication network for administrative support of the administration of the network. So can with the subnet mainly administrative or control information are processed. For example, a router or a message communication Functions can be arranged in an SDH / SONET or OTN network node.

Everyone Network nodes in the communication network require an overview of the topology or transmission paths in the communication network. This is necessary to forward data packets properly to be able to. In this case, an information is stored in the network node, which Network node via which Port or the connected transmission path respectively Connection fastest, shortest or according to other criteria is to achieve the most optimal.

router or network nodes with functions according to layer 3 of the OSI model use often the so-called shortest path algorithm. This will be the port with the shortest Path to a network node detected and stored. An example therefor is the so-called OSPF protocol (Open Shortest Path First). This is a so-called link state protocol. Here are so-called link Status information of the connection between two network nodes generated and this information is distributed in the system. The link status information are used with so-called hello packets or protocol data units, short (Hello) PDUs. Each network node determines link status Information of his connections and distributes these by means of such called Link State PDUs. Each network node evaluates the link state PDUs of the other network nodes, determined from a topology of the communication network and calculated using the so-called Dijkstra algorithm the shortest transmission paths to every network node.

The OSPF protocol is event-driven. A change in the topology of the Network, for example if a port or a connection respectively a link fails or disturbed is detected by a (neighbor) network node. This information is distributed via Link State PDUs in the network. Each network node leads to it a recalculation of its transmission paths by.

One Switch uses often no explicit method for determining transmission paths. Often it will used a method in which detected at the ports (MAC) addresses and saved. Thus, data packets become known addresses only on the respective port and unknown addresses on all ports sent out. This has the disadvantage that before handensein of Looping data packets infinitely in the network circling.

Around Avoiding sanding is common uses the so-called spanning tree protocol. This protocol The task is to recognize grinding, an optimal, balanced Find tree with short paths in the communication network and loops break up by breaking individual links.

In US Pat. No. 6,147,971 is an optimized routing method using a minimum hop number for use in PNNI (Private Network Node Interface) ATM (Asynchronous Transfer Mode) Networks described.

In US Patent Publication 2003/0225916 A1 discloses the implementation a data link layer protocol in a network node with several Network interfaces or devices described.

task The present invention is an easy way for the determination of transmission paths specify.

These The object is achieved by a method having the features of the patent claim 1 and a network node with the features of claim 9 solved.

• – einerseits das Topologie-Paket mit seiner Netzknotenadresse erweitert und dieses an seinen verbleibenden Ports ausgesendet, d.h. an allen Ports außer dem das Topologie-Paket emp fangenden Port, und
• – andererseits aus dem empfangenen Topologie-Paket die Netz knotenadresse des dieses Topologie-Paket initiierenden Netzknoten und eine zur Anzahl der eingetragenen Netzknotenadressen proportionale Hopzahl ermittelt und beide mit einer Kennzeichnung des dieses Topologie-Paket empfangenden Ports in einer Tabelle gespeichert. Diese wird zuerst nach den Netzknotenadressen der Topologie-Pakete initiierenden Netzknoten und dann nach deren Hopzahl sortiert. So kann aus der Tabelle für eine Netzknotenadresse der Port mit der kleinsten Hopzahl zu diesem Netzknoten ermittelt werden.

According to the invention, it is proposed that each network node ent its own network node address ent initiate holding topology packets and submit them to each port. A network node receiving such a topology packet on a port checks this for its own node address. If no own node address is missing:

• On the one hand, the topology packet with its network node address extended and this sent out on its remaining ports, ie on all ports except the topology packet emp catching port, and
• On the other hand, from the received topology packet, the network node address of the network node initiating this topology packet and a hop number proportional to the number of registered network node addresses are determined and both stored in a table with an identification of the port receiving this topology packet. This is sorted first according to the network node addresses of the topology packets initiating network nodes and then by their hop number. Thus, from the table for a network node address, the port with the lowest hop number for this network node can be determined.

Of the Advantage of the method is that a simple way for the determination of transmission paths given is.

advantageous Embodiments of the invention are specified in the subclaims.

In an advantageous embodiment of the invention are received Topology packets containing their own node address are discarded. This has the advantage of avoiding infinite circling topology packets become.

In a further advantageous embodiment of the invention is in the multi-column table in the first column Network node address of the network node initiating a topology packet, in the second column the hop number, in the third column the marking of the port receiving this topology packet and in the fourth Column entered the reception time of the topology package. This has the advantage that the reception time for further evaluations for disposal stands.

In an advantageous embodiment of the invention are entries in the Table exceeding a certain time period deleted. This has the advantage of having old entries or entries which were not renewed were deleted and the table is kept up to date to misinsert data packets on failed connections to avoid.

In An advantageous embodiment of the invention is a topology package generated regularly. This has the advantage of having failed connections or ports by no longer transmitted Topology Packages, resulting in entries in the table by deleting not updated entries be removed, detected and incorrect transmissions avoided.

In An advantageous embodiment of the invention is a topology package after a change generated in the network. This has the advantage of making network changes faster be recognized.

In An advantageous embodiment of the invention are in the table a network node compared the network node addresses. With agreement two network node addresses, an information is given or stored, that the communication network has a loop. This has the Advantage that loops are detected in the communication network. These Design is especially for Ethernet networks with MAC address "switching" of importance.

In An advantageous embodiment of the invention are in the table a network node compared the network node addresses. With agreement two network node addresses become their port identifiers compared. If there is no match an information is issued or stored that the network node in a loop. This has the advantage of being recognized whether a network node is in a loop.

In An advantageous embodiment of the invention is in a network node a proportional to the number of accumulated hops of the table priority of the network node determined. This priority of the network node is in the inserted topology packets of the network node. One in a loop located network node compares its priority with the transmitted priorities the other nodes in the loop. If his priority the highest is, becomes a port of the network node in the loop disabled. This has the advantage that the loop is split is avoided and so.

One embodiment The invention will be explained in more detail below with reference to the drawing. there shows:

1 a schematic representation of a communication network,

2 a table of a network node.

1 shows a communication network, comprising 14 network nodes Adr1, ..., Adr14. The ports of the network nodes are connected to each other bidirectionally. Network node Adr1, port 2 is with network node Adr2, Port 1 connected. Its port 2 is in turn connected to port 1 of network node Adr5. Its port 2 is connected to port 1 of the network node Adr12. Its port 2 is connected to port 1 of network node Adr11. Its port 2 is connected to port 2 of network node Adr10. Its port 1 is connected to port 2 of network node Adr9. Its port 1 is connected to port 2 of the network node Adr8. Its port 1 is connected to port 2 of the network node Adr7. Its port 1 is connected to port 2 of the network node Adr6. Its port 1 is connected to port 2 of the network node Rdr4. Its port 1 is connected to port 2 of the network node Adr3. Its port 1 is connected to port 1 of the network node Adr1. Furthermore, port 3 of the network node Adr5 is connected to port 3 of the network node Adr6. Network node Adr12 Port 3 is also connected to port 1 of the network node Adr13, whose port 2 is connected to port 1 of the network node Adr14.

Everyone Network node has its own unique network node address in the communication network. This For example, an IP address, MAC address, NSAP address or be another address. In the example, the marking of the network nodes Adr1, ..., ADr14 be equal to the node address, so also Adr1, ..., Adr14.

Everyone Network node initiates topology packets or topology PDUs with their own Node address and returns or sends to each port every port. these can correspond to the following pattern: "Topology PDU: == Adr1", for network nodes Adr1.

• 1.) einerseits das Topologie-Paket mit seiner Netzknotenadresse erweitert, beispielsweise an das Ende des Paketes die eigene Netzknotenadresse hinzugefügt, und dieses an seinen verbleibenden Ports ausgesendet, d.h. an allen Ports außer dem das Topologie-Paket empfangenden Port, und
• 2.) andererseits aus dem empfangenen Topologie-Paket die Netzknotenadresse des dieses Topologie-Paket initiierenden Netzknoten und eine zur Anzahl der eingetragenen Netzknotenadressen proportionale Hopzahl ermittelt und beide mit einer Kennzeichnung des dieses Topologie-Paket empfangenden Ports in einer Tabelle gespeichert. Diese wird zuerst nach den Netzknotenadressen der Topologie-Pakete initiierenden Netzknoten und dann nach deren Hopzahl sortiert. So kann aus der Tabelle für eine Netzknotenadresse der Port mit der kleinsten Hopzahl zu diesem Netzknoten ermittelt werden.

A network node that receives a topology packet at a port checks this for its own node address. If no own node address is missing:

• 1.) on the one hand, the topology packet with its node address extended, for example, added to the end of the packet's own node address, and this sent out at its remaining ports, ie all ports except the topology packet receiving port, and
• 2.) On the other hand, from the received topology packet determines the network node address of this topology packet initiating network node and a number of registered network node addresses proportional hop number and stored both with a label of the port receiving this topology packet in a table. This is sorted first according to the network node addresses of the topology packets initiating network nodes and then by their hop number. Thus, from the table for a network node address, the port with the lowest hop number for this network node can be determined.

Example for 1.):

node Adr2 initiates a topology packet with its own node address Adr2, for example: "topology PDU: == Adr2 "This Topology package is delivered at ports 1 and 2. node Adr5, whose port 1 is connected to port 2 of network node Adr2, add this Topology packet adds its own node address Adr5 at the end, For example: "topology PDU: == Adr2, Adr5 ". This so-extended topology package is provided by network node Adr5 delivered at its ports 2 and 3. It will not work on port though 1 because the topology package was originally received here.

From Port 2 arrives at port 1 of the network node Adr12 and from port 3 to port 3 of the network node Adr6.

Here it is processed in an analogous way and so each network node reach the communication network.

By the entries in the topology packet, any connection between two network nodes be recognized.

As already written, each network node checks a received topology packet for its own node address. A network node will receive its own topology packet if Loops are present in the communication network. An information, that a loop in the communication network is available, can be evaluated the received topology packets are performed, and upon receipt a self-initiated topology packet, i. with the own network node address, a corresponding loop information issued or stored become.

received self-initiated topology packages are not forwarded and discarded to avoid circling these in the network.

in the For example, node Adr2 will have its own topology packet that is on the port 1 was sent, received at port 2 and vice versa.

Example for 2.):

One received topology package is used for a table. The Table has several columns, where in the first column Network node address of the network node initiating a topology packet, in the second column the hop number, in the third column the marking of the port receiving this topology packet and in the fourth column the reception time of the topology packet is entered.

In 2 is such a table, in the example for the network node Adr5 shown. The first column for the node address is marked 'Address'. The second column for the hop count is 'Hops', the third column for the receiving port is 'Port' and the fourth column for the receive time is marked 'Refresh time'.

Out a received topology packet that does not have its own node address contains becomes the node address of the topology packet initiating Network node determined and entered in the column 'Address' in the table. Farther becomes a proportional to the number of registered network node addresses Hop count determined. This can be the number of registered network node addresses be. The hop number is entered under 'Hops' and stands for the distance or the number of jumps to the initiating network node.

Of There will also be a label for the port on the topology package received was registered under 'Port'. The reception time of the topology packet is entered under 'Refresh time'.

2 shows a part of the entries for network node Adr5 according to the communication network 1 ,

These Table will be / was first after the node addresses and then after sorted by the hop count, so for each node address continues to be the least hop port in front. This port is the path with the least number of intermediate ones Network nodes.

In A meshed or looped net will be more as one entry per node address. The first one Entry gives the port with the least number of intermediate ones Network node.

By The marking of the reception time can be entries in the table, the one exceed certain time span, deleted become. These are the entries checked regularly in the table and if necessary deleted. Thus, the table is always kept up to date.

For example The network nodes can generate a topology packet every 10 seconds become. Then the table is checked regularly and entries that have not been updated for 20 to 30 seconds will be deleted.

It is an advantage for that a network node address are different entries in the table, the thus checked separately and at timeout individually deleted become. Thus, there are often replacement routes to disposal, if the primary Way, with the smallest hop count, has failed.

If the topology packages more often can be generated faster, the network can change or failed Connections react. On the other hand, it increases the computational effort because of the higher Number of packets to be evaluated.

If less often Topology packets are generated, the network takes longer to with changes to adapt or to become stable.

which Time behavior is always used, it should be in all network nodes be equal.

The Topology packages can generated both regularly and irregularly become. Likewise, a network can be realized that is event-controlled works and generates topology packets after changes in the network. In this case, the deletion can the entries a period of time, i. the fourth column in the table Refresh time is not mandatory. Posts in the table are then evaluated after evaluation Topology package possibly deleted.

The inventive method is capable of loops in the net by evaluating the table too detect. For this purpose, in the table of a network node, the network node addresses compared with each other and match two network node addresses the communication network has a loop and gives a corresponding one Signal off or stores it.

Also can be determined if a network node is in a loop located. For this purpose, in the table of a network node, the network node addresses with each other compared. With agreement two network node addresses whose port identifiers are compared. If there is no match an information is given or stored that the Network node is in a loop.

Furthermore, loops in the network can be meaningfully separated in certain network configurations. For this purpose, a priority proportional to the number of accumulated hops of the table is determined in the network node, which is inserted into initiated topology packets, for example at the beginning of each topology packet. A looping network node compares its priority with the transmitted priorities of the other loop nodes in the loop. If its priority is the highest, one of its in-loop ports will be disabled. This prevents a loop in the network. This procedure leads primarily to networks with branches, stubs or parti eller meshing to success.

The transmitted priority of the initiating network node can be entered in a fifth column of the table so that this information is virtually permanent in the network node is present. Similarly, information that is a network node in a loop, to be inserted in topology packages. This information may also be in a sixth column of the table be registered. So is a simpler evaluation by query the table possible.

The inventive method settles in a network node for transmitting a data packets Integrate communication network. This includes funds that way that they are the method or parts of the described Perform the procedure can.

Of the Another advantage of the invention is that a simple method for determining the transmission path indicated with the least number of intermediate network nodes is. For this purpose, in one embodiment, only 2 time values are correct Setting needed. The first time value is the time constant for the regular transmission of the Topology packages. The second time value is the verification time the table for the deleting of the entries.

Claims (9)

Method for determining transmission paths of a data packet-transmitting communications network, comprising a plurality of network nodes each having a unique network node address and whose ports are bidirectionally connected to each other, characterized in that each network node initiates the topology packets containing its own network node address and delivers it to each port, a network node which receives a topology packet at a port, checks this with regard to its own network node address and, in the absence of its own network node address, on the one hand broadens the topology packet with its network node address and transmits it to its remaining ports and, on the other hand, uses the received topology packet Node address of the network node initiating this topology packet and a hop number proportional to the number of registered network node addresses and both identified with an identification of the port receiving this topology packet in a T teller, which is first sorted by the node addresses of the topology packets initiating network node and then by their hop number, so that is determined from the table for a network node address of the port with the smallest hop number to this network node. Method according to claim 1, characterized in that that received topology packets that own network node address contain, be discarded. Method according to claim 1 or 2, characterized that the table has multiple columns, where in the first column the node address of the network node initiating a topology packet, in the second column the hop number, in the third column the marking of the port receiving this topology packet and in the fourth Column the receive time of the topology packet is entered. Method according to claim 1, 2 or 3, characterized that entries in the table that exceed a certain period of time. Method according to claim 1, 2, 3 or 4, characterized that a topology package generates regularly becomes. Method according to claim 1, 2, 3 or 4, characterized that a topology packet is generated after a change in the network. Method according to one of the preceding claims, characterized characterized in that in the table of a network node, the network node addresses compared with each other and match two network node addresses the communication network has a loop. Method according to claim 7, characterized in that that in the table of a network node, the network node addresses with each other be compared, that at agreement two network node addresses whose port identifiers with each other be compared and in the absence of agreement information is issued that the network node is in a loop. Network node for transmitting a data packets Communication network comprising means designed in such a way that it is a method according to one of the preceding claims 1 to 8 execute can.