TW462167B - Communications network and method of routing messages within the network - Google Patents

Abstract

A communications network for sending messages between a plurality of nodes with unique identities comprises a server unit. Each node (1) is provided with first storing means for storing a list of all nodes neighbouring said node, which are directed to it and second storing means containing a data field (2) which forms an array (4) of cells (5) indexed for all nodes in the network. A third storing means is provided in the server unit for forming a temporary queue (3). A method of routing messages within the network includes transforming a graph-representation of the network into a tree model with the destination node as root. This tree is the shortest which the graph can be transformed into.

Description

4 6 2 16 7 V. Description of the Invention (1) Technical Field The present invention relates to a communication network including a server unit for transmitting messages between a plurality of nodes having unique identification codes. The present invention also relates to messages in the communication network. The method of routing is related. Technical Background For example, in a packet-switched network, message packets are transmitted by the outgoing node through a number of intermediary nodes before reaching the destination node. The basic problem of this operation is how to shorten the transmission path length between the transmitting and receiving nodes, that is, the shortest path problem. The theoretically best solution to this problem is the algorithm of Dj 1 kstra. However, such an idealized mathematical solution cannot be practically applied in its theoretical form. In the future, the technical solution includes the use of distributed routing tree information, each node of which has a = routing tree representing the shortest path connected to all connected nodes, see US Patent No. 4.987. 536. Other solutions = with central path decision node, see US Patent No. 5.608.721. The solutions provided by Pingxian technology are not very cost-effective. Summary of the invention Early or bidirectional network to any node to the path ', and the main purpose of the invention of the capital stored in each node is to change from one to the other-a cooking node Finding the shortest message is the least. solution

4 6 2 16 7 V. Description of the invention (2)------ Store i in a list; the second store contains-data block η: node identification code; the third misstore device is located at The central unit is used to form a temporary storage queue, which is presented in the renderer of this communication network. &Amp; Μ 扪 雔, 雔 In the embodiment, the data row of the second storage device. This array can be connected to the network. The nodes in the road are indexed. The number of nodes in the array network is proportional and each of its ten bits is counted as an identification code that can store a single node. The distance between groups of points is available-the stated distances are proportional. And the number of virtual nodes

In addition, the negative transference between the two nodes can be represented by a group of hypothetical virtual nodes. The number of virtual transfer points in the card is proportional to the transfer load. The method routes messages in a communication network to transmit messages between nodes with unique identification codes and calculates the shortest path between the first node and the second node. It is characterized by the following steps, including forming an all-disk A list of the adjacent node identification codes related to the first node, the second node ’s unique code is placed in the-queue, a termination instruction is written into a data column to identify the second node as the second node, And write the node's identity = the inscription and the data block of all its adjacent nodes, unless the data is recorded in the identification, whenever there is a record will be written in the node's capital, the identification of the node is exhausted Place it in the queue and move the Z in front of the queue out of the 'After visiting all nodes in the list by all adjacent nodes of that node' repeat the following steps, including selecting the queue = j end Node identification code and write its identification code to all its adjacent nodes = Be • Material 5. Inventory (3) ~~ "--column, until the node visited is the first node. The advantage of the above method is to check the nodes in the network that are not connected to a certain node. It is worth mentioning that the step of the inspection method includes placing the identification code of the node (η) in a queue of 0 and labeling the node, selecting the front-end node of the queue and labeling its adjacent nodes that have not been labeled. Once when a new node, do n’t = accept, after the note, the node will be placed in the queue. After completing the visit to the queue, after all the nodes in the identifier list of all adjacent nodes of the node, the node will The queue is removed and placed into a collection, repeating the steps of selecting nodes from the front of the queue and marking adjacent nodes that have not been annotated until the list is empty, and the generated collection is from all nodes related to the node Subtracting the resulting set, the result is a set of nodes that are not connected to the aforementioned nodes. A method for routing messages in a communication network. It is used to send messages between many nodes with unique identification codes and calculate the shortest path between any two nodes in the network. Its characteristic steps include forming a routing and routing task. — A list of identification codes that are directly transmitted to all nodes adjacent to the node, and put the identification of the node into a queue. 'Write a stop command to this node to an identification code for all nodes in the network. Note the array data for indexing. Take the node at the front of the list and write its identification code in the column of all its adjacent nodes, unless this column has been recorded, whenever a record is written to a node In the data column, the ID of the visited node is placed in the queue, and the node ID at the front of the queue is moved out of 'After all nodes in the list of all adjacent nodes of the node are visited,' it is repeated.

Page 7 4 6 2 1 67 V. Description of the invention (4) The steps include the steps of taking out the front-end node identification code and writing its identification code into the data bar of all adjacent nodes until the queue is empty, and then Repeat for all nodes in the network. The advantage of the above method is that you can check whether a node in the network is connected to other nodes. It is worth mentioning that the steps of the inspection method include: placing the node's identification code in a queue and labeling the node, taking out the front-end node of the queue and labeling its unlabeled neighboring node 1 each time a new label is marked. When the node's identification code is reached, the new node will be placed in the queue. After all nodes in the list of all adjacent node identification codes of the node at the front of the queue have been visited, the node will be taken out of the queue and placed. Into a set, select and repeat the above-mentioned take-out front-end nodes and mark the adjacent nodes that have not been marked until the queue is empty, and subtract the generated set from the set made by all nodes related to the node. The result is a set of nodes that are not connected to the node. BRIEF DESCRIPTION OF THE DRAWINGS The present invention will be explained in more detail in a non-limiting manner with reference to the attached drawings, in which: FIG. 1 is a schematic diagram of the shortest path calculated from any node to another node according to the present invention; 1The queue used when calculating the path. Figure 3 illustrates the calculated shortest path between any two nodes graphically; and Figure 4 is a flowchart illustrating the calculation of the shortest path from all nodes in the network to a node η. Detailed description of the invention

462167 V. Description of the invention (5) The purpose of the present invention is to find the shortest path from any node n1 to another arbitrary node η 2 in a two-way network. Each node must keep the following information: • A "neighbor list" consisting of a node 1 and all its neighbors. This list will be entered into the algorithm. Each node must have its own list. • A data block 2 used to annotate the node identifier "forward to the target" block. This column will be filled in by the algorithm and will contain the information required for routing. Each node must have its own data block Contents. A server unit maintains a π queue " 3, this queue is a temporary storage queue, and will be deleted after the execution of the algorithm is completed. This algorithm is used to calculate from point η 1 to node η 2 The shortest path includes the following steps: • Step 1: Put η2 (not nl) into queue 3. Write a stop instruction, that is, the symbol ", advance to the target at π of node n2" field. • Step 2: Select the node at the front of the queue, and write the identification code of the node and all its adjacent nodes (that is, the nodes in its neighbor list) into the "Go To Target" field list, and write the conditions For this node must have never been written. If the node has already been written, it cannot be written repeatedly. If the visited node is η 1, the algorithm ends. If the queue is empty and node η 1 is not visited, it means that nodes η 1 and η 2 are not connected. • Step 3: Every time we write (and only we can write) to the node's "forward to target" stop, put the node in queue 3 ^ • Step 4: After completing the visit to the node " Neighbor List " All Nodes

Page 9 462167 V. Description of the invention (6) ------ After ’the node is taken out of the queue. • Step 5: Go to step 2. • End When the algorithm has been executed, the shortest path from node “to node n2 is: • Go to node 01 ′, and go to node 1 indicated by column 2 in the target □ • Keep reading the same way until I reach node n2. The algorithm described above actually transforms the graphical representation of the network into a tree-like module with the destination node as the root. The tree transformed from the graph is the shortest (shortest (M) shows the results produced by the algorithm after the shortest pass from node i to node 8. Figure 2 shows the queue used in the process to find the shortest path. This algorithm can also be used to view all nodes that are not connected to node n. Each time when a node is taken out of the queue, 'the node must be placed in a " visited node " set: Step 丨: Node η Place it in queue 3 and mark the node. • Step 2. Select the front-end node of the queue and mark the adjacent nodes that it has not yet marked. Step 3. Each time you mark a new node, put it in the queue. 〇 • Step 4. After completing the manufacturing The node (on the front end of queue) of " neighbor list " after all the nodes, the node to remove the section ,, have visited Step 5: Step 2 Aberdeen column 3 until point set is empty °.

4 62 16 7 V. Description of Invention (7) • End. When the implementation of this algorithm is completed, it will be "Mouth: All the nodes known by Gan Minglang Queren" minus its set of known knowledge points ", and the result is the stuffing Machine ^ point set. 丨 Horse and knowing point η are not connected. When a node wants to send a message to the target 铕 椹 ^, L, Sun knowing point (node 8 in Figure 1), this root node is only one earlier. The path goes back up. This path is true. Let the horse take a short path, as shown in Figure 1: The first node n8 is placed at the root (the section of the bismuth shovel branch J is placed from the point A to the point X). 'Then point n4 is placed under node n8, n4i? LaA ,, .... 々n4 is the only possible choice (after Ning Γ node n8), and then the node "乂 = below (in order 仵In the column), finally, place the node nj 5 below the Lang point η6 'and place the knot blessing below the node center (place :: "in the queue). #Then, the shortest path from node to node ⑽ is: η 1 to η 5 to η 3 to η 4 to η 8. The present invention can also find the shortest path from any point 111 to another arbitrary node η2 in a two-way network with minimum information. Each node should contain the following information: • A node's "neighbor list" consisting of all its neighboring nodes, and the "neighbor list" of all nodes depicts the topology of the network. This list is entered into the algorithm, and each node must have its own list content. • A "forward to the target" array data block 4 is used to index all node notes in the network. The size of the array should be N. There must be enough space for each grid 5 in the array to store the identification code of a node 1 (such as a number), for example: an array of integers [1 ... N]. This data block will be filled by the algorithm and will only contain the information needed for routing. Each node must have a property

Page U 4 6 2 1 6 7 V. Description of the invention (8) In his own data block. A server unit maintains a "queue" 3, which is a temporary queue and will be deleted after the algorithm is executed. Figure 4 illustrates that the algorithm is used to calculate all nodes in a network. The shortest path, and each node η (η = 1 * ·· N) contains the following steps: • Step 1: Place η in the queue and write a stop instruction, that is, the symbol M-"in the node" Go to the target [η] in the field. • Step 2: Select the node at the front of the queue and write its identification code to all of its neighboring nodes (nodes in its neighbor list). η] π In the list, the write condition must be that the grid has not been written. If the grid has written data, it cannot be written repeatedly. • Step 3: Every time I visit a node and write (Only write) When the "Go Forward to Target" field of the node, place the visited node identifier in the Suining column. • Step 4: After all nodes of the node's neighbor list π have been visited, remove them from the queue. Step 5: If the queue is empty, node η is completed, otherwise go to step 2. End. *) In Figure 4, the word "unvisited" means a node which "goes forward to the target [η Γ has not been written, where η represents the first node placed in the queue. The resulting inner loop is the shortest path from all nodes to node n. The resulting outer loop is the shortest path from all nodes to all other nodes. When the algorithm is completed and the network information has been updated to obtain

Page 12 462167

When the shortest path from the point n 1 to the node η 2 is: • Read the node nl's "forward to the target" block [η2] and call it ρ., Advance to ρ, and then visit its "forward to the target" 襕Bit [η2] holds the same steps until node η2 has been reached. "The content of bit [η2] before the target of node [η2] is a sign". When the inner loop, the shortest path to a node will be calculated. We can check whether this node is connected to all other nodes (these nodes must be known by 诙). To achieve this, each time a node is removed from the column, the node must be moved into a collection. When the part of the algorithm 'point η has been executed, all the nodes known by the node will be subtracted from the previous set, and the output will be a sum node + connected node 0 union. · Figure 3 shows the results of the algorithm after updating all nodes' to the target bit list. While the marriage information is being transmitted, each node only needs to save a minimum of moving forward to the target "field area or list. ) ^ The above algorithm is used in "optimization", and is the same as Distra's algorithm in the calculation of the two = most path. Algorithms always end execution. Μ are: i. In the performance algorithm, the assumption is that the load on the interesting link between all adjacent nodes can be interpreted as two adjacent nodes (the length between the links. Low load can be interpreted as between nodes The shorter the distance is, the longer it is. &Amp; < The person with j Q can assume a group of virtual nodes at each pair of physical nodes.

Page 13 462167 V. Description of the Invention (ίο) Therefore, considering the different distances between two nodes in the algorithm, the virtual node depends on the distance between the two nodes. For example, if the distance between node 1 and node 5 is 3 meters, we can assume that there are two virtual nodes (called nodes 1 51 and 1 52) in between. Once these virtual nodes (representing the distance between any two adjacent nodes) have been placed in the graph, the 'algorithm can be executed in the normal situation. The only difference is that the "Forward" field of each node must be updated according to the identifier of the physical node closest to the specified virtual node. When the data structure is completed, the virtual node data generated for calculation purposes will not be known. Even though the network is bidirectional in general, the algorithm can be applied to unidirectional circles with some modifications. The difference is that the nodes in a unidirectional graph may not know which nodes are directly passed to it, and the node may only know which nodes it is straight (as in the " adjacent list ', above). To overcome this problem, one called " direct to my list " will be attached to each node. This array contains the identification moms of all neighboring nodes that are directly passed. This array can be removed after the algorithm execution is completed [eg: for any other reason for retention. The array: calculated from the "neighbor list" attached to the network nodes, such as the following calculation = • Start • Step 1: Place any node in the queue. • Step 2: Select the end node (node n) and (〇 Put it in the direct transmission (in its 1-home list " node " directly into the list ". Every time you go to a sergeant) Tian Tiandian, direct transmission

4 6 2 1 6 7 Case No. 88120779 Same month i. Amendment Five 'invention description (11) My list " When updating, check whether the node has not been marked. If it was annotated, label the node and place it in the back end of the queue (if the section was previously annotated, it cannot be placed in the queue). • Step 3: Remove node η from the queue. If the queue is not empty, go to step 2. •End. After the execution of this algorithm is completed, each node in the graph will have the information that those nodes send to that node (" directly sent to my list "). Once the `` direct transmission to my list '' of these nodes in the graph has been calculated, the shortest path of all nodes in a single graph can be obtained by referring to the same algorithm as the two-way network case. The only difference is that the Use "quote directly to my list" instead of "neighbor list in a two-way network" to the network. It is certain that the present invention is not limited to the above specific embodiments, as the present invention can be Various changes are made within the scope of the patent application. Component symbol description 1 Node 2 Data block 3 Temporary queue 4 Array 5 cells to the un-pointed row inward increase the annoyance | | > 1 3 mentioned

O: \ 61 \ 61547.ptc Page 15 2001,05.02.015

Claims (1)

462167 _ Case No. 88120779 VI. Scope of Patent Application Years & Months> Revision 1. A communication network for transmitting messages between a plurality of nodes (1) with unique identification codes, the network includes a server unit, which It is characterized in that each node (1) has the following storage device: a first storage device for storing a list of neighboring nodes directly transmitted to the node; a second storage device including a data column (2) for An identification code of a node is noted; and the third storage device is placed in the server unit to form a temporary storage queue (3). Device Dragon ^ Please refer to the Internet Heart Committee a a member of the jujube 疋 疋 appeal ^ 0 ^ 3 · year node ^ the number I ί;] Μ .: 1: ¾ between the repair center 4 mention points 5. As applied The ratios in the assets such as the sum of Shen Jian and the distance between Ru Shen Transmission Project and Ru Shen Loss Please refer to the special shed. ○ Please distance away, please load the pass, and please include the number of the pass and The transmission 6 is a communication network of the first item in the Yili range, in which: the second storage forms an array (4) of a grid (5), which can be used to index the points, the size of the array and The number of nodes in the network, and each grid position (5) is used to store the communication network of item 1 or 2 of a node. Among them: any pair is represented by a group of hypothetical virtual nodes, and the proportion of virtual nodes. The communication network of item 1 or 2 of the profit range, where: the two nodes are represented by a group of hypothetical virtual nodes, where the virtual node load is a 'proportional β communication range of the item 3 of the profit range, where: A group of hypothetical virtual node representations, where the virtual node load is proportional. Send a message to a group of nodes with a unique identifier (η 1 O: \ 61 \ 61547.ptc Page 1 2001.05.02.017 462167 _Case No. 88120779 f0 Death Month > Day Amendment_ VI. Patent Application Scope, η 2, ...) to route messages within the communication network A method for calculating a shortest path between a first node (η 1) and a second node (η2), which is characterized by the following steps: forming a list of identification codes of all neighboring nodes directly transmitted to the first node; Put the identification code of the second node (η2) into a queue (3); write a stop command in the data column (2) to note the identification code of the second node (η2); select the node from the front of the queue And write its identification code into the data column of all adjacent nodes of the node, unless the data block of the adjacent node has been written; each time when the data block of a node is written, the node The identification code of the queue is placed in the queue; after visiting all the nodes in the list of all adjacent nodes of the node, the identification code of the node at the front of the queue is removed from the queue; and repeating from the front of the queue Select the node ID and identify it Bar code data writing step of all its neighboring nodes, the nodes until the first node visited so far (η 1). 7. The method as described in item 6 of the patent application scope, includes inspecting nodes that are not connected to a node (1) in the network. 8. If the method of applying for item 7 of the patent scope *, the review process includes the following steps: Place the identification code of the node (1) in the queue (3) and mark the node 0: \ 61 \ 61547, ptc Page 2 2001.05.02.018 4 6 2 16 7 _ Case No. 88120779_fp year 2nd month_ί ± ^ _ VI. Scope of patent application Select this node from the front of the queue and mark it as not affected Labeled neighboring nodes: Each time a new node identifier is labeled, the node is placed in the queue; after all nodes in the list of identifiers of all adjacent nodes in the queue's front-end node are visited, the node is automatically identified. Remove the node from the queue and place it in the set of visited nodes; Repeat the steps of selecting the front-end node of the queue and labeling its adjacent nodes that have not been labeled until the queue is empty; and know the node (η) The set of all nodes minus the generated set of visited nodes. The result is a set of nodes that are not connected to the aforementioned node (η). 9. The method according to any of claims 6-8 of the scope of patent application, further comprising the following steps:-expressing the distance between each pair of nodes and / or transmission load by a group of hypothetical virtual nodes, the group The number of virtual nodes is proportional to the distance and / or transmission load. 1 0. — A method for routing information in a communication network for transmitting messages between a plurality of nodes with unique identification codes, used to calculate the shortest path between all two nodes in the network, which is characterized by the following steps: Forming a list of identification codes of all neighboring nodes directly transmitted to the node; placing the identification codes of the node (1) in a queue (3); O: \ 61 \ 61547.ptc Page 3 2001.05. 02.019 462167 _ Case No. 88120779 amended on the following day of the month_ Sixth, the patent application scope stop order is written in a grid (5), the grid is located in the array (4) The data block has its index code [η] to index the identification code notes of all nodes (1) in the network; select the identification code of the queue front-end node and write the identification code into the column of all its adjacent nodes In the grid (5) of the internal index [η], unless the data block has been previously recorded; each time a record is written into the grid of [η], which is searched by the data column of a node (1) (5), the identification code of the visiting node is placed in the queue; after visiting all the nodes in the list of all adjacent nodes of the node, the identification code of the front-end node of the queue is removed from the queue; Repeat the steps of selecting the node identification code from the front end of the queue and writing its identification code in the data block of all its adjacent nodes until the queue is empty; and repeat the above steps for all nodes in the network. 1 1. The method according to item 10 of the scope of patent application includes checking whether a node in the network is connected to other nodes (η 1, η 2, ...). 12. The method according to item 11 of the patent application scope, wherein the review process includes the following steps: placing the identification code of the node (1) into the queue (3) and marking the node; selecting the front-end node of the queue and marking it has not been affected. Labeled adjacent nodes Each time when a new node's identification code is marked, the node identification code will be placed in the queue; after all nodes of the identification list of all adjacent nodes of the node at the front of the queue have been visited. , Remove the node from the queue and place it in a created O: \ 61 \ 61547, ptc Page 4 2001.05. 02. 020 4 6 2 16 7 _ Case No. 88120779 VI. The scope of application for patents is revised within the set of visited nodes; repeated selection of nodes from the front end of the queue And labeling its adjacent nodes that have not been labeled until the queue is empty; and subtracting the generated set of visited nodes from the set of all nodes known to that node (η), the resulting result is the same as that of the node Disconnected nodes (η 1, η 2, ...). 43. The method according to any one of claims 10-12, further comprising the following steps:-the distance and / or transmission between each of the two nodes The load can be represented by a group of hypothetical virtual nodes, and the number of virtual nodes in the group is proportional to the distance and / or transmission load. 2001.05. 02. 021