KR100421414B1 - Method for searching and updating using node structure of multiway search tree - Google Patents

Abstract

1. TECHNICAL FIELD OF THE INVENTION

The present invention relates to a tree search and update method using a node structure of a multiple search tree and a computer readable recording medium having recorded thereon a program for realizing the method.

2. Technical problem to be solved by the invention

In the present invention, regardless of the number of keys used by a node in a multiple search tree, only one pointer is written to a node so that only one node pointer, the number of keys, a root port, a key, a 2-bit flag, a port, and an area port are used. To match the cache line size to speed up the multiple search tree, increasing the number of keys that can be stored in the node in network address searches, to speed up the search, and to enable maximum valid matching in the bittree. It is an object of the present invention to provide a method of searching and updating a tree using a tree structure of a tree and a computer-readable recording medium recording a program for realizing the method.

3. Summary of Solution to Invention

The present invention searches for an initial bit array using an upper bit segment of a destination IP address, moves to a root node indicated by a node pointer, and searches for multiple paths with a lower bit offset of the destination IP address to reach a leaf node. A first step of determining whether Exact Matching has been performed before; A second step of returning the corresponding port and ending the lookup if the match is complete before the leaf node arrives as a result of the determination of the first step; And a third step of returning or storing the area port of the key in a register according to the node where the search has reached, if the result of the determination of the first step is not complete match before the leaf node arrives.

4. Important uses of the invention

The present invention is used for address search of routers and switch systems

Description

Method for searching and updating using node structure of multiway search tree}

The present invention relates to a tree search and update method using a node structure of a multiple search tree. More particularly, the present invention relates to the number of keys that can be stored in a node by using only one pointer number required in a node of a bittree. Retrieval time by reducing the number of levels of tree paths by increasing the number of tree paths. A computer readable recording medium having recorded thereon a program for realizing the method.

In the structure of a general gigabit router system as shown in FIG. 1, an interface performs a function of matching between a form for using a packet inside the system and a form of data to be used in a link, and forwarding. The forwarding engine (FE) 14 classifies and processes the received packets. In other words, the forwarding engine 14 determines where the destination of the packet is and decides on which link to send it. Then, the switching fabric 12 performs a function of sending data to the actual link by using the information determined by the forwarding engine 14.

The address searching method includes an exact match method and an LPM (Longest Prefix Match) method according to how an address is used. The perfect matching method is used when the bits of all addresses match, such as the Ethernet MAC address, and the maximum valid matching method is to find the longest matching data starting from the first bit with the key to be searched for. Is used).

2 illustrates a data structure according to a general cache line.

That is, Figure 2 shows a multiple search tree with a 32-byte cache line used by Lampson. Each node consists of keys K1, K2 ... and pointers P0, P1, P2, ..., and each node consists of 32 bytes.

The reason why Lampson uses multiple search trees is to make the most of the hierarchical memory structure that modern processor structures generally take. In other words, in order to overcome the ever-increasing memory speeds and processor speed differences, an increasing amount of caches are installed in the processor. Therefore, when an arbitrary memory address is accessed during the operation of the cache, the entire cache line including the memory address is processed at the same time. Is copied into the cache and the data in the cache is processed at the processor's speed. Therefore, by making one node of the multiple search tree one cache line size, the number of branches can be greatly increased without additional main memory access.

Since the above-described multiple search tree is designed in consideration of the memory hierarchy between the disk and the main memory, the disk block read from the disk into the main memory at one time in consideration of the low speed memory and the relatively high speed main memory is used. Accordingly, the number of branches must be increased to the speed of main memory without additional slow disk access. A representative example of such a multiple search tree is a tree structure of a bit tree, which is described later in FIG. This will be explained in detail later.

3 is a diagram illustrating a structure in which three IP paths are extended to an IP region.

As shown in FIG. 3, the IP paths in the routing table are represented by a Route Prefix and a Prefix Length. In order to perform the maximum valid matching method using the Btree on the IP path, an additional function is required. First, assuming three IP paths limited to 6 bits, three IP paths of 1 *, 101 *, and 10101 * should be represented by IP pairs in order to express corresponding IP ranges. That is, in the case of 1 *, '100000' (Start) and '111111' (End) are represented, 101 * is '101000' (Start) and '101111' (End), and 10101 * is '101010' (Start ) And '101011' (End).

That is, the destination IP address input between each IP path corresponds to the range of the corresponding IP path. For example, if the destination IP address input in the lookup is '101110', it corresponds to '101000' and '101111'. These pairs of IP paths are keyed into the multi-path tree and must be updated identically.

4 is a diagram illustrating a situation when adding a new prefix in the structure extended to the IP region illustrated in FIG. 3.

That is, FIG. 4 shows matters to be considered when a prefix queue (Prefix Q) is newly added in an area consisting of IP path pairs, and a prefix that is within this range by addition of a prefix queue (Prefix Q) is shown. These are affected by the update of the prefix queue (Prefix Q).

Unique Prefix is a 32-bit prefix and is marked as Start and End at the same time, and means an IP path having no area. The addition of the prefix queue (Prefix Q) adds an area between queue start (Q Start) and start 1 (Start (1)), between end 1 (End (1)) and start 3 (Start (3)). Prefix queues (area C) between region (B), region (C) between End (3) and a single prefix, and region (D) between a single prefix and Q End. Prefix Q) is further updated simultaneously.

In the case of multiple navigation trees, each key has a port number that is used for exact matching and an area port number that represents a larger area than itself, so in this case the keys that need to be modified are Q Start and Queue. The ports and zone ports of Q End and the zone ports of End 1 (End (1)), End 3 (End (3)) and Unique Prefix should be updated.

5 is a diagram illustrating a tree structure of a general bit tree.

As shown in Fig. 5, the bit tree (Btree) is an improved version of the binary tree, and is mainly used for algorithms for searching or searching for data. That is, the Btree is compared with the root node and divided into left and right according to its size is the same as the binary tree, but the next node is faster by comparing several at once according to its size and the depth of the tree. There is an advantage to reduce.

However, in the general case where Btree is used, it is also used to configure the index file stored on disk to access the database stored on disk. The key for accessing this database is composed of a string and the result of comparison in this node. As a result, the problem of using a lot of memory compared to the length of the pointer to point to the next node to be connected.

Accordingly, the present invention has been proposed to solve the above-mentioned problems, and includes a tree search and update method for searching and updating a tree using a node structure of a multiple search tree, and a computer recording a program for realizing the method. The purpose is to provide a readable recording medium.

That is, the present invention uses only one pointer written to a node regardless of the number of keys used by a node in the multiple search tree, so that the node pointer, the number of keys, the root port, the key, the 2-bit flag, the port, By matching the area port with the cache line size, it speeds up the multiple search tree, increasing the number of keys that can be stored in the node in network address search, and thus speeds up the search speed and enables maximum valid matching in the Btree. An object of the present invention is to provide a method of searching and updating a tree using a node structure of a multiple search tree, and a computer-readable recording medium recording a program for realizing the method.

1 is a diagram illustrating an embodiment of a general gigabit router.

2 illustrates a data structure according to a general cache line.

3 is a diagram illustrating a structure in which three IP paths are extended to an IP region.

4 is a view showing a situation when adding a new prefix to the structure extended to the IP region shown in FIG.

5 is a diagram illustrating a tree structure of a general bittree.

6 is an embodiment diagram of a data structure of a multiple search tree according to the present invention;

FIG. 7 is an embodiment diagram of a node structure of multiple search trees in the data structure shown in FIG. 6; FIG.

FIG. 8 is an embodiment diagram of the structure of a key in the node structure of the multiple search tree shown in FIG.

9 is a diagram showing a tree condition in a node structure of a multiple search tree according to the present invention.

FIG. 10A illustrates an update process by modifying a key value of a tree in a node structure of a multiple search tree according to the present invention. FIG.

FIG. 10B illustrates an update process by inserting and deleting key values of a tree in a node structure of a multiple search tree according to the present invention. FIG.

11 is a flowchart illustrating a tree search method using a node structure of a multiple search tree according to the present invention.

12 is a flowchart illustrating a tree update method using a node structure of a multiple search tree according to the present invention.

The tree search method of the present invention for achieving the above object is a tree search method applied to a tree search system using a node structure of a multiple search tree, the initial bit array using the upper bit segment of the destination IP address A first step of searching for a path and moving to a root node indicated by a node pointer to search for multiple paths with a lower bit offset of the destination IP address to determine whether Exact Matching is performed before the leaf node arrives; A second step of returning the corresponding port and ending the lookup if the match is complete before the leaf node arrives as a result of the determination of the first step; And a third step of returning or storing the area port of the key in a register according to the node that the search has reached if the result of the determination of the first step is that the match is not completed before the leaf node arrives. .

Meanwhile, the tree update method of the present invention is a tree update method applied to a tree update system using a node structure of a multiple search tree, and performs an update in an initial array according to a prefix length of a corresponding entry, or a node pointer is rooted. Determining whether the node is instructed; A second step of creating a new root node and inserting a new key value into the generated root node after confirming that the node pointer does not indicate the root node as a result of the determination of the first step; A third step of determining whether a corresponding key value exists by moving to a root node of a corresponding entry when the node pointer indicates a root node as a result of the determination of the first step; As a result of the determination of the third step, after confirming that the key value does not exist, the new key pair is inserted, and a port number of an IP prefix of a range larger than the input IP path is searched and found within the corresponding IP path range. Deleting a key pair from the tree after performing the update of the tree at; And a fifth step of deleting a key pair from the tree after performing a tree update from a searched key value to a found key pair according to the modification and deletion of the update as a result of the determination of the third step. It is characterized by including.

Meanwhile, in order to search for multiple search trees, the present invention provides a multiple search tree search system having a processor to search for an initial bit array using an upper bit segment of a destination IP address to a root node indicated by a node pointer. A first function of moving to search for multiple paths with a lower bit offset of the destination IP address to determine whether Exact Matching is made before the leaf node arrives; A second function of returning a corresponding port and ending the lookup if a match is complete before the leaf node arrives as a result of the determination in the first function; And a program for realizing a third function of returning or storing the area port of the key in a register according to the node at which the search has arrived if the result of the determination in the first function is that a full match is not achieved before the leaf node arrives. Provide a computer-readable recording medium for recording.

In addition, the present invention, in order to update the multiple search tree, in the multiple search tree update system with a processor to perform an update in the initial array according to the prefix length of the entry, or determine whether the node pointer indicates the root node A first function of doing; A second function of creating a new root node and inserting a new key value into the generated root node after confirming that the node pointer does not indicate the root node as a result of the determination in the first function; A third function of determining whether a corresponding key value exists by moving to a root node of a corresponding entry when a node pointer indicates a root node as a result of the determination in the first function; As a result of the determination in the third function, after confirming that the key value does not exist, the new key pair is inserted and the port number of the IP prefix of a range larger than the input IP path is searched to confirm that the key value is not deleted. A fourth function of deleting a key pair from the tree after performing an update of the tree within the tree; And a fifth function of deleting a key pair from the tree after performing a tree update from a searched key value until a corresponding key pair is found according to the modification and deletion of the update as a result of the determination in the third function. A computer readable recording medium having recorded thereon a program for realizing the present invention is provided.

The above objects, features and advantages will be apparent from the following detailed description taken in conjunction with the accompanying drawings. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

6 is an embodiment diagram of a data structure of a multiple search tree according to the present invention.

That is, FIG. 6 is an example of configuring a multipath tree composed of an initial 16-bit array and an 8-way search tree. "Flag indicates whether 16 bits is a node pointer or port number," P "flag indicates whether a key is entered or not," U "flag indicates whether the key is 16 bits long or not, and" S "flag indicates Indicates whether the key is a start pointer or end point.

The destination IP address entered into the router retrieves the initial 16-bit array using the upper 16-bit segment. If the first "I" flag is set to 'FALSE', since there is no information on the node pointer, the corresponding port number is returned and the lookup ends. If the "I" flag is 'TRUE', it moves to the root node address of the multipath tree indicated by the node pointer, starts searching the multipath tree using the low offset of the IP address, and returns the port number from the tree. The lookup ends. This initial 16-bit array IP expansion method is the same as that used by Lampson, but differs greatly in terms of node configuration and updates in a multipath tree.

The "P" flag, "U" flag, and "S" flag of the initial 16-bit array are used when updating the 16-bit array table. The "P" flag is set to 'TRUE' when a corresponding entry is entered into the routing table, and the "U" flag is a single prefix when the entry is entered into the routing table, i.e. a single prefix. It is set to 'TRUE' only for (Unique Prefix).

FIG. 7 is a diagram illustrating an embodiment of a node structure of a multiple search tree in the data structure shown in FIG. 6.

That is, FIG. 7 illustrates an embodiment of configuring an 8-way search tree by applying a node structure to a bit tree, that is, one node is 32 bytes in total, and seven keys. (K1 to K7), one node pointer Po, the number of keys, Num, and the Root Port, which are matched with 32 byte cache lines. Here, all the keys K1 to K7 are composed of 32 bits, and the pointer Po is composed of 16 bits, respectively.

Herein, the node pointer Po indicates an address of a lower node and, in the case of a leaf node, indicates 'null'. The number of keys, Num, displays information about the number of keys in the node. The root port shows the port number that corresponds to the entry in this 16-bit array and is used only at the root node. Set to 'null' on the node. In addition, the root port refers to the root port in the root node when the key in the lower node belongs to the same range as the entry of the 16-bit array. Therefore, it is used to prevent the necessity of updating keys referring to the root port in the lower node when the root port is changed due to the update of the entry of the 16-bit array. The remaining bits are used to store the key value and the port number corresponding to the key value.

FIG. 8 is a diagram illustrating the structure of a key in the node structure of the multiple search tree shown in FIG.

That is, FIG. 8 is a detailed structure of one key in seven keys K1 to K7 composed of 32 bits each, and includes a 16-bit key number, a 2-bit flag, a 7-bit port, and a 7-bit area port. do.

Here, the port is compared with the lower offset of the IP requesting the search and the 16-bit key value in the node. The port is used when the exact matching, which is the same value as the key, is used. Used if greater than the key value.

The two-bit flag is used to update the tree. The "U" flag is set when the prefix length of the key is 32 bits, and the "S" flag is the entry or end point in the IP address range. Indicates cognition.

9 is a diagram illustrating a tree condition in a node structure of a multiple search tree according to the present invention.

As shown in FIG. 9, the pair of K1 in the key is K7, the pair of K2 is K6, and the pair of K3 is K5. K4 is a unique prefix that is an independent key and does not affect the stack execution process.

10A is a diagram illustrating an update process by modifying a key value of a tree in a node structure of a multiple search tree according to the present invention.

That is, FIG. 10A illustrates a process of updating a key according to modification by the condition of the tree illustrated in FIG. 9. Assuming that the port number of the key K2 should be modified, first, K2 is searched to find a path passed in the search process. The node address and the passed position information are stored in the register existing for each level (N = 0). At this time, since K2 was found at the second position of the lower first node, the port and range port of K2 are modified, stored in the stack, and then moved to the right key (K3). Each time a move is made, the "U" flag and the "S" flag stored in each key are examined to determine whether the stack is stored.

Here, if the "S" flag is set, it is stored in the start point stack and moved to the next key. If the "S" flag is not set, the corresponding key can be removed from the end point stack to maintain the range of IP to which each key belongs.

That is, since K3 has the "S" flag set, the flag of K4 is examined using the information of the registers present at each level in order to be stored on the stack and move to the upper node. Since K4 is an independent IP with the "U" flag set, increase the register N by "1" without changing the stack, and move to the second lower node to check the flag of K5.

Since K5 does not have the "S" flag set, it removes K3 from the stack, modifies the area port of K5 to the updated port number of K2, and moves to its key pair K6. Since K6 is the end point of the update, only the port of K6 is modified and the update is finished. If the found key exists in the branch node, the above update process is performed by moving to the first key of the nearest lower right node. In this example, the ports of K2 and zones, the ports of K5, and ports of K6 are modified.

On the other hand, the number of registers present in each higher level except the last level is affected by the number of keys present in the corresponding entry of the 16-bit array. According to Lampson, the maximum number of keys in an 16-bit entry is 336 if the maximum number of keys in a 16-bit entry is 336 as a result of examining the mae-east routing table with the largest number of entries. 3 is sufficient, but a recent mae-east routing table has shown that up to 5 levels can sometimes be reached beyond the number of 336 keys, so it is advisable to have up to 4 registers. .

10B is a diagram illustrating an update process by inserting and deleting key values of a tree in a node structure of a multiple search tree according to the present invention.

That is, FIG. 10B is a process of updating a key according to insertion and deletion based on the condition of the tree shown in FIG. 9. First, since a range of IP paths larger than a corresponding key pair needs to be found, the key moves to the left side of the inserted key. Examine the "S" flag and zone port value. The stack is adjusted according to the state of the "S" flag of each key, and the structure of the stack operates in the reverse direction of FIG. 10A to find the key with the first "S" flag set without an IP path pairing. At this time, the key corresponds to a larger range of keys including a key to be deleted or newly inserted. If the "S" flag is not set during the movement, and the Range Port value encounters a key referring to null or the root port, the movement is terminated.

On the other hand, if the range port value is null, it means that there is no key in the range larger than the key for updating, and the range port value indicates the root port. If so, the entry that contains the key to update is the entry (Root Port).

That is, since K1 on the left side of K2 is the first key (key including K2) that does not form an IP pair, update is performed from K1 to K6. In the case of new insertion, the update is completed when K6 is reached, but in the case of deletion, the update is terminated by deleting the corresponding K2 and K6 after the above process is completed.

When deleting or inserting a key as above, move to the left of the key to find a key in a range larger than the current key, and move to the right around the end key of the key pair to form a key pair. It is also possible to find a port on a key whose first S flag is not set and perform an update.

11 is a flowchart illustrating a tree search method using a node structure of a multiple search tree according to the present invention.

First, an initial 16-bit array is searched using the upper 16-bit segment of the input destination IP address (101). In operation 102, it is determined whether the "I" flag of the corresponding entry is "TRUE" in the searched initial 16-bit array.

As a result of the determination in step 102, if the "I" flag of the entry is not 'TRUE', that is, 'FALSE', the corresponding port number is returned and the loop ends (108).

As a result of the determination in the step 102, if the "I" flag of the entry is 'TRUE', the multi-path search is performed with the lower 16-bit offset of the input destination IP address when the node pointer indicates the root node. In step 103, it is determined whether the leaf node is fully matched before reaching 104.

As a result of the determination in step 104, if a perfect match is reached before the leaf node arrives (108), the corresponding port number is returned (108), and the loop ends. On the other hand, as a result of the determination in step 104, if a complete match is not reached before the leaf arrives, it is determined whether the arrival of the search is a branch node or a leaf node (105).

As a result of the determination in step 105, if the arrival of the search is a branch node, each time the branch node passes through the branch node, the region port value of the key passing through is stored in a register and moved to the next load (107). The loop returns to step 104 to determine if a perfect match is reached before reaching the loop.

On the other hand, as a result of the determination in step 105, if the arrival of the search is a leaf node, that is, if there is no smallest key value closest to the leaf node, the corresponding key value exists in the intermediate node. Since the corresponding area port is being stored in the register during the retrieval since the key value has already passed during execution, the area port stored in this register is returned and the present loop ends (106).

12 is a flowchart illustrating a tree update method using a node structure of a multiple search tree according to the present invention.

That is, FIG. 12 is a flowchart in which an IP prefix performs an update by a routing protocol, wherein the update can be classified into three types of modification, insertion, and deletion, and slightly different updates are performed according to each classification. .

First, in the initial state 200, it is first determined whether the first length of the prefix of the corresponding entry is smaller than 16 (201).

As a result of the determination in step 201, if the length of the prefix is smaller than 16 bits, the update is simply performed using the "P" flag, the "U" flag, and the "S" flag in the initial 16-bit array (202). . The method of updating this 16-bit array table is the same as the fourth method proposed by "Gupta".

On the other hand, as a result of the determination in step 201, if the length of the prefix of the entry is longer than 16 bits, it is necessary to update in the multi-path tree of the entry, so is the "I" flag of the entry set? Determine (203).

As a result of the determination in step 203, if the "I" flag of the entry is not set, that is, if the node pointer is null, it is determined whether to delete the key (204). The node is created, a new key value is inserted and the update is terminated (205). If the key is deleted, an error is returned (206) because the corresponding key does not exist.

On the other hand, if it is determined in step 203 that the "I" flag of the corresponding entry is set (when the node pointer indicates the root node), it moves to the root node of the entry and checks whether the corresponding key value exists through key search. It is determined whether or not (207, 208).

As a result of the determination in step 208, if the key value does not exist, it is determined whether the key is deleted (209). If the key is deleted, the key is not present, and an error is returned (210). On the other hand, if the key is not deleted, a new pair of keys is inserted (211), a port number of an IP prefix of a range larger than the input IP route is found (212), and a tree within the corresponding IP route range is found. Proceeds to the process of performing an update of (214).

As a result of the determination in step 208, if there is a key value, it is determined whether the update is Modify (213). That is, the process 213 determines if the update method is a modification or deletion if a key value exists, and determines the order sequentially. In the case of modification, the tree is updated from the searched key value until the corresponding key pair is found. In step 214, it is determined whether the key is deleted (215).

In this case, as a result of the determination in step 215, if the key is deleted, the key pair is deleted from the tree, and then terminated (216). If the key is not deleted, the loop ends immediately.

On the other hand, if it is determined in step 213 that the update is not a fix, that is, if it is a delete, the port of a larger range of IP prefixes is found (212), and the tree is updated to the corresponding key pair. The process proceeds to step 214 and loops.

The method of the present invention as described above may be implemented as a program and stored in a computer-readable recording medium (CD-ROM, RAM, ROM, floppy disk, hard disk, magneto-optical disk, etc.).

The present invention described above is not limited to the stated embodiments and the accompanying drawings, and various substitutions, modifications, and changes are possible in the art without departing from the technical spirit of the present invention. It will be clear to those of ordinary knowledge.

In the present invention as described above, since the multi-path tree is generated for each 16-bit segment of the IP by combining the initial 16-bit array and the multi-path tree, the level of the tree can be reduced and used in one node of the multi-path tree. By using only one number of pointers, the number of keys can be increased to further reduce the level of the tree, thereby speeding up the search and speeding up the forwarding engine.

In addition, the present invention is an update method utilizing the advantage that it is easy to move between nodes of the same level when updating a multi-path tree because the lower node is stored in a contiguous memory location due to the use of one pointer in the node. Updates can be made at, which will speed up the update.

Claims (13)

In a tree search method applied to a tree search system using a node structure of a multiple search tree, Search the initial bit array using the upper bit segment of the destination IP address, go to the root node indicated by the node pointer, and search for multiple paths with the lower bit offset of the destination IP address, so that a full match (before the leaf node arrives) A first step of determining whether or not Exact Matching is performed; A second step of returning the corresponding port and ending the lookup if the match is complete before the leaf node arrives as a result of the determination of the first step; And A third step of returning or storing the area port of the key in a register according to the node that the search has reached if the result of the determination of the first step is not complete match before the leaf node arrives; Tree search method using the node structure of the multiple search tree comprising a. The method of claim 1, A fourth step of retrieving the initial bit array using the upper bit segment of the destination IP address and returning the corresponding port to a register when the node pointer does not exist and ending the lookup Tree search method using the node structure of the multiple search tree further comprising. The method of claim 2, The third step, A fifth step of determining whether the arrival of the search is a branch node or a leaf node; A sixth step of returning an area port of a corresponding key if the search arrives as a leaf node, and a port area stored in a register if the key does not exist; And As a result of the determination of the fifth step, when the search reaches the branch node, the seventh step of storing the region port value of the key passing through the branch node in a register and moving to the next node to search for the multi-path Tree search method using the node structure of the multiple search tree comprising a. The method according to any one of claims 1 to 3, The node structure of the multiple search tree is One node pointer (Po) to tell the address of the next node, a num to display information about the number of keys, a root port used at the root node, a number of keys and a number of Tree search method using node structure of multiple search tree including area ports. The method of claim 4, wherein The node pointer indicates the address of the lower node and, in the case of a leaf node, null. And the loop port indicates a port number corresponding to an entry of a bit array. The method of claim 4, wherein The key is The maximum effective match of the IP sub-offset that requested the search, the key value in the node, the port used in case of exact matching of the same value as the key value, and the key value closest to the key value. Prefix Marching), a node of a multiple search tree comprising a region port used in the case of a prefix, a flag set according to a prefix length of a key, and a flag indicating whether the entry point or the end point is included in the IP address range of the corresponding entry. Tree search method using structure. In a tree update method applied to a tree update system using a node structure of a multiple search tree, A first step of performing an update on an initial array or determining whether a node pointer indicates a root node according to the prefix length of the entry; A second step of creating a new root node and inserting a new key value into the generated root node after confirming that the node pointer does not indicate the root node as a result of the determination of the first step; A third step of determining whether a corresponding key value exists by moving to a root node of a corresponding entry when the node pointer indicates a root node as a result of the determination of the first step; As a result of the determination of the third step, after confirming that the key value does not exist, the new key pair is inserted, and a port number of an IP prefix of a range larger than the input IP path is searched and found within the corresponding IP path range. Deleting a key pair from the tree after performing the update of the tree at; And A fifth step of deleting the key pair from the tree after performing a tree update from a searched key value to a found key pair according to the modification and deletion of the update if the key value exists as a result of the determination in the third step Tree update method using the node structure of the multiple navigation tree comprising a. The method of claim 7, wherein The five steps, A sixth step of deleting a key pair from the tree after performing a tree update from the searched key value until a corresponding key pair is found if the update is modified because the corresponding key value exists; And If the up data is deleted because the corresponding key value exists, search for the port number of the IP prefix with a range larger than the entered IP path, perform a tree update to the corresponding key pair, and then delete the key pair from the tree. step Tree update method using the node structure of the multiple navigation tree comprising a. The method according to claim 7 or 8, The node structure of the multiple search tree is One node pointer (Po) to tell the address of the next node, a num to display information about the number of keys, a root port used at the root node, a number of keys and a number of Tree update method using node structure of multiple search trees including area ports. The method of claim 9, The node pointer indicates the address of the lower node and, in the case of a leaf node, null. And the loop port indicates a port number corresponding to an entry of a bit array. The method of claim 9, The key is The maximum effective match of the IP sub-offset that requested the search, the key value in the node, the port used in case of exact matching of the same value as the key value, and the key value closest to the key value. Prefix Marching), a node of a multiple search tree comprising a region port used in the case of a prefix, a flag set according to a prefix length of a key, and a flag indicating whether the entry point or the end point is included in the IP address range of the corresponding entry. Tree update method using structure. In order to search multiple search trees, in a multiple search tree search system with a processor, Search the initial bit array using the upper bit segment of the destination IP address, go to the root node indicated by the node pointer, and search for multiple paths with the lower bit offset of the destination IP address, so that a full match is made before the leaf node arrives. A first function of determining whether or not Exact Matching); A second function of returning a corresponding port and ending the lookup if a match is complete before the leaf node arrives as a result of the determination in the first function; And A third function of returning or storing the area port of the key in a register according to the node at which the search arrives, when the result of the determination in the first function is that a complete match is not achieved before the leaf node arrives. A computer-readable recording medium having recorded thereon a program for realizing this. In order to update multiple search trees, in a multiple search tree update system having a processor, A first function of performing an update on an initial array or determining whether a node pointer points to a root node according to the prefix length of the entry; A second function of creating a new root node and inserting a new key value into the generated root node after confirming that the node pointer does not indicate the root node as a result of the determination in the first function; A third function of determining whether a corresponding key value exists by moving to a root node of a corresponding entry when a node pointer indicates a root node as a result of the determination in the first function; As a result of the determination in the third function, after confirming that the key value does not exist, the new key pair is inserted and the port number of the IP prefix of a range larger than the input IP path is searched to confirm that the key value is not deleted. A fourth function of deleting a key pair from the tree after performing an update of the tree within the tree; And A fifth function of deleting a key pair from the tree after performing a tree update from a searched key value until a corresponding key pair is found according to the modification and deletion of the update as a result of the determination in the third function; A computer-readable recording medium having recorded thereon a program for realizing this.