KR20010019620A - Data Search Method Using Binary Linked List - Google Patents

Abstract

PURPOSE: A data search method is provided to store data with a binary linked list structure including an upper pointer and a lower pointer, and search the data by using the binary linked list so that it can reduce a search time. CONSTITUTION: The method comprises steps of reading data stored in the first register among the binary linked list of a corresponding data memory(S61, S62), comparing the recorded data with a search data(S63), reporting that a corresponding search data is found if the comparison result is identical, and stopping the search process(S64), checking if the recorded data is over the search data in the case that the comparison result is not identical(S65), reading the register directed by the lower pointer in the case that the recorded data is over the search data(S66), reading the register directed by the upper pointer in the case that the recorded data is under the search data(S67), comparing the recorded data with the search data in the memory area directed by the upper or lower pointer(S68), reporting that a corresponding search data is found if the comparison result is identical, and stopping the search process(S64), and repeating the checking and comparison processes till finding the corresponding search data or no data in the upper or the lower pointer.

Description

Data Search Method Using Binary Linked List}

The present invention relates to a data retrieval method using a binary linked list. In particular, a pointer of a linked list is divided into an upper pointer and a lower pointer, and the data is searched using the upper or lower pointer according to a result of comparing the search data with the recorded data. The present invention relates to a data retrieval method using a binary linked list.

In general, a compression function is used to retrieve the existence of desired data and the location of the data from a memory in which a plurality of data are stored. The compression function compresses a large amount of data according to address information of the compressed memory. Has the function of making form data.

That is, a structure of a memory having a compressed memory for searching for a corresponding material is a data memory area in which actual data exists and a compressed memory area in which address information of the data memory is recorded, as shown in FIG. 1. Are distinguished.

Here, the address information of the compressed memory is the remainder obtained by dividing the data to be searched by the polynomial as shown in FIG. 2, and the actual content of the compressed memory has the address information of the data memory as a result of the compression function.

For example, as shown in FIG. 1, only the lower 6-bit value is allocated to a corresponding compressed memory area to store data in a compressed form that is a result of a compression function. The data memory is connected as shown in FIG. It has a double structure that stores a pointer to a linked list structure and actual data.

Meanwhile, a memory search operation using a conventional compression function will be described in detail as follows.

First, a compressed memory address having passed through a compression function is obtained from specific input data, and then a data memory address is extracted from the compressed memory using the corresponding compressed memory address.

Then, the corresponding data memory area is read using the data memory address extracted from the compressed memory to check whether there is data to be searched (hereinafter referred to as 'search data').

That is, since there is a connection list pointer and the actual recorded data in the memory area corresponding to the corresponding data memory address, it is checked whether the corresponding record data and the search data are matched and matched. As shown in Fig. 3, the data recorded in the next address area indicated by the corresponding connection list pointer is compared with the search data to sequentially check whether the search data exists in the corresponding data memory.

In this case, if the data of the predetermined address area indicated by the linked list pointer and the search data match, that is, if the search data exists, it is reported and the search is continuously performed until the corresponding linked list pointer becomes blank. If the search data does not exist, it will report that the search data does not exist.

However, in the related art, when data search time is lengthened in proportion to the depth of a linked list, when the data exists in a short time and needs to be used as an internal control signal, normal processing may not be performed as the search time is lengthened.

For example, when performing high-speed switching in an Ethernet switch, which is a network equipment, switching information is requested using the MAC address of an incoming packet. When the searching time is long when switching information is requested using the corresponding MAC address, the switching delay is increased. As a result, there was a case in which the network equipment could not perform a normal switching process.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and its object is to construct a binary linked list divided into an upper pointer and a lower pointer so that data can be managed and then searched by comparing the corresponding search data with the recorded data. If the data is small, the data of the memory area indicated by the lower pointer is searched, and if the search data is large, the operation of searching the data of the memory area indicated by the upper pointer is repeated, thereby shortening the data search time.

1 shows a structure of a general memory having a compressed memory.

2 shows a conventional compressed memory and a data memory.

3 illustrates a conventional data memory retrieval procedure.

4 illustrates a compressed memory and a data memory according to the present invention.

5 is a diagram illustrating a data memory retrieval procedure according to the present invention.

6 is an operational flowchart for implementing a data retrieval method using a binary linked list according to the present invention.

7 illustrates a typical Ethernet switch and memory.

8A is a diagram illustrating a conventional data retrieval procedure, and (B) is a diagram illustrating a data retrieval procedure according to the present invention.

A feature of the present invention for achieving the above object includes the steps of: reading recorded data using a data memory address extracted from a compressed memory; Confirming whether the read record data and the search data match; Checking whether the recorded data is larger than the search data if they do not match; Reading recorded data indicated by the upper or lower pointer of the binary linked list according to the checked result; Confirming whether the read record data and the search data match; If there is a match, the process includes reporting that the search data exists.

The reading of the recording data indicated by the upper or lower pointer of the binary linked list may include reading the recording data indicated by the lower pointer when the recording data is larger than the search data, and reading the recording data indicated by the lower pointer. In this case, the recording data indicated by the higher pointer is read.

And in the process of checking whether the read record data and the search data match, if the record data and the search data do not match and the record data is larger than the search data, checking whether the lower pointer of the binary linked list is blank. Wow; The method may further include reporting that the search data does not exist in the case of a blank. If the record data and the search data do not match and the record data is smaller than the search data, the upper pointer of the binary linked list is blank. Confirming; If blank, the method may further include reporting that no search data exists.

In addition, the reporting of the absence of the search data may include recording the search data in a predetermined area of the data memory, and recording address information in which the search data is recorded in a blank pointer. It is done.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

In the present invention, the structure of the compressed memory for shortening the data retrieval time is the same as in the related art, and the structure of the binary linked list is changed from the structure of the linked list existing in the conventional data memory. As shown in FIG. 4, the linked list pointer of the data structure existing in the corresponding data memory is composed of a binary linked list divided into an upper pointer and a lower pointer, and data is obtained by using the binary pointer information of the upper pointer and the lower pointer. The search will be performed.

In this case, in order to perform data search using the corresponding binary pointer information, as shown in FIG. 5, the search area of the linked list is divided according to the size of the data to be searched. By comparing the search data recorded in the data memory, if the search data is larger than the recording data, the data of the area indicated by the upper pointer is searched. On the contrary, if the search data is smaller than the recording data, the data of the area indicated by the lower pointer is searched. Will be searched.

That is, the read operation of the compressed memory when searching data using the binary linked list according to the present invention is the same as the conventional method, and since performing the binary search based on the search data in the corresponding data memory is distinguished from the conventional method, the corresponding compressed memory A search operation for a corresponding data memory region by extracting a data memory address from the compressed memory using the address will be described in detail with reference to FIG. 6.

First, when performing a search for a corresponding data memory area using the data memory address extracted from the compressed memory, after reading the data recorded in the first stage from the binary linked list of the corresponding data memory (steps S61 and S62), The read record data and the search data are compared to confirm whether they match (step S63).

At this time, if the record data and the search data match, it is reported that the search data exists (step S64), and then the data search operation is terminated. If the record data and the search data do not match, the record data and the search data are not matched. The data is compared again to confirm whether the record data is larger than the search data (step S65).

If the record data is larger than the search data, the recorded data is read in the data memory area indicated by the lower pointer of the binary linked list from which the record data is read (step S66), and the record data is smaller than the search data. The data written in the data memory area indicated by the upper pointer of the binary linked list which has read the write data is read (step S67).

Thereafter, the write data read out from the data memory area indicated by the corresponding upper or lower pointer is compared with the search data to check whether the matching data is matched (step S68). (Step S64), the data search operation is ended, and if the data does not match, the record data and the search data are compared again to confirm whether the record data is larger than the search data (step S69).

If the record data is larger than the search data, check whether the lower pointer of the binary linked list from which the record data has been read is blank (step S70). If the record data is not blank, the data recorded in the data memory area indicated by the corresponding lower pointer is checked. The flow returns to step S66 of reading out the sequence to perform the repetitive operation.

By the way, if the record data is smaller than the search data, it is checked whether the upper pointer of the binary linked list from which the record data has been read is blank (step S71), and if it is not blank, the data recorded in the data memory area indicated by the higher pointer is blank. The process returns to step S67 of reading and the repetitive operation is performed.

On the other hand, when the lower pointer is blank in step S70 or the upper pointer is blank in step S71, the search operation is terminated after reporting that there is no corresponding search data (step S72). At this time, the search data is recorded in a predetermined area of the data memory, but the address information in which the search data is recorded is recorded in a blank upper or lower pointer.

For example, in the case of an Ethernet switch supporting a plurality of network ports, as shown in FIG. 7, a 6-byte MAC address inside the packet for high-speed switching of an Ethernet packet, which is network information transmitted from each port, by configuring a memory. Requesting switching information using Ethernet MAC address, which is input data to request switching information on data memory using MAC address, is '123456ABCDEF' (hexadecimal), and data recorded in memory is '23456789ABCD' Assuming (hexadecimal), since the corresponding MAC address is smaller than the write data, the write data of the data memory area indicated by the lower pointer is read to check whether the matched data is the same.

If it does not match, check whether the corresponding write data is larger than the MAC address which is the input data, and according to the result, read the write data in the data memory area indicated by the lower or upper pointer to check whether the write data matches the corresponding search data. By repeating this operation, it is possible to search the data faster than in the conventional data search operation.

In other words, in the conventional data search operation, when the search data is 'bcde' as shown in FIG. 8 (a) of the accompanying drawings, the data search operation should be performed by searching along the pointer of the linked list until the corresponding pointer becomes 'blank'. Although it was found that the search data does not exist, when the data search is performed using the binary linked list according to the present invention, since the search data 'bcde' is larger than the recorded data '1234' in the first stage as shown in FIG. Therefore, the data is compared with 'a34c' in the area indicated by the upper pointer. At this time, the upper pointer is 'blank' in a state where the search data is large and do not match with each other. .

Then, the search data 'bcde' is recorded in the data memory, and the address information in which the search data is recorded is recorded in the 'blank' of the upper pointer in which 'a34c' is recorded.

In addition, the embodiments according to the present invention are not limited to the above-described embodiments, and various alternatives, modifications, and changes can be made within the scope apparent to those skilled in the art.

As described above, the present invention configures a binary linked list divided into an upper pointer and a lower pointer to manage data, and then compares the corresponding search data with the recorded data, and the memory indicated by the lower pointer when the search data is small. By retrieving the data of the area and retrieving the data of the memory area indicated by the upper pointer when the search data is large, the corresponding data retrieval time can be shortened.

Claims (5)

Reading recorded data using the data memory address extracted from the compressed memory; Confirming whether the read record data and the search data match; Checking whether the recorded data is larger than the search data if they do not match; Reading recorded data indicated by the upper or lower pointer of the binary linked list according to the checked result; Confirming whether the read record data and the search data match; And if it matches, reporting the presence of search data. The method of claim 1, The process of reading the write data indicated by the upper or lower pointer of the binary linked list may include reading the write data indicated by the lower pointer when the record data is larger than the search data, and reading the record data indicated by the lower pointer than the search data. A data retrieval method using a binary linked list, characterized by reading the record data indicated by a pointer. The method of claim 1, Determining whether the lower pointer of the binary linked list is blank when the recorded data and the search data do not match and the record data is larger than the search data in the process of checking whether the read record data and the search data match each other; And if it is blank, reporting the search data does not exist. The method of claim 1, Determining whether the upper pointer of the binary linked list is blank when the record data and the search data do not match and the record data is smaller than the search data in the process of checking whether the read record data and the search data match each other; And if it is blank, reporting the search data does not exist. The method according to claim 3 or 4, The reporting of the absence of the search data may include recording the search data in a predetermined area of the data memory and recording address information on which the search data is recorded in a blank pointer. Data retrieval method using binary linked list.