JP2003157187A - Algorithm for extracting difference between two pieces of binary information - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the processing efficiency and to shorten the processing time for extracting the difference between pieces of binary information. SOLUTION: Binary information is not compared byte by byte but is compared in the unit of blocks to retrieve parts of coincidence between pieces of information because information like an image is not rewritten in one byte unit but is rewritten in a unit of cluster of bytes in many cases. In a comparison in the unit of blocks, it is easy to make the compared part into hardware, and the processing speed of an algorithm can be increased by making it into hardware. Comparison in the unit of blocks is terminated at a fixed number of bytes (for example, 1,024 bytes). It is preferable that this fixed number of bytes is changed in accordance with classifications of information. This change realizes reduction of difference information extracted and increase of the processing speed.

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of comparing two binary files to extract a difference, instead of comparing one byte at a time, and comparing blocks. Further, the present invention relates to a method for speeding up processing by terminating comparison in block units at a fixed value. 2. Description of the Related Art When an arbitrary file such as a changing image file or data file is transmitted on the Internet or an intranet or the like, only the difference from the previous data is transmitted, rather than retransmission of the entire data. The technique of increasing the speed has been used conventionally. However, since the difference is extracted after comparing one byte at a time, there is a limit to speeding up. [0003] An object of the present invention is to increase the efficiency of difference extraction and reduce the time. "Preparation" A counter for counting the number of matching bytes (referred to as A for description) is set to zero. The old information and the new information to be compared are prepared, and counters for counting the number of bytes from the head of the information are prepared (these are referred to as B and C, respectively, for the following description). The values of B and C are also set to 0. [Repeat processing] (1) Both counters of the number of bytes from the head (B, C)
If both of them exceed the last byte of information, the repetition processing is terminated. At this time, the matching counter (A) is set to 0
If it is larger than the value, the value is recorded in the difference information. (2) When the value obtained by adding one to the number of bytes (B, C) from the beginning exceeds the number of the last byte, the repetition processing is terminated. At this time, if the coincident counter (A) is larger than 0, the value is recorded in the difference information. In addition, if the old information has been removed from the new information, record the number of bytes removed in the difference information, and if the old information has new information, add the number of bytes added. The added information is recorded as difference information. (3) One byte of information indicated by the number of bytes (B, C) from the head is compared. (3.1) When the compared information matches, both B and C are 1
And returns to (1). (3.2) When the compared information is different, the value of the counter (A) that matches is larger than 0. This value is recorded in the difference information, and the value of A is returned to 0. Subsequently, the following processing is performed. (4) A fixed number of bytes (for example, 16 bytes) of new information is set as a block, and a byte string having the same size as the block is compared with old binary information. (4.1) Old binary information and new binary information
If 6 bytes match, record the number of matching bytes (16 bytes) and return to (1) (4.2) If some or all of the block contents do not match, start the old binary information block The position is moved by one byte, and the comparison of (4) is performed. If the movement of the start position exceeds the total value (for example, 1024 bytes), the block comparison is stopped and the process proceeds to (5). (5) Until the old binary information matches the new binary information, or The comparison is performed until one reaches the end of the information. (6) The number of bytes compared in the difference information and the value of the compared new binary information are recorded by the amount compared in (5), and the process returns to (1). DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A program for realizing the present invention is installed in a computer. With this program, it becomes possible to extract differences and restore files of each version based on the extracted differences. [0012] Hardware for implementing the present invention is added to a computer. With this hardware, it is possible to extract differences and restore files of each version based on the extracted differences. DESCRIPTION OF THE PREFERRED EMBODIMENTS When a part of image information is changed, old image information is held at two sites, and a difference between new image information and old image information is extracted using the present algorithm. By transferring only the difference information to the other, the image information can be transferred without sending the whole new image information. As a result, only the difference can be transmitted without transmitting the entire large image information, so that the communication amount can be reduced. When storing information in which part of image information has been changed, only the old image information and the difference information obtained by the present algorithm are stored, and the entire changed image information is stored. The changed image information can be saved without any change. By using this, when managing the version of the information (version management), the management can be realized without saving the entire data for each version. According to the present invention, it is possible to speed up the difference extraction by extracting the difference efficiently. As a result, when transmitting information such as moving image data that changes every moment, it becomes possible to extract and transmit the difference from the original information in a shorter time than before. Based on the extracted difference, the time required to generate an arbitrary version of the file can be greatly reduced.
Only the extracted differences are stored, and the version can be managed by restoring an arbitrary version.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is an algorithm diagram according to an embodiment of the present invention.

Claims (1)

Claims: 1. When extracting a difference between two arbitrary pieces of binary information, the binary information is compared in units of blocks (for example, 16 bytes) instead of comparing one byte at a time. Search for the matching part, and further compare in block units for a certain fixed value (change the value according to the type of data)
The present invention relates to all algorithms and implementation methods for reducing difference information to be extracted and reducing the processing speed.