JP6257221B2 - Communication apparatus and data transfer control method - Google Patents

Description

  The present invention relates to a computer system for compressing and communicating information.

  The following methods are known as methods for compressing or decompressing data.

[Dictionary compression]
The lexicographic compression is a method of compressing a data string by registering an alphabetical sequence appearing in a data string as a word in the dictionary and replacing it with a dictionary registration number when the same word appears next time. For example, in the technique derived from LZ78 and LZ78 described in Non-Patent Document 1, the same dictionary is dynamically generated when compressing data in a compression device or the like and when decompressing data in a decompression device or the like. There is a feature that. Therefore, only the post-compression data string needs to be passed from the compression side to the decompression side.

  FIG. 14 is an explanatory diagram showing a procedure for compressing and expanding data using conventional lexicographic compression.

  FIG. 14 shows an example in which the data string “ababa” is compressed and decompressed by LZ78. In LZ78, a dictionary is generated by compressing data based on a simple rule “add known word + 1 alphabet as a new word to the dictionary”.

  As shown in FIG. 14, dictionaries are generated according to the dictionary creation procedure 1401 to procedure 1404 during compression, and dictionaries are generated according to the dictionary creation procedure 1411 to procedure 1414 during decompression, and the compression side and expansion side dictionaries are the same. It becomes a dictionary.

  Here, the dictionary refers to a collection of pairs of alphabetical sequences appearing in the data string and headings corresponding thereto. Furthermore, the alphabetical sequence that appears in the data string is called a word.

[Entropy code]
The entropy code is a method of compressing data by assigning code words having different lengths based on the appearance frequency of alphabets appearing in a data string. For example, in the dynamic encoding method represented by the dynamic Huffman code (Adaptive Huffman-coding) described in Non-Patent Document 2, the same frequency table is dynamically generated on the compression side and the expansion side as with the LZ78. There is a feature.

  FIG. 15A, FIG. 15B, and FIG. 15C are explanatory diagrams showing a procedure of encoding using a conventional Huffman code. Here, encoding using the Huffman code described in Non-Patent Document 3 will be described using the alphabet “a, b, c, d” having a bias in appearance frequency as an example.

The code length when simply encoding four characters is 2 bits, but as shown in FIG. 15B, the average code length is lowered to 1.8 bits by assigning a short code to “a” having a large number of appearances.

  Here, the frequency table (FIG. 15A) refers to a collection of pairs of alphabets and alphabet strings appearing in the data and information for calculating the frequency such as the appearance frequency or the number of appearances. In addition, there is an encoding method and a data structure characterized by handling a frequency table in a form that facilitates entropy encoding, such as a Huffman tree in a Huffman code (FIG. 15B), and here, these are also referred to as a frequency table.

[Communication using compression]
As background art in the technical field of the present invention, there is Non-Patent Document 4 (for example, refer to “2. Compression Process”) related to compression of IP (Internet Protocol) packets. Further, Patent Document 1 relating to data stream compression exists as another background art of this technical field.

  In US Pat. No. 6,057,089, "Encoding is a step of determining a target segment size, a step of determining a window size, a step of identifying a fingerprint in a symbol window at an offset in input data, and an offset. Determining whether to designate as a cut point, and segmenting the input data as indicated by the set of cut points. Is a referenced segment or a non-referenced segment, replaces the segment data of each referenced segment with a reference label, and, if necessary, a persistent segment of each referenced segment Stores the reference binding in the store " To have.

US Patent Application Publication No. 2004/0174276

Ziv, Jacob; Lempel, Abraham (1978), "Compression of individual sequences via variable-rate coding". IEEE Transactions on Information Theory 24 (5): pp 530-536. Vitter, Jeffrey (1987), "Design and analysis of dynamic Huffman codes". Journal of the ACM, 34 (4): pp 825-845 Huffman, David (1952), "A Method for the construction of minimum-redundancy codes". Proceedings of the I.R.E .: pp 1098-1102 RFC2393 (1998), "IP payload compression protocol (IPComp)" O'neil, Elizabeth (1993), "The LRU-K page replacement algorithm for database disk buffering". ACM SIGMOD Conf .: pp. 297-306

  In the technique described in Non-Patent Document 4, compression can be performed only for each IP packet. The size of the IP packet is equal to or smaller than the MTU (Maximum Transmission Unit) of the communication path, and is generally 1500 bytes in the Internet environment. Therefore, since the payload has a size smaller than that, the technique described in Non-Patent Document 1 does not increase the compression rate, and the effect of reducing the communication amount is low.

  The technique described in Patent Document 1 cannot determine whether or not a segment is known by a simple calculation. In addition, it is possible to increase the effect of reducing the amount of communication by increasing the capacity of the persistent segment store. However, since the HDD (Hard Disk Drive) having a low capacity unit price is generally used for the persistent segment store, when searching for data The I / O speed (the random access speed of the HDD is about 100 to 200 Mbps) is lower than the I / O speed (for example, 1 Gbps) of the communication path. Therefore, with the technique described in Patent Document 1, it is possible to increase the effect of reducing the amount of communication, but there is a problem that the communication speed decreases in a network that handles a sufficiently wide communication band.

  That is, a problem of a computer system that communicates by compressing information is to achieve both a high compression rate of information and a high communication speed.

In order to solve the above problems, for example, the configuration described in the claims is adopted. A communication device comprising a processor, a main storage device connected to the processor, and a network interface connected to the processor, and transferring data transmitted and received between a plurality of computers, wherein the main storage device is a communication destination Each including a temporary area for storing a plurality of compression / decompression information for compressing or decompressing data, and storing a copy compression / decompression information that is a copy of the compression / decompression information used for communication with an arbitrary communication destination , The communication apparatus includes a compression / decompression unit that compresses or decompresses received data using the compression / decompression information, and the compression / decompression unit associates information for specifying a communication destination with identification information for compression / decompression information to be used. and supervises as, when receiving the data from the first communication destination, from among a plurality of compression and decompression information used for communication between the first communication destination, especially the first communication destination A first decompression information associated with the information to identify the first replication decompression information is a duplicate of the first compression and expansion information stored in the temporary area, the first replication compression Using decompression information , the data received from the first communication destination is compressed or decompressed, the compressed data or the decompressed data is transmitted, and the compression result of the data received from the first communication destination or based on expansion results, the first update the replication compression and decompression information, determines whether it meets the switching condition, when it is determined not to satisfy the switching condition, the first replication decompression information When the first compression / decompression information is updated by overwriting the first compression / decompression information and it is determined that the switching condition is satisfied, a plurality of items are used for communication with the first communication destination. Of the compression / decompression information Select compression and decompression information, by associating the first information and the second decompression information specifying the communication destination, switches the compression and expansion information used for communication between the first communication destination It is characterized by that.

  According to the present invention, the compression / decompression unit can switch a plurality of pieces of compression / decompression information for each communication destination based on the data compression rate. Thereby, a high compression rate of data can be realized. In addition, since the compression / decompression information is stored in the main storage device, it is possible to avoid a decrease in the I / O speed and increase the communication speed.

  Problems, configurations, and effects other than those described above will become apparent from the description of the following examples.

It is explanatory drawing which shows an example of a structure of the computer system in Example 1 of this invention. It is a block diagram which shows the structural example of the compression / decompression apparatus in Example 1 of this invention. It is explanatory drawing which shows an example of the communication destination-dictionary correspondence information in Example 1 of this invention. It is explanatory drawing which shows an example of the communication destination-dictionary correspondence information in Example 1 of this invention. It is explanatory drawing which shows an example of the dictionary management information in Example 1 of this invention. It is a flowchart explaining the transfer process of the IP packet performed when the compression / decompression module in Example 1 of this invention receives an IP packet. It is a flowchart explaining the transfer process of the IP packet performed when the compression / decompression module in Example 1 of this invention receives an IP packet. It is a flowchart explaining the update process of the dictionary management information which the compression / decompression module in Example 1 of this invention performs. It is a block diagram which shows the structural example of the compression / decompression apparatus in Example 2 of this invention. It is explanatory drawing which shows an example of the communication destination-dictionary frequency table corresponding | compatible information in Example 2 of this invention. It is explanatory drawing which shows an example of the dictionary frequency table management information in Example 2 of this invention. It is a flowchart explaining the transfer process of the IP packet performed when the dictionary entropy compression expansion module in Example 2 of this invention receives an IP packet. It is a flowchart explaining the transfer process of the IP packet performed when the dictionary entropy compression expansion module in Example 2 of this invention receives an IP packet. It is a flowchart explaining the update process of the dictionary frequency table management information which the dictionary entropy compression / decompression module in Example 2 of this invention performs. It is explanatory drawing which shows an example of the dictionary management information in Example 3 of this invention. It is a flowchart explaining the deletion process of the dictionary in Example 3 of this invention. It is a flowchart explaining the deletion process of the dictionary in Example 3 of this invention. It is explanatory drawing which shows the procedure of the compression and expansion | extension of data using the conventional lexicographic compression. It is explanatory drawing which shows the procedure of the encoding using the conventional Huffman code. It is explanatory drawing which shows the procedure of the encoding using the conventional Huffman code. It is explanatory drawing which shows the procedure of the encoding using the conventional Huffman code.

  Hereinafter, examples will be described with reference to the drawings.

  In the present invention, description will be made assuming that at least one of a dictionary and a frequency table is used as information used when data (IP packet payload) is compressed or expanded. However, the present invention is not limited to this, and any information may be used as long as the information is used for compressing or expanding data.

  In the first embodiment, a method, apparatus, and system for compressing or expanding communication contents using a dictionary having the highest compression rate from among a plurality of dictionaries according to the communication contents will be described.

  FIG. 1 is an explanatory diagram illustrating an example of a configuration of a computer system according to the first embodiment of this invention.

  The computer system shown in FIG. 1 includes client apparatuses 101 and 102, a server 103, and compression / decompression apparatuses 111, 121, and 131. The client apparatuses A 101 and B 102 and the server 103 are connected via the wide area network 104. Note that the client device A 101, the client device B 102, and the server 103 may be connected via a network other than the wide area network 104 such as a LAN or a SAN.

  In this embodiment, the client apparatus A 101 belongs to the subnet 112 of “192.168.1.0/24”, and the IP address is “192.168.1.101”. Further, the client apparatus B102 belongs to the subnet 122 of “192.168.2.0/24”, and the IP address is “192.168.2.102”. Further, the server 103 belongs to the subnet 132 of “192.168.10.0/24”, and the IP address is “192.168.10.103”.

  The client apparatus A 101 and the client apparatus B 102 transmit and receive data to and from the server 103 via the wide area network 104. The present invention is not limited to the type and content of data to be transmitted / received. The client device A 101, the client device B 102, and the server 103 are general computers, and include hardware such as a CPU (not shown), a memory (not shown), and a network interface (not shown).

  The compression / decompression device A111, the compression / decompression device B121, and the compression / decompression device C131 compress data to be transmitted / received, and decompress the compressed data. Here, in this invention, it uses so that a compression / expansion apparatus may become a pair. For example, the compression / expansion device A111 and the compression / expansion device C131 are used as a pair, and the compression / expansion device B121 and the compression / expansion device C131 are used as a pair.

  It should be noted that the compression / decompression apparatus on the client device side and the compression / decompression apparatus on the server side may have an n-to-one configuration as shown in FIG. 1 (2 to 1 in FIG. 1), or an n-to-n configuration. Also good.

  FIG. 2 is a block diagram illustrating a configuration example of the compression / decompression apparatus according to the first embodiment of the present invention. FIG. 2 shows a configuration example of the compression / decompression apparatus A111. Note that the compression / expansion device B121 and the compression / expansion device C131 have the same configuration.

  The compression / decompression device A111 includes a CPU 201, a main storage device 202, a secondary storage device 203, and two network interfaces (NICs) 204 and 205.

  The CPU 201 executes a program stored in the main storage device 202. When the CPU 201 executes the program, the function of the compression / decompression apparatus A111 can be realized.

  The main storage device 202 stores a program executed by the CPU 201 and information necessary for executing the program. The main storage device 202 provides a work area where the program temporarily stores data. The main storage device 202 may be a memory such as a DIMM, for example.

  In this embodiment, the main storage device 202 stores a compression / decompression module 221, communication destination-dictionary correspondence information 222, and dictionary management information 223. The compression / decompression module 221 is a program module realized by the compression / decompression module program 231. Details of the communication destination-dictionary correspondence information 222 will be described later with reference to FIGS. 3A and 3B, and details of the dictionary management information 223 will be described later with reference to FIG.

  The CPU 201 reads the compression / decompression module program 231 from the secondary storage device 203 when the compression / decompression device A 111 is activated, and expands the compression / decompression module program 231 on the main storage device 202. Further, the CPU 201 implements the compression / decompression module 221 by executing the expanded compression / decompression module program 231. The compression / decompression module 221 expands communication destination-dictionary correspondence information 222 and dictionary management information 223 described later on the main storage device 202.

  The secondary storage device 203 stores programs and information. The secondary storage device 203 may be, for example, an HDD or an SDD. In this embodiment, the secondary storage device 203 stores a compression / decompression module program 231.

  NIC1 (204) and NIC2 (205) are interfaces for connecting to other devices. In this embodiment, NIC1 (204) is connected to a base-side device, that is, client device A101, and NIC2 (205) is connected to wide area network 104.

  Here, the compression / decompression module 221 will be described.

  The compression / decompression module 221 compresses data based on a predetermined compression method, and decompresses the compressed data. In the present embodiment, it is assumed that compression processing and decompression processing are executed based on a lexicographic compression method such as LZ78 and LZW derived from LZ78.

  The compression / decompression module 221 of this embodiment compresses the payload of the IP packet received from the NIC1 (204) based on the communication destination-dictionary correspondence information 222 and the dictionary management information 223, and the compressed payload from the NIC2 (205). The IP packet including is transferred to the communication destination device. The compression / decompression module 221 expands the payload of the IP packet received from the NIC 2 (205) based on the communication destination-dictionary correspondence information 222 and the dictionary management information 223, and includes the payload expanded from the NIC 1 (204). The IP packet is transferred to the communication destination device.

  The present invention is not limited to lexicographic compression methods such as LZ78 and LZW, and the compression / decompression on the receiving side can be performed without separately distributing the dictionary from the compression / decompression module 221 on the transmission side to the compression / decompression module 221 on the receiving side. Any compression method may be used as long as the module 221 can execute decompression processing.

  3A and 3B are explanatory diagrams illustrating an example of the communication destination-dictionary correspondence information 222 according to the first embodiment of this invention. 3A shows the communication destination-dictionary correspondence information 222 held by the compression / decompression device C131, and FIG. 3B shows the communication destination-dictionary correspondence information 222 held by the compression / expansion device A111.

  The communication destination-dictionary correspondence information 222 holds information indicating a correspondence relationship between the IP address range of the communication destination device and a dictionary used for communication with the communication destination device. Specifically, the communication destination-dictionary correspondence information 222 includes an ID 301, a communication destination 302, a device ID 303, a dictionary name 304, a dictionary number 305, a current dictionary number 306, a compression rate short-term average 307, and a compression rate long-term average 308. Is done.

  ID 301 is an identification number for identifying each row of the communication destination-dictionary correspondence information 222.

  The communication destination 302 is an IP address range in which data is compressed or expanded using a common dictionary. More specifically, the communication destination 302 indicates an IP address range that specifies a network to which a device that is a destination of an IP packet belongs. For example, in the case of a network system in which a plurality of networks such as a LAN are connected via the wide area network 140, the communication destination 302 stores a range of IP addresses indicating the network.

  A subnet mask is designated as the IP address range for the communication destination 302 in the rows 332 and 333 in FIG. 3A. In addition, the IP address of each device is designated as the IP address range for the communication destination 302 in the row 312 of FIG. 3B.

  It should be noted that the expression format of the value stored in the communication destination 302 may be any expression format as long as it can define an IP address range in which data is compressed or expanded using a common dictionary. The value stored in the communication destination 302 is assumed to be set by the administrator by means of a setting file or the like, but any setting method may be used. For example, a method of setting the value of the communication destination 302 by automatically recognizing each other during communication and exchanging information between the compression / decompression apparatuses can be considered.

  The device ID 303 is an ID of a compression / decompression device arranged on the device side corresponding to the communication destination 302. The value stored in the device ID 303 is assumed to be input by the administrator after the compression / decompression device is started, but any setting method may be used. For example, a method is conceivable in which the compression / decompression device automatically recognizes each other during communication and sets the value of the device ID 303 by exchanging information between the compression / decompression devices.

  The dictionary name 304 is a name of a dictionary used in communication with a device corresponding to the communication destination 302. The value of the dictionary name 304 is determined by the compression / decompression module 221.

  The number of dictionaries 305 is the total number of dictionaries defined with a dictionary name corresponding to the dictionary name 304. In this embodiment, one or more dictionaries are generated for one dictionary name.

  The current dictionary number 306 is the number of the currently used dictionary among the dictionaries corresponding to the dictionary name 304.

  The compression rate short-term average 307 represents a moving average of the data compression rate in the communication with the device corresponding to the communication destination 302 in the latest short-term.

  The length of the averaged period is assumed to be set by the administrator as a user-defined value by means of a setting file or the like. However, an appropriate value may be set in the apparatus in advance as a manufacturer-defined value. . In addition, as a way of expressing the period, a method expressed by the number of packets such as 100 packets is assumed here, but a method expressed by a time such as 1 minute may be used. As a method for calculating the average, a simple arithmetic average is assumed here, but a weighted average or a geometric average that increases the influence of the latest value may be used.

  The compression ratio long-term average 308 represents the moving average of the data compression ratio in the communication with the communication destination 302 over the latest long-term.

  The length of the averaged period is assumed to be given by the administrator as a user-defined value by means of a setting file or the like, but an appropriate value may be set in the apparatus in advance as a manufacturer-defined value. In addition, as a method of expressing the period, a method of expressing by the number of packets such as 100,000 packets is assumed, but a method of expressing by a time such as 1,000 minutes may be used. Moreover, as a method of taking the average, a simple arithmetic average is assumed here, but a weighted average or a geometric average that increases the influence of the latest value may be used.

  As illustrated in FIG. 3A, the communication destination-dictionary correspondence information 222 of the compression / decompression device C131 holds a row 332 corresponding to the client device A101 and a row 333 corresponding to the client device B102. On the other hand, as illustrated in FIG. 3B, the communication destination-dictionary correspondence information 222 of the compression / decompression device A <b> 111 holds a row 312 corresponding to the server 103.

  Here, line 332 and line 312 are lines relating to communication between the compression / decompression apparatus C131 and the compression / decompression apparatus A111. The paired compression / decompression apparatuses hold the same plurality of dictionaries, and compress or expand IP packets using the same dictionary. Therefore, the values of the number of dictionaries 305, the current dictionary number 306, the compression rate short-term average 307, and the compression rate long-term average 308 in the rows 333 and 312 are the same.

  FIG. 4 is an explanatory diagram illustrating an example of the dictionary management information 223 according to the first embodiment of this invention. FIG. 4 shows dictionary management information 223 held by the compression / decompression apparatus C131.

  The dictionary management information 223 is information for managing dictionary data held by the compression / decompression apparatus. Specifically, the dictionary management information 223 includes an ID 401, a dictionary name 402, a dictionary number 403, and dictionary data 404.

  ID 401 is an identification number for identifying each row of the dictionary management information 223.

  The dictionary name 402 is the same as the dictionary name 304 of the communication destination-dictionary correspondence information 222.

  The dictionary number 403 is an identification number of a currently used dictionary among dictionaries defined by a dictionary name corresponding to the dictionary name 402.

  In the present embodiment, an identification number is assigned only to the number of dictionaries 305 in the dictionary defined by the dictionary name corresponding to the dictionary name 304. Therefore, the dictionary number 403 stores the identification number of the currently used dictionary. For example, when the number of dictionaries 305 is “5”, there are five dictionaries defined by the dictionary name corresponding to the dictionary name 304, and numbers from “1” to “5” are assigned to each dictionary. .

  The dictionary data 404 is dictionary data of a dictionary corresponding to the dictionary name 402 and the dictionary number 403. In this embodiment, it is assumed that the internal data represented by the compression / decompression module 221 is stored in the dictionary data 404. The present invention is not limited to this, and the dictionary data 404 may store dictionary data in a tabular format as shown in FIG. When dictionary data is managed separately from the dictionary management information 223, the dictionary data 404 may store a pointer for accessing the dictionary data.

  The dictionary management information 223 is stored on the main storage device 202 for speeding up. Therefore, an upper limit is set for the size of each dictionary data 404. The upper limit of the size of the dictionary data 404 is a value set by the administrator using a setting file or the like. For example, “4 MB” is set. The upper limit of the size of the dictionary data 404 described above is an example, and may be a value larger or smaller than “4 MB”. Note that the upper limit of the size of the dictionary data 404 may be set using other means such as embedding in the compression / decompression module program 231 in advance.

  As shown in the communication destination-dictionary correspondence information 222, the compression / expansion device C131 has a dictionary name in communication with the subnet 112 of the communication destination “192.168.1.0/24” that passes through the compression / expansion device A111. In the communication with the communication destination “192.168.2.0/24” passing through the compression / decompression apparatus B121, the dictionary name is “Dictionary B”. Keep six dictionaries. Therefore, the dictionary management information 223 includes five rows 4401 to 4405 of the dictionary A and six rows 4406 to 4411 of the dictionary B.

  Using the communication destination-dictionary correspondence information 222 and the dictionary management information 223, the compression / decompression module 221 can uniquely acquire a dictionary used for communication with the communication destination device. For example, when the server 103 transmits an IP packet to the client apparatus A 101, the compression / expansion module 221 of the compression / expansion apparatus C131 acquires a dictionary according to the following procedure.

  The compression / decompression module 221 acquires the IP address of the communication destination device (client device A 101) from the IP header of the IP packet. Here, the IP address “192.168.1.101” is acquired.

  Next, the compression / decompression module 221 acquires a line in which the IP address acquired from the IP packet is included in the communication destination 302 from the communication destination-dictionary correspondence information 222. Here, row 332 is acquired.

  Next, the compression / decompression module 221 acquires the values of the dictionary name 304 and the current dictionary number 306 in the acquired row 332. Here, “Dictionary A” is acquired as the dictionary name, and “5” is acquired as the current dictionary number.

  Next, the compression / decompression module 221 refers to the dictionary management information 223 and searches for a row in which the dictionary name 402 and the dictionary number 403 match the values acquired from the dictionary name 304 and the current dictionary number 306 in the row 332. Further, the compression / decompression module 221 acquires dictionary data from the dictionary data 404 of the searched row. Here, row 4405 is retrieved.

  Next, details of processing executed by the compression / decompression module 221 will be described.

  5A and 5B are flowcharts for explaining IP packet transfer processing executed when the compression / decompression module 221 according to the first embodiment of the present invention receives an IP packet.

  FIG. 5A shows an IP packet transfer process accompanied by a compression process, which is executed when an IP packet is received from NIC1 (204). FIG. 5B shows an IP packet transfer process accompanied by a decompression process, which is executed when an IP packet is received from the NIC 2 (205).

  Any IP packet transfer processing is the same as the processing other than the compression processing or decompression processing, and will be described by comparison.

  First, IP packet transfer processing accompanied with compression processing will be described with reference to FIG. 5A.

  When the compression / decompression module 221 receives an IP packet from the NIC 1 (204) (step S511), the compression / decompression module 221 acquires a dictionary applied to the destination device of the IP packet. Specifically, a dictionary is acquired according to the procedure from step S512 to step S514.

  First, the compression / decompression module 221 obtains a destination IP address (Destination IP Address) from the IP header of the received IP packet (step S512).

  Next, the compression / decompression module 221 searches the communication destination 302 for a line including the destination IP address acquired in step S512 from the communication destination-dictionary correspondence information 222, and further includes the dictionary name 304 and the current dictionary of the line. The value of number 306 is acquired (step S513).

  The compression / decompression module 221 searches the dictionary management information 223, searches for a line in which the dictionary name 402 and the dictionary number 403 match the values of the dictionary name 304 and the current dictionary number 306 acquired in step S513, and further The dictionary data stored in the line dictionary data 404 is acquired and stored in the temporary area (step S514).

  The compression / decompression module 221 compresses the payload of the received IP packet using the dictionary data stored in the temporary area (step S515). At this time, the compression / decompression module 221 adds a new word to the dictionary data stored in the temporary area based on the transmitted communication content. The compression / decompression module 221 calculates a short-term compression rate and a long-term compression rate of the IP packet.

  When no new word is detected, no new word is added to the dictionary data. Also, when a new word is added and the size of the dictionary exceeds the upper limit of the size of the dictionary data 404, no new word is added. In addition, as another countermeasure for the case where the dictionary size exceeds the upper limit of the size of the dictionary data 404 by adding a new word, a method of deleting the least used word is also known.

  The compression / decompression module 221 transmits the IP packet including the compressed payload from the NIC 2 (205) to the device corresponding to the destination IP address (step S516).

  Thereafter, the compression / decompression module 221 executes a process for updating the dictionary management information 223 (step S517), and ends the process.

  The details of the update process of the dictionary management information 223 will be described later with reference to FIG. 8, but the dictionary data updated in step S515 is written back from the temporary area to the dictionary management information 223 in the update process of the dictionary management information 223. It is.

  Next, IP packet transfer processing accompanied with decompression processing will be described with reference to FIG. 7B.

  When the compression / decompression module 221 receives an IP packet from the NIC 2 (205) (step S521), the compression / decompression module 221 acquires a dictionary applied to the transmission destination apparatus of the IP packet. Specifically, a dictionary is acquired according to the procedure from step S523 to step S524.

  First, the compression / decompression module 221 acquires a source IP address (Source IP Address) from the IP header of the received IP packet (step S522).

  Next, as in step S513, the compression / decompression module 221 searches the communication destination-dictionary correspondence information 222 for a line including the transmission source IP address acquired in step S522 in the communication destination 302, and further, The values of the dictionary name 304 and the current dictionary number 306 are acquired (step S523).

  Similar to step S514, the compression / decompression module 221 refers to the dictionary management information 223, and finds a line in which the dictionary name 402 and the dictionary number 403 match the values of the dictionary name 304 and the current dictionary number 306 acquired in step S523. Further, the dictionary data stored in the dictionary data 404 is acquired from the row and stored in the temporary area (step S524).

  The compression / decompression module 221 expands the payload of the IP packet using the dictionary data stored in the temporary area (step S525). At this time, the compression / decompression module 221 adds a new word to the dictionary data stored in the temporary area based on the transmitted communication content. The compression / decompression module 221 calculates a short-term compression rate and a long-term compression rate of the IP packet. When no new word is detected, no new word is added to the dictionary data.

  The compression / decompression module 221 transmits the IP packet including the decompressed payload from the NIC1 (204) to the device corresponding to the destination IP address (step S526).

  Thereafter, the compression / decompression module 221 executes a process for updating the dictionary management information 223 (step S527), and ends the process.

  Details of the updating process of the dictionary management information 223 will be described later with reference to FIG. 8, but the dictionary data updated in step S525 is written back from the temporary area to the dictionary management information 223 in the updating process of the dictionary management information 223. It is.

  FIG. 6 is a flowchart illustrating the dictionary management information 223 update process executed by the compression / decompression module 221 according to the first embodiment of this invention.

  When the compression / decompression module 221 starts updating the dictionary management information 223 (step S601), the compression / decompression module 221 determines whether the size requirement is satisfied and the compression rate requirement is satisfied. (Step S602). Here, the determination process in step S602 will be described.

  The size requirement is that the size of the dictionary is greater than or equal to the first threshold. In this embodiment, the size of the dictionary assumes the number of words registered in the dictionary, but the number of bytes of the dictionary may be used.

  The first threshold value is a value set by the administrator using a setting file or the like, and for example, “30,000” is set as the number of words. The first threshold value described above is an example, and may be a value greater than “30,000” or a smaller value. The first threshold may be set using other means such as embedding in the compression / decompression module program 231 in advance.

  The compression ratio requirement is that the ratio of the short-term average compression ratio and the long-term average compression ratio (the ratio of the short-term average compression ratio to the long-term average compression ratio) is equal to or less than the second threshold. That is, the compression rate requirement is to determine whether the data after compression using the current dictionary or the data before being expanded using the current dictionary is data compressed to a sufficiently small size. . The short-term average compression ratio and the long-term average compression ratio are values that can be acquired from the short-term compression ratio average 307 and the long-term compression ratio average 308 of the communication destination-dictionary correspondence information 222.

  The second threshold is a value set by the administrator using a setting file or the like, and for example, “0.4 (40%)” is set as the ratio. The second threshold value described above is an example, and may be a value greater than “0.4” or a smaller value. Note that the second threshold may be set using other means such as embedding in the compression / decompression module program 231 in advance.

  The determination process in step S602 is a process for determining whether or not an appropriate dictionary is used for the current communication content. Specifically, the following determination is made.

  If the size requirement is satisfied and the compression rate requirement is satisfied, it indicates that an appropriate dictionary is not used for the current communication content. Therefore, in this case, the compression / decompression module 221 needs to use another dictionary or generate a new dictionary.

If the size requirement is satisfied but the compression rate requirement is not satisfied, this indicates that a dictionary appropriate for the current communication content is used.

  If the size requirement is not met, it indicates that the number of words in the newly generated dictionary is not sufficient. Immediately after a new dictionary is generated in step S631 described later, the compression rate does not increase because the number of words in the dictionary is not sufficient. Therefore, it is determined whether or not the size requirement is satisfied so as not to change the dictionary and generate a new dictionary until the number of words in the newly generated dictionary increases.

  The above is the description of the determination process in step S602.

  When it is determined that at least one of the size requirement and the compression rate requirement is not satisfied, the compression / decompression module 221 updates the dictionary data 404 of the dictionary management information 223 (step S611). This indicates that dictionary data appropriate for the current communication content is being used, or that new dictionary data has just been generated. Therefore, the compression / decompression module 221 updates the current dictionary and continues to use the dictionary.

  Specifically, the compression / decompression module 221 converts the latest dictionary data in which a word is newly added in step S515 or step S525 into the dictionary data 404 of the line retrieved from the dictionary management information 223 in step S514 or step S524. Store.

  The compression / decompression module 221 updates the compression rate short-term average 307 of the communication destination-dictionary correspondence information 222 (step S612), and updates the compression rate long-term average 308 of the communication destination-dictionary correspondence information 222 (step S613). The process ends. Specifically, the following processing is executed.

  In step S513 or step S523, the compression / decompression module 221 adds the short-term compression rate of the IP packet calculated in step S515 or step S525 to the compression rate short-term average 307 of the row acquired from the communication destination-dictionary correspondence information 222. Store.

  In addition, the compression / decompression module 221 performs the long-term compression of the IP packet calculated in step S515 or step S525 to the long-term compression ratio long-term average 308 of the row acquired from the communication destination-dictionary correspondence information 222 in step S513 or step S523. Stores the rate.

  The above is description of the process of step S612 and step S613.

  In step S602, when it is determined that the size requirement is satisfied and the compression rate requirement is satisfied, the compression / decompression module 221 acquires dictionary data different from the dictionary data currently used from the dictionary management information 223, The dictionary data is stored in the temporary area (step S621). That is, the compression / decompression module 221 acquires a dictionary that is not currently used.

  Specifically, the compression / decompression module 221 determines that the value of the dictionary name 402 from the dictionary management information 223 matches the dictionary name acquired in step S513 or step S523, and the value of the dictionary number 403 is set in step S513 or step S523. All lines different from the obtained current dictionary number are extracted. Further, the compression / decompression module 221 acquires dictionary data from the extracted dictionary data 404 of the row, and stores the acquired dictionary data in the temporary area.

  The compression / decompression module 221 calculates the compression rate of the payload of the IP packet using the dictionary data acquired in step S621 (step S622).

  Specifically, the compression / decompression module 221 compresses the payload of the IP packet received in step S511 or the IP packet transmitted in step S526, using the acquired dictionary data. Thereby, the compression / decompression module 221 can calculate the compression rate of the payload of the IP packet.

  The compression / decompression module 221 determines whether there is dictionary data in which the payload compression rate using other dictionary data is higher than the long-term average of the payload compression rate using the currently used dictionary (step S623). ). That is, it is determined whether the data after compression using another dictionary is data compressed to a sufficiently small size.

  Specifically, the compression / decompression module 221 compares the payload compression rate calculated in step S622 with the value of the long-term average compression rate 308 of the row acquired in step S513 or step S523, and calculates the calculated payload. It is determined whether or not the compression ratio is larger than the value of the compression ratio long-term average 308.

  When it is determined that there is no dictionary data whose payload compression rate using other dictionary data is higher than the long-term average of the payload compression rate using the currently used dictionary, the compression / decompression module 221 selects the dictionary management information. A line corresponding to the new dictionary data is added to 223 (step S631).

  At this time, the compression / decompression module 221 sets an arbitrary value for the ID 401 of the added row, sets the dictionary name acquired in step S513 or step S523 for the dictionary name 402, and sets the dictionary number 403 for step S513 or step. A value obtained by adding “1” to the value of the dictionary number 305 in the row acquired in S523 is set. Note that the dictionary data 404 of the added line is empty dictionary data.

  The compression / decompression module 221 updates the dictionary number 305 and the current dictionary number 306 of the communication destination-dictionary correspondence information 222 (step S632), and further updates the compression rate short-term average 307 to “0%” (step S633). Thereafter, the process proceeds to step S613.

  Specifically, the compression / decompression module 221 adds “1” to the value of the dictionary number 305 in the row acquired in step S513 or step S523. Further, the compression / decompression module 221 stores the updated value of the number of dictionaries 305 in the value of the current dictionary number 306 of the row.

  If it is determined in step S623 that there is dictionary data in which the compression rate of the payload using other dictionary data is higher than the long-term average of the compression rate of the payload using the currently used dictionary, the compression / decompression module 221 The current dictionary number 306 of the communication destination-dictionary correspondence information 222 is updated (step S641), the compression rate short-term average 307 is further updated (step S642), and then the process proceeds to step S613. This indicates that an IP packet having a communication content similar to the past was transmitted. In this case, communication is performed by switching to other dictionary data.

  Specifically, the compression / decompression module 221 sets the value of the current dictionary number 306 of the line acquired in step S513 or step S523 to a compression rate larger than the long-term average compression rate of the payload using the currently used dictionary. Update to dictionary data number. Further, the compression / decompression module 221 updates the value of the compression rate short-term average 307 of the row to the compression rate of the IP packet calculated in step S515 or step S525.

  If there is a plurality of dictionary data having a compression rate larger than the long-term average of the payload compression rate using the currently used dictionary in step S622, the compression / decompression module 221 selects the dictionary data with the highest compression rate. Shall.

  In the first embodiment, the network configuration for connecting the compression / decompression apparatuses has been described. However, the present invention is not limited to this. For example, a network configuration may be employed in which a compression device having only a compression function is connected to the client device side, and a decompression device having only the decompression function is connected to the server side.

  In the first embodiment, the compression / decompression apparatus is described as a separate apparatus from the client apparatus and the server, but the present invention is not limited to this. For example, the client device and the server may have a function included in the compression / decompression device.

  In the first embodiment, it is determined whether to switch the dictionary or to generate the dictionary using the compression ratio short-term average and the compression ratio long-term average. However, the present invention is not limited to this. For example, it may be determined that the compression ratio requirement is satisfied when the absolute value of the compression ratio short-term average is equal to or smaller than the second threshold.

As described above, according to the first embodiment, the compression / decompression apparatus holds a plurality of dictionaries, which are information for compressing or expanding IP packets, for one communication destination, and responds to communication contents. Thus, the communication content can be compressed or expanded using a dictionary with a high compression rate.

  Thereby, the compression rate of data transmitted / received to / from each communication destination device can be increased. Further, communication that can realize an optimal compression rate is possible by dynamically switching dictionaries or newly generating dictionaries according to communication contents.

  Furthermore, since a small-sized dictionary is used, all the dictionaries can be stored on the main storage device 202. Therefore, the I / O speed at the time of search can be increased.

  The second embodiment is different in that the compression rate is further increased by combining the entropy code with the lexicographic compression used by the compression / decompression apparatus according to the first embodiment. Hereinafter, the second embodiment will be described focusing on differences from the first embodiment.

  Since the configuration of the computer system of the second embodiment is the same as that of the first embodiment, description thereof is omitted. However, in the second embodiment, the configurations of the compression / expansion device A111, the compression / expansion device B121, and the compression / expansion device C131 are different.

  Hereinafter, the configuration of the compression / decompression apparatus according to the second embodiment will be described with reference to FIGS. The compression / decompression apparatus according to the second embodiment is different from the compression / decompression apparatus according to the first embodiment in that a function related to an entropy code is added. Therefore, the compression / decompression apparatus according to the second embodiment will be described focusing on differences from the compression / decompression apparatus according to the first embodiment.

  FIG. 7 is a block diagram illustrating a configuration example of the compression / decompression apparatus according to the second embodiment of the present invention. FIG. 7 shows a configuration example of the compression / decompression apparatus A111. Note that the compression / expansion device B121 and the compression / expansion device C131 have the same configuration.

  Since the hardware configuration of the compression / decompression apparatus A111 is the same as that of the first embodiment, the description thereof is omitted.

  The secondary storage device 203 according to the second embodiment stores a dictionary entropy compression / expansion module program 731 for realizing the dictionary entropy compression / expansion module 721.

  The CPU 201 reads the dictionary entropy compression / expansion module program 731 from the secondary storage device 203 and expands it on the main storage device 202 when the compression / expansion device A111 is activated. Further, the CPU 201 implements the dictionary entropy compression / expansion module 721 by executing the expanded dictionary entropy compression / expansion module program 731. The dictionary entropy compression / decompression module 721 expands communication destination-dictionary frequency table correspondence information 722 and dictionary frequency table management information 723, which will be described later, on the main storage device 202.

  The dictionary entropy compression / decompression module 721 compresses the payload of the IP packet received from the NIC1 (204) based on the communication destination-dictionary frequency table correspondence information 722 and the dictionary frequency table management information 723, which will be described later, and from the NIC2 (205). The IP packet including the compressed payload is transferred to the communication destination device. The compression / decompression module 221 expands the payload of the IP packet received from the NIC2 (205) based on the communication destination-dictionary frequency table correspondence information 722 and the dictionary frequency table management information 723, and expands the NIC1 (204). The IP packet including the payload is transferred to the communication destination device.

  FIG. 8 is an explanatory diagram illustrating an example of the communication destination-dictionary frequency table correspondence information 722 according to the second embodiment of the present invention. FIG. 8 shows communication destination-dictionary frequency table correspondence information 722 held by the compression / decompression apparatus C131.

  The communication destination-dictionary frequency table correspondence information 722 is obtained by changing the communication destination-dictionary correspondence information 222 of the first embodiment so as to manage the correspondence between the communication destination device, the dictionary, and the frequency table. Specifically, the dictionary name 304, dictionary number 305, and current dictionary number 306 of the communication destination-dictionary correspondence information 222 are changed to a dictionary frequency table name 804, a dictionary frequency table number 805, and a current dictionary frequency table number 806. The The other information strings are the same as those in the first embodiment.

  In the second embodiment, since the frequency table is switched simultaneously with the switching of the dictionary, the dictionary and the frequency table are managed on the same line. However, the dictionary and the frequency table may be switched at different timings. In this case, the information column of the communication destination-dictionary frequency table correspondence information 722 includes the dictionary name and the frequency table name, the number of dictionaries for managing those numbers, the number of frequency tables, the current dictionary number, and the current frequency table number. It is.

  FIG. 9 is an explanatory diagram illustrating an example of the dictionary frequency table management information 723 according to the second embodiment of this invention. FIG. 9 shows dictionary frequency table management information 723 held by the compression / decompression apparatus C131.

  The dictionary frequency table management information 723 has a structure in which frequency table data 905 is added to the dictionary management information 223 of the first embodiment in order to manage the frequency table.

  In the dictionary frequency table management information 723, the dictionary name 402 and the dictionary number 403 of the dictionary management information 223 of the first embodiment are changed to the dictionary frequency table name 902 and the dictionary frequency table number 903. The other information strings are the same as those in the first embodiment.

  As shown in communication destination-dictionary frequency table correspondence information 722, the compression / decompression device C113 uses a dictionary in communication with the subnet 112 of the communication destination “192.168.1.0/24” that passes through the compression / decompression device A111. Five dictionary data and five frequency table data constituting the dictionary frequency table management information whose frequency table name is “dictionary frequency table A” are stored, and the communication destination “192.168.2.0” passing through the compression / decompression device B121 is held. / 24 "subnet communication 122 holds six dictionary data and frequency table data constituting dictionary frequency table management information whose dictionary frequency table name is" dictionary frequency table B ". Therefore, the dictionary frequency table management information 723 includes five rows 9401 to 9405 of the dictionary frequency table A and six rows 9406 to 9411 of the dictionary frequency table B.

  The dictionary entropy compression / decompression module 721 can uniquely acquire a dictionary and a frequency table used for communication with a communication destination device using the communication destination-dictionary frequency table correspondence information 722 and the dictionary frequency table management information 723. . For example, when the server 103 transmits an IP packet to the client apparatus A 101, the dictionary entropy compression / expansion module 721 of the compression / expansion apparatus C131 acquires dictionary data and frequency table data according to the following procedure.

  The dictionary entropy compression / decompression module 721 acquires the IP address of the communication destination device (client device A 101) from the IP header of the IP packet. Here, the IP address “192.168.1.101” is acquired.

  Next, the dictionary entropy compression / decompression module 721 acquires a line in which the acquired IP address is included in the communication destination 302 from the communication destination-dictionary frequency table correspondence information 722. Here, row 832 is acquired.

  Next, the dictionary entropy compression / decompression module 721 acquires the values of the dictionary frequency table name 804 and the current dictionary frequency table number 806 in the acquired row 832. Here, “dictionary frequency table A” is acquired as the dictionary frequency table name, and “5” is acquired as the current dictionary frequency table number.

  Next, the dictionary entropy compression / decompression module 721 refers to the dictionary frequency table management information 723, and the dictionary frequency table name 902 and the dictionary frequency table number 903 are obtained from the row 832 and the dictionary frequency table name 804 and the current dictionary frequency table. The line that matches the value of the number 806 is searched. Here, line 9405 is searched.

  Further, the dictionary entropy compression / decompression module 721 acquires dictionary data and frequency table data from the dictionary data 404 and frequency table data 905 in the searched row 9405.

  Next, details of processing executed by the dictionary entropy compression / expansion module 721 will be described.

  FIGS. 10A and 10B are flowcharts illustrating IP packet transfer processing executed when the dictionary entropy compression / expansion module 721 according to the second embodiment of the present invention receives an IP packet.

  FIG. 10A shows an IP packet transfer process with a compression process, which is executed when an IP packet is received from NIC1 (204). FIG. 10B shows an IP packet transfer process accompanied by a decompression process, which is executed when an IP packet is received from NIC2 (205).

  Note that the IP packet transfer process executed by the dictionary entropy compression / decompression module 721 is basically the same procedure as the IP packet transfer process executed by the compression / decompression module 221. Therefore, the following description focuses on the differences. explain.

  When the dictionary entropy compression / decompression module 721 receives an IP packet from the NIC1 (204) (step S511), the dictionary entropy compression / expansion module 721 acquires a destination IP address (Destination IP Address) from the IP header of the IP packet (step S512).

  Next, the dictionary entropy compression / decompression module 721 searches the communication destination 302 for the line containing the destination IP address acquired in step S512 from the communication destination-dictionary frequency table correspondence information 722, and further, the dictionary frequency of the line. The values of the table name 804 and the current dictionary frequency table number 806 are acquired (step S1013).

  The dictionary entropy compression / decompression module 721 refers to the dictionary frequency table management information 723, and the dictionary frequency table name 902 and the dictionary frequency table number 903 are stored in the dictionary frequency table name 804 and the current dictionary frequency table number 806 acquired in step S1013. The line that matches the value is searched, and the dictionary data and frequency table data stored in the dictionary data 404 and the frequency table data 905 are acquired from the line and stored in the temporary area (step S1014).

  The dictionary entropy compression / decompression module 721 compresses the payload of the received IP packet using the dictionary data and frequency table data stored in the temporary area (step S1015). At this time, the dictionary entropy compression / decompression module 721 adds a new word to the dictionary data stored in the temporary area and updates the appearance frequency of the frequency table data based on the transmitted communication content. The dictionary entropy compression / decompression module 721 calculates a short-term compression rate and a long-term compression rate of the IP packet. When no new word is detected, no new word is added to the dictionary data.

  The dictionary entropy compression / decompression module 721 transmits an IP packet including the compressed payload from the NIC 2 (205) to a device corresponding to the destination IP address (step S516).

  Thereafter, the dictionary entropy compression / decompression module 721 executes an update process of the dictionary frequency table management information 723 (step S1017), and ends the process.

  Details of the updating process of the dictionary frequency table management information 723 will be described later with reference to FIG. 11, but the dictionary data and the frequency table data updated in step S1015 are stored in the temporary area in the updating process of the dictionary frequency table management information 723. The dictionary frequency table management information 723 is written back.

  Next, an IP packet transfer process accompanied by a decompression process will be described with reference to FIG. 10B.

  When the dictionary entropy compression / decompression module 721 receives an IP packet from the NIC 2 (205) (step S521), the dictionary entropy compression / expansion module 721 acquires a source IP address (Source IP Address) from the IP header of the received IP packet (step S522).

  Next, as in step S513, the dictionary entropy compression / expansion module 721 searches the communication destination-dictionary frequency table correspondence information 722 for a line including the transmission source IP address acquired in step S522 in the communication destination 302. Further, the values of the dictionary frequency table name 804 and the current dictionary frequency table number 806 in the row are acquired (step S1023).

  Similar to step S1014, the dictionary entropy compression / decompression module 721 refers to the dictionary frequency table management information 723, and the dictionary frequency table name 902 and the dictionary frequency table number 903 are the dictionary frequency table name 804 acquired in step S1023 and the current one. A line that matches the value of the dictionary frequency table number 806 is searched, and the dictionary data and frequency table data stored in the dictionary data 404 and the frequency table data 905 are acquired from the line and stored in the temporary area (step S1024). ).

  Similar to step S1015, the dictionary entropy compression / decompression module 721 decompresses the payload of the received IP packet using the dictionary data and frequency table data stored in the temporary area (step S1025). At this time, the dictionary entropy compression / expansion module 721 adds a new word to the dictionary data stored in the temporary area and updates the appearance frequency of the frequency table data based on the transmitted communication content. The dictionary entropy compression / decompression module 721 calculates a short-term compression rate and a long-term compression rate of the IP packet. When no new word is detected, no new word is added to the dictionary data.

  The dictionary entropy compression / decompression module 721 transmits an IP packet including the decompressed payload from the NIC 1 (204) to a device corresponding to the destination IP address (step S526).

  Thereafter, the dictionary entropy compression / decompression module 721 executes an update process of the dictionary frequency table management information 723 (step S1027), and ends the process.

  Note that the updating process of the dictionary frequency table management information 723 will be described later with reference to FIG. 11. However, the dictionary data and the frequency table data updated in step S1025 are updated from the temporary area in the updating process of the dictionary frequency table management information 723. The frequency table management information 723 is written back.

  FIG. 11 is a flowchart for explaining the updating process of the dictionary frequency table management information 723 executed by the dictionary entropy compression / expansion module 721 according to the second embodiment of the present invention.

  The update process of the dictionary frequency table management information 723 executed by the dictionary entropy compression / decompression module 721 is basically the same procedure as the update process of the dictionary management information 223 executed by the compression / decompression module 221. The explanation will focus on the differences.

  When the dictionary entropy compression / decompression module 721 starts updating the dictionary frequency table management information 723 (step S1101), the dictionary entropy compression / decompression module 721 determines whether the size requirement is satisfied and the compression rate requirement is satisfied (step S602).

  When it is determined that at least one of the size requirement and the compression rate requirement is not satisfied, the dictionary entropy compression / decompression module 721 updates the dictionary data 404 and the frequency table data 905 of the dictionary frequency table management information 723 (step S1111). ). This indicates that appropriate dictionary data and frequency table data are used for the current communication content, or that new dictionary data and frequency table data have just been generated. Therefore, the dictionary entropy compression / decompression module 721 continues to use the current dictionary data and frequency table data.

  Specifically, the dictionary entropy compression / expansion module 721 generates the dictionary data in which words are newly added in step S1015 or step S1025 and the frequency table data in which the frequency is updated, and the dictionary frequency table management information in step S1014 or step S1024. Stored in the dictionary data 404 and the frequency table data 905 of the row retrieved from 723.

  The dictionary entropy compression / decompression module 721 updates the compression rate short-term average 307 of the communication destination-dictionary frequency table correspondence information 722 (step S1112), and updates the compression destination long-term average 308 of the communication destination-dictionary frequency table correspondence information 722. (Step S1113) and the process is terminated. Specifically, the following processing is executed.

  In step S1013 or step S1023, the dictionary entropy compression / decompression module 721 sets the short-term average of the IP packet calculated in step S1015 or step S1025 to the line compression rate short-term average 307 acquired from the communication destination-dictionary frequency table correspondence information 722. Stores the compression ratio.

  In addition, the dictionary entropy compression / decompression module 721 calculates the IP packet calculated in step S1015 or step S1025 to the long-term average compression rate 308 of the row acquired from the communication destination-dictionary frequency table correspondence information 722 in step S1013 or step S1023. Stores the long-term compression ratio.

  The above is the description of the processing in step S1112 and step S1113.

  If it is determined in step S602 that the size requirement is satisfied and the compression rate requirement is satisfied, the dictionary entropy compression / decompression module 721 determines from the dictionary frequency table management information 723 that the dictionary is different from the dictionary data currently used. Data and frequency table data different from the currently used frequency table data are acquired, and the dictionary data and frequency table data are stored in the temporary area (step S1121). That is, the dictionary entropy compression / decompression module 721 acquires dictionary data and frequency table data that are not currently used.

  Specifically, the dictionary entropy compression / decompression module 721 determines that the value of the dictionary frequency table name 902 from the dictionary frequency table management information 723 matches the dictionary frequency table name acquired in step S1013 or step S1023, and the dictionary frequency table. All rows in which the value of the number 903 is different from the current dictionary frequency table number acquired in step S1013 or step S1023 are extracted. Further, the dictionary entropy compression / decompression module 721 acquires dictionary data and frequency table data from the extracted dictionary data 404 and frequency table data 905, and stores them in the temporary area.

  The dictionary entropy compression / decompression module 721 calculates the compression rate of the payload of the IP packet using the dictionary data and frequency table data acquired in step S1121 (step S1122).

  Specifically, the dictionary entropy compression / decompression module 721 compresses the payload of the IP packet received in step S511 or the IP packet transmitted in step S526, using the acquired dictionary data and frequency table data. Thereby, the dictionary entropy compression / decompression module 721 can calculate the compression rate of the payload of the IP packet.

  The dictionary entropy compression / decompression module 721 determines that the compression rate of the payload compressed using the dictionary data and frequency table data acquired in step S1121 is the payload compression rate using the dictionary data and frequency table data currently used. It is determined whether dictionary data and frequency table data higher than the long-term average exist (step S1123).

  Specifically, the dictionary entropy compression / decompression module 721 compares the payload compression rate calculated in step S1122 with the value of the long-term average compression rate 308 of the row acquired in step S1013 or step S1023. It is determined whether or not the compression rate of the payload is larger than the value of the compression rate long-term average 308.

  When it is determined that there is no dictionary data and frequency table data whose payload compression rate using other dictionary data is higher than the long-term average of the payload compression rate using the dictionary data and frequency table data currently used, The dictionary entropy compression / decompression module 721 adds new dictionary data and a row corresponding to the frequency table data to the dictionary frequency table management information 723 (step S1131).

  At this time, the dictionary entropy compression / decompression module 721 stores an arbitrary value in the ID 401 of the added row, stores the dictionary frequency table name acquired in step S1013 or step S1023 in the dictionary frequency table name 902, and stores the dictionary frequency A value obtained by adding “1” to the value of the dictionary frequency table count 805 in the row acquired in step S1013 or step S1023 is stored in the table number 903. Note that empty dictionary data is stored in the dictionary data 404 of the added row, and empty frequency table data is stored in the frequency table data 905.

  The dictionary entropy compression / decompression module 721 updates the dictionary frequency table number 805 and the current dictionary frequency table number 806 of the communication destination-dictionary frequency table correspondence information 722 (step S1132), and further sets the compression rate short-term average 307 to “0%”. (Step S1133), and then the process proceeds to step S1113.

  Specifically, the dictionary entropy compression / decompression module 721 adds “1” to the value of the dictionary frequency table number 805 in the row acquired in step S1013 or step S1023. Further, the dictionary entropy compression / decompression module 721 stores the updated value of the dictionary frequency table number 805 in the current dictionary frequency table number 806 of the row.

  In step S1123, it is determined that there is dictionary data and frequency table data in which the compression rate of the payload using other dictionary data is higher than the long-term average compression rate of the payload using the dictionary data and frequency table data currently used. In this case, the dictionary entropy compression / decompression module 721 updates the current dictionary frequency table number 806 of the communication destination-dictionary frequency table correspondence information 722 (step S1141), and further updates the compression rate short-term average 307 (step S1142). Thereafter, the process proceeds to step S613. This indicates that an IP packet having a communication content similar to the past was transmitted. In this case, communication is performed by switching to other dictionary data and frequency table data.

  Specifically, the dictionary entropy compression / decompression module 721 uses the value of the current dictionary frequency table number 806 of the line acquired in step S1013 or step S1023 as the payload of the dictionary data currently used and the frequency table data. The dictionary frequency table number corresponding to the dictionary data and frequency table data having a compression rate larger than the long-term average compression rate is updated. Further, the dictionary entropy compression / decompression module 721 updates the value of the compression rate short-term average 307 of the row to the compression rate of the IP packet calculated in step S1015 or step S1025.

  In step S1122, if there are a plurality of dictionary data and frequency table data having a compression ratio larger than the long-term average of the payload compression ratio using the dictionary data and frequency table data currently used, the dictionary entropy compression / decompression module 721 The dictionary data and frequency table data with the highest compression rate are selected.

  In the second embodiment, the IP packet is compressed or expanded using the dictionary and the frequency table. However, the IP packet may be compressed or expanded using only the frequency table. That is, an embodiment using a frequency table instead of the dictionary in the first embodiment may be used.

  As described above, according to the second embodiment, the compression / decompression apparatus holds a plurality of dictionaries and a frequency table for a communication destination, and uses a dictionary and a frequency table having a high compression rate according to communication contents. Communication contents can be compressed.

  Thereby, the compression rate of the data transmitted / received to / from each communication destination can be increased. Further, by dynamically switching the dictionary and the frequency table according to the communication content, or by newly generating the dictionary and the frequency table, it is possible to realize communication capable of realizing the optimum compression rate.

  Furthermore, since the dictionary is generated from the communication content (payload) included in the IP packet, the dictionary becomes a small size, and all the dictionaries can be stored on the main storage device 202. Therefore, the I / O speed at the time of search can be increased.

  The third embodiment is different from the first embodiment in that the storage capacity of the memory of the compression / decompression apparatus has an upper limit, so that a storage capacity for storing the dictionary is secured by deleting a dictionary that is less necessary during operation. . Hereinafter, the third embodiment will be described focusing on differences from the first embodiment. Although the present embodiment is an extension of the system of the first embodiment, the system of the second embodiment can be expanded in the same manner to cope with a case where there is an upper limit in the storage capacity for storing the dictionary and the frequency table.

  When there is an upper limit on the storage capacity for storing the dictionary, the compression / decompression module 221 may not be able to generate a new dictionary or store a new dictionary due to insufficient storage capacity in the update processing of the dictionary management information 223. .

  Therefore, the compression / decompression module 221 deletes a dictionary with low necessity for each dictionary corresponding to the dictionary name 402 of the dictionary management information 223 at an appropriate timing.

  In addition, as a timing which deletes a dictionary, periodic timing, such as a day unit, can be considered, for example. Other examples of timing for deleting a dictionary include timing when a new line is added to the communication destination-dictionary correspondence information 222, timing when a new dictionary is generated, and the like.

  Further, in this embodiment, a low-necessity dictionary is specified using the LRU-k method, which is a cache memory page replacement algorithm disclosed in Non-Patent Document 5. In this embodiment, for simplicity, the parameter k of LRU-k is described as “k = 1”, but k may be “2” or more. In addition to LRU, other page replacement algorithms such as FIFO and 2Q can be used.

  FIG. 12 is an explanatory diagram illustrating an example of the dictionary management information 223 according to the third embodiment of this invention. The dictionary management information 223 shown in FIG. 12 corresponds to LRU-1 having the parameter k = 1 as described above.

  In the dictionary management information 223 of the third embodiment, an information string of the last access time 1205 is newly added. The last access time 1205 stores the time when the dictionary data was updated. For example, the time when the dictionary data is updated in step S611 is stored, and the time when the line corresponding to the new dictionary data is added to the dictionary management information 223 is stored in step S631.

  In addition to the last access time 1205, the dictionary management information 223 corresponding to LRU-2 is obtained by adding an information string for managing the access time immediately before the last access time. Further, instead of the last access time 1205, a queue is virtually assumed, and an information string for managing the number representing the order in the queue is added, whereby the dictionary management information 223 corresponding to the FIFO is obtained. Furthermore, assuming two or three virtual queues, and adding two or three columns for managing numbers representing the order of them, dictionary management information 223 corresponding to 2Q is obtained.

  FIGS. 13A and 13B are flowcharts illustrating dictionary deletion processing according to the third embodiment of the present invention.

  In the following description, it is assumed that the value of the dictionary name 402 of the dictionary management information 223 corresponding to LRU-1 is designated as “a”. Note that any one of the values of the dictionary name 304 of the communication destination-dictionary correspondence information 222 is assigned to the value “a” of the dictionary name 402. Therefore, a list of values of the dictionary name 304 is acquired from the communication destination-dictionary correspondence information 222, and a dictionary having the dictionary name “a” is repeatedly executed on the list, whereby a dictionary for each communication destination is obtained. Dictionaries that are less needed are deleted.

  When the compression / decompression module 221 starts the deletion process of the dictionary (step S1301), the compression / decompression module 221 acquires a line whose dictionary name 402 is “a” from the dictionary management information 223 corresponding to LRU-1 (step S1302). For example, the line “a = dictionary A” is acquired.

  The compression / decompression module 221 refers to the last access time 1205 of the acquired row, and deletes the row with the oldest last access time 1205 from the dictionary management information 223 (step S1303). Here, when specifying a dictionary with low necessity using a method other than LRU-1, such as FIFO or 2Q, the processing contents are changed to correspond to each method.

  The compression / decompression module 221 updates the number of dictionaries 305 and the current dictionary number 306 in the row where the dictionary name 304 of the communication destination-dictionary correspondence information 222 is “a” (step S1304).

  Specifically, the compression / decompression module 221 subtracts “1” from the value of the number of dictionaries 305 in the row whose dictionary name 304 is “a”. With respect to the value of the current dictionary number 306 in the row, if the deleted dictionary is a dictionary that matches the value of the current dictionary number 306, the value of the current dictionary number 306 is subtracted by “1”, and the deleted dictionary becomes the current dictionary number. When the dictionary does not match the value of the dictionary number 306, the current dictionary number 306 is not particularly updated.

  Next, the compression / decompression module 221 identifies the opposing device that holds the deleted dictionary (step S1305). Here, the opposite device indicates a destination compression / decompression device that performs communication using the deleted dictionary.

  Specifically, the compression / decompression module 221 refers to the row where the dictionary name 304 of the communication destination-dictionary correspondence information 222 is “a”, and acquires the device ID of the opposite device from the device ID 303 of the row.

  The compression / decompression module 221 transmits a message including information on the row deleted from the dictionary management information 223 to the opposite device (step S1306). As information of the deleted line, information such as a dictionary name 402 and dictionary number 403 paired (a, dictionary number) can be considered.

  The compression / decompression module 221 of the opposite device that has received the message starts a dictionary deletion process (step S1311).

  The compression / decompression module 221 deletes the line in which the dictionary name 402 and the dictionary number 403 match (a, dictionary number) included in the message from the dictionary management information 223 corresponding to LRU-1 (step S1312).

  The compression / decompression module 221 updates the number of dictionaries 305 and the current dictionary number 306 in the row where the dictionary name 304 of the communication destination-dictionary correspondence information 222 is “a” (step S1313).

  Specifically, the compression / decompression module 221 subtracts “1” from the value of the number of dictionaries 305 in the row whose dictionary name 304 is “a”. With respect to the value of the current dictionary number 306 in the row, if the deleted dictionary is a dictionary that matches the value of the current dictionary number 306, the value of the current dictionary number 306 is subtracted by “1”, and the deleted dictionary becomes the current dictionary number. When the dictionary does not match the value of the dictionary number 306, the current dictionary number 306 is not particularly updated.

  As described above, according to the third embodiment, the compression / decompression module 221 deletes a dictionary that is less necessary during operation so that the storage capacity for storing the dictionary does not exceed the upper limit, thereby increasing the storage capacity. It is possible to manage the dictionary according to the upper limit of.

  In each embodiment, an example using control by software has been described. However, part of the control can be realized by hardware.

  Although the present invention has been described in detail with reference to the accompanying drawings, the present invention is not limited to such specific configurations, and various modifications and equivalents within the spirit of the appended claims Includes configuration.

101, 102 Client device 103 Server 104 Wide area network 111, 121, 131 Compression / decompression device 112, 122 Subnet 201 CPU
202 Main storage device 203 Secondary storage device 221 Compression / decompression module 222 Communication destination-dictionary correspondence information 223 Dictionary management information 231 Compression / decompression module program 721 Dictionary entropy Compression / decompression module 722 Dictionary frequency table correspondence information 723 Dictionary frequency table management information 731 Dictionary entropy Compression / decompression module program 804 Dictionary frequency table name 805 Dictionary frequency table number 806 Current dictionary frequency table number 905 Frequency table data 905 Dictionary frequency table number

Claims (10)

A communication device comprising a processor, a main storage device connected to the processor, and a network interface connected to the processor, and transferring data transmitted and received between a plurality of computers,
The main storage device
Stores multiple compression / decompression information for compressing or decompressing data for each communication destination,
Including a temporary area for storing duplicate compression / decompression information, which is a copy of the compression / decompression information used for communication with an arbitrary communication destination ,
The communication device includes a compression / decompression unit that compresses or decompresses received data using the compression / decompression information,
The compression / decompression unit
Manage the information identifying the communication destination and the identification information of the compression / decompression information to be used ,
When data is received from the first communication destination, the first information associated with the information specifying the first communication destination is selected from among a plurality of compression / decompression information used for communication with the first communication destination . 1 compression / decompression information is specified, and first copy compression / decompression information that is a copy of the first compression / decompression information is stored in the temporary area,
Using the first copy compression / decompression information , compress or decompress the data received from the first communication destination ,
Sending the compressed data or the decompressed data;
Updating the first copy compression / decompression information based on the compression result or decompression result of the data received from the first communication destination;
Determines whether satisfies the switching condition,
When it is determined that the switching condition is not satisfied, the first compression / decompression information is updated by overwriting the first compression / decompression information on the first compression / decompression information,
When it is determined that the switching condition is satisfied, second compression / decompression information is selected from a plurality of compression / decompression information used for communication with the first communication destination, and the first communication destination is specified. The communication apparatus switches the compression / decompression information used for communication with the first communication destination by associating the information to be performed with the second compression / decompression information . The communication device according to claim 1,
The compression / decompression unit
Calculating a compression rate of data using the first copy compression / decompression information , calculating an index for evaluating a size of data compressed using the first compression / decompression information based on the compression rate;
Determining whether the size of the first compression / decompression information is equal to or larger than a first threshold and the index is equal to or smaller than a second threshold ;
A communication apparatus , wherein when the size of the first compression / decompression information is equal to or larger than a first threshold and the index is equal to or smaller than a second threshold, it is determined that the switching condition is satisfied .   The communication device according to claim 2,
  The compression / decompression unit
  When it is determined that the switching condition is satisfied, compression of data stored in the main storage device and compressed using each of a plurality of compression / decompression information used for communication with the first communication destination A communication apparatus that calculates a rate and selects compression / decompression information with the compression rate larger than a third threshold as the second compression / decompression information.   The communication device according to claim 3,
  The compression / decompression unit
  If there is no compression / decompression information for which the calculated compression rate is greater than a third threshold, new third compression / decompression information is generated,
  A communication apparatus, characterized in that information for specifying the first communication destination is associated with identification information for the third compression / decompression information.   The communication device according to claim 1,
  The compression / decompression unit
  Access indicating information for specifying the communication destination, identification information of a plurality of compression / decompression information used for communication with the communication destination, and time when accessing a plurality of compression / decompression information used for communication with the communication destination Manage time in association,
  In communication with the first communication destination, the compression / decompression information with the oldest access time is selected from the plurality of compression / decompression information associated with the first communication destination, and the compression / decompression information is selected. Deleting from the main storage,
  A communication apparatus that notifies the communication apparatus included in each of the plurality of communication destinations of information regarding the deleted compression / decompression information.   A data transfer control method in a communication device that includes a processor, a main storage device connected to the processor, and a network interface connected to the processor, and transfers data transmitted and received between a plurality of computers,
  The main storage device
  Stores multiple compression / decompression information for compressing or decompressing data for each communication destination,
  Including a temporary area for storing duplicate compression / decompression information, which is a copy of the compression / decompression information used for communication with an arbitrary communication destination,
  The communication device includes a compression / decompression unit that compresses or decompresses received data using the compression / decompression information,
  The compression / decompression unit manages information specifying a communication destination and identification information of compression / decompression information to be used in association with each other,
  The data transfer control method includes:
  When the compression / decompression unit receives data from the first communication destination, information for specifying the first communication destination from among a plurality of compression / decompression information used for communication with the first communication destination A first step of specifying the first compression / decompression information associated with the first compression / decompression information, which is a copy of the first compression / decompression information, in the temporary area;
  A second step in which the compression / decompression unit compresses or decompresses data received from the first communication destination using the first copy compression / decompression information;
  A third step in which the compression / decompression unit transmits the compressed data or the decompressed data;
  A fourth step in which the compression / decompression unit updates the first copy compression / decompression information based on a compression result or a decompression result of data received from the first communication destination;
  A fifth step of determining whether or not the compression / decompression unit satisfies a switching condition;
  When the compression / decompression unit determines that the switching condition is not satisfied, the first compression / decompression information is updated by overwriting the first compression / decompression information on the first compression / decompression information. A sixth step;
  When it is determined that the compression / decompression unit satisfies the switching condition, second compression / decompression information is selected from a plurality of compression / decompression information used for communication with the first communication destination, and the first A seventh step of switching compression / decompression information used for communication with the first communication destination by associating information specifying one communication destination with the second compression / decompression information;
  A data transfer control method comprising:   The data transfer control method according to claim 6,
  The fifth step includes
  The compression / decompression unit calculates a compression rate of data using the first copy compression / decompression information, and evaluates a size of data compressed using the first compression / decompression information based on the compression rate. Calculating a metric,
  The compression / decompression unit determines whether the size of the first compression / decompression information is equal to or greater than a first threshold and the index is equal to or smaller than a second threshold;
  Determining that the compression / decompression unit satisfies the switching condition when the size of the first compression / decompression information is equal to or larger than a first threshold and the index is equal to or smaller than a second threshold. A data transfer control method characterized by the above.   The data transfer control method according to claim 7,
  In the seventh step, the compression / decompression unit stores data stored in the main storage device and compressed using each of a plurality of compression / decompression information used for communication with the first communication destination. And a step of selecting the compression / decompression information with the compression ratio larger than a third threshold as the second compression / decompression information.   The data transfer control method according to claim 8, wherein
  The seventh step includes
  When there is no compression / decompression information for which the calculated compression ratio is greater than a third threshold, the compression / decompression unit newly generates third compression / decompression information; and
  A data transfer control method comprising: a step in which the compression / decompression unit associates information specifying the first communication destination with identification information of the third compression / decompression information.   The data transfer control method according to claim 6,
  The compression / decompression unit includes information identifying the communication destination, identification information of a plurality of compression / decompression information used for communication with the communication destination, and a plurality of compression / decompression information used for communication with the communication destination. Associating and managing access times indicating access times;
  In the communication with the first communication destination, the compression / decompression unit selects the compression / decompression information with the oldest access time from among a plurality of compression / decompression information associated with the first communication destination. Deleting the compression / decompression information from the main storage device;
  A data transfer control method comprising: a step in which the compression / decompression unit notifies another communication device included in each of the plurality of communication destinations of information regarding the deleted compression / decompression information.

Images