JP3848209B2 - Data transfer device, data transfer method and program - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a data transfer apparatus, a data transfer method, and a program for transferring data for another apparatus.
[0002]
[Prior art]
2. Description of the Related Art A client / server type information system is widely used, which includes a server that provides various services via a network and a client that requests a desired service from the server. In particular, a World Wide Web system (or simply referred to as Web) composed of a Web server and a client that communicate using the HTTP protocol on the Internet is a client-server type information system that is very widely used. Normally, a server program operates on the server, and a predetermined tool (program) such as a browser operates on the client. The contents of services provided on the Internet are also diverse, providing information such as text, still images, moving images, audio (such as homepages, e-mails, digital contents, etc.) and programs via the network. Various services such as distribution or transfer services, electronic store services for selling products, reservation services for seats and rooms, mediation services for various contracts, etc., and new forms of services one after another Has appeared.
[0003]
By the way, in a client server type information system such as the Web, the service is basically provided by transferring data between the client and the server regardless of the form of the provided service. Is done. Therefore, the capacity (bandwidth) of the network used for communication between the client and the server tends to become a bottleneck of the entire system. Therefore, a cache technique is usually used to reduce the load on the network.
[0004]
In the case of a Web system, many browsers or the like that operate on a client use a cache mechanism, and cache recently accessed data. On the Web, access is made by designating information and services with names called URLs, so the cache on the client is the information that has been requested from the Web server in the past and the data returned as a result of the service that can be cached The URL is recorded in the cache in association with the URL. In this case, when there is a request for information or a service with the same URL as that in the cache, and it can be determined that the response data in the cache is not stale, by returning the data, Communication between can be eliminated.
[0005]
When there are a plurality of users in a corporate office LAN, a research institution LAN, or a home LAN, a proxy server may be placed between the LAN and the Internet, and a cache mechanism may be provided in the proxy server. Many. The cache in the client (for example, the browser cache) operates as a cache dedicated to the client user, while the proxy server cache on the LAN operates as a shared cache for a plurality of client users. For this reason, in the latter case, the cache is effective even when accessing a URL accessed by another person (another client) in the past.
[0006]
Now, on the Web, communication is performed between a client and a server using a protocol called HTTP. The HTTP protocol is a combination of a “request message” sent from the client to the server and a “reply message” that returns a response from the server to the client.
[0007]
The request message consists of a “request header” and a “request body”. The request header includes a URL for designating information to be accessed and a service, a method name indicating the type of access, and other various information necessary for access. The data to be sent to the server is entered in the request body. The data contained in the request body is also called “request data”.
[0008]
The reply message includes a “reply header” and a “reply body”. Information such as the status of the processing result is entered in the reply header, and data such as the requested information and the processing result of the requested service is entered in the reply body. The data contained in the reply body is also called “reply data”.
[0009]
As a request message method, there are a “GET method” that reads information on the server, a “PUT method” that writes data held by the user to the server, and a “POST method” that sends back the result processed in response to the request. It is the main thing used to access services. In addition, a method such as DELETE is defined.
[0010]
In many cases, the request body of the request message of the GET method and the reply body of the reply message of the PUT method are empty. The request body of the POST method request message contains information used for processing on the server side as necessary, and the reply body of the POST method reply message contains data of the processing result.
[0011]
Data read from the server by the GET method can be divided into “dynamic data” generated on the server side each time it is read, and “static data” that returns data already stored on the server side as it is. Of these, the contents of dynamic data may be different each time the same URL is read, and in many cases, the server sends back a specification indicating that caching is impossible in the header of the reply message. Therefore, it is the static data portion that is the target of caching in the Web data. This static data can be divided into “shared data” that can be referenced by an unspecified number of users and “private data” that performs access control so that only specific users can access by authenticating users. . The former shared data can be cached in any cache. However, the latter private data cannot be cached in a shared cache such as a proxy server (because private data must be authenticated and sent back by the server). However, in the case of a personal cache such as a browser, even private data can be cached.
[0012]
Since the POST method returns the result of processing on the server side, in general, the server sends back the result with a specification indicating that caching is impossible in the header of the reply message. Therefore, it is not normally subject to caching.
[0013]
Since the PUT method sends data to the server, the cache does nothing.
[0014]
[Problems to be solved by the invention]
The conventional Web cache targets static content as a cache target. In the past, information and services published on the Web were not so frequently updated, and many were published to an unspecified number of people, so the proportion of static content was very high. Cache technology was also effective in reducing network load.
[0015]
However, as a system in which a user accesses information and services on a server via a network using a Web browser, such as a Web-based ASP (Application Service Provider), has been popularized as described below. The data that cannot be handled is increasing.
-Since the user is authenticated and the user who can access is restricted, there is much private data.
-Many dynamic data are generated by referring to the backend database.
-In many cases, the POST method is used for form processing and search.
-The PUT method is often used to share information within a group.
As a result, the cache technology alone has not functioned effectively as a technique for reducing the load on the network.
[0016]
The present invention has been made in consideration of the above circumstances, and provides a data transfer device, a data transfer method, and a program provided with a cache technology and a compression technology that can further reduce the load on the network connecting the data transfer devices. The purpose is to provide.
[0017]
[Means for Solving the Problems]
The present invention relates to a data transfer device that relays communication between a server device and another data transfer device that relays data transmitted from the server device to a client device, and receives the data from the server device to another data transfer device. Correspondence between transmitted first complete data, first fingerprint generated based on the first complete data, and first identification information indicating that the first complete data is complete data A holding means for attaching and holding, a receiving means for receiving the second complete data transmitted by the server device to the client device, and a difference that is a difference between the first complete data and the second complete data The data, the second fingerprint generated based on the second complete data, and the second identification information indicating that the difference data is difference data, Causes held in lifting means and said difference data and the second identification information, characterized by comprising a transmitting means for transmitting the client device as a destination.
The present invention also relates to another data transfer device that relays data transmitted from the server device to the client device and a data transfer device that relays communication with the client device, and receives the client from the other data transfer device and First complete data transmitted to the device, a first fingerprint generated based on the first complete data, and first identification information indicating that the first complete data is complete data; Storing means for associating and holding, difference data that is a difference between the first complete data and the second complete data, which is destined for the client device, and that the difference data is difference data. Receiving means for receiving the second identification information, and generating the second complete data using the first complete data and the differential data, and the differential data, The second fingerprint generated on the basis of the complete data and the second identification information are associated with each other and held in the holding unit, and the second complete data is transmitted to the client device as the destination And transmitting means.
[0023]
The present invention relating to the apparatus is also established as an invention relating to a method, and the present invention relating to a method is also established as an invention relating to an apparatus.
Further, the present invention relating to an apparatus or a method has a function for causing a computer to execute a procedure corresponding to the invention (or for causing a computer to function as a means corresponding to the invention, or for a computer to have a function corresponding to the invention. It is also established as a program (for realizing) and also as a computer-readable recording medium on which the program is recorded.
[0024]
According to the present invention, the correspondence between the data and its name is held between the data transfer devices, and the data having this correspondence is transferred by transferring the corresponding name instead of transferring the data body. The amount of data transferred between transfer devices can be reduced. For example, even if the reply message of the GET method is private data, it can be compressed between fingerprints and transferred between data transfer apparatuses. For example, even if the reply message of the GET method is dynamic data, if the content is the same, it can be compressed between fingerprints and transferred between data transfer apparatuses. Further, for example, even in the case of the POST method, if the result is the same data, the data can be transferred between the data transfer devices by compressing the data with a fingerprint.
[0025]
Further, according to the present invention, since the name corresponding to the data is not held, the data is transferred. Cost Even if the corresponding name cannot be transferred instead, the compressed data expressed by compressing the data using the name corresponding to the stored reference data is transferred between the data transfer devices. The amount of transfer data can be reduced. For example, if the reply data of the GET method or the POST method is partially different from the previously accessed data, the data amount can be reduced by differential transfer. Also, for example, if the PUT method or POST method request data is partially different from the previously sent data, the amount of data can be reduced by differential transfer.
[0026]
Further, according to the present invention, when data compression is possible, the storage device (memory, hard disk, etc.) can be effectively used by storing the compressed data instead of the data.
[0027]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the invention will be described with reference to the drawings.
[0028]
In the following, a case where the WAN is the Internet and the client is connected to the user office LAN and the HTTP protocol is used will be described as an example. Even if the client is installed in a home LAN other than the office, for example, or a protocol using a protocol other than HTTP is applicable.
[0029]
FIG. 49 shows a basic configuration example of a computer network system to which the present invention is applied. In this configuration example, a local area network (LAN) 12 in the ASP server center 2 and a local area network (LAN) 16 in the user office 4 are connected via a wide area network (WAN) 14 such as the Internet or a dedicated line. The server 20 and the client 50 can communicate with each other via the LAN 12, the WAN 14, and the LAN 16. One or more servers are connected to the LAN 12, and one or more clients are connected to the LAN 16.
[0030]
The Web-based ASP provides services by various application programs from the server 20 installed in the center 2 via the WAN 14, and the user accesses these services using a Web browser on a client installed in the office 4. To do. In such a form of use, the effective communication capacity (bandwidth) of the network connecting the intra-office LAN 16 and the intra-center LAN 12, particularly the WAN 14 such as the Internet, is lower than that of the LAN 12 or LAN 16. There is a problem that a communication delay occurs as a bottleneck and the response performance of the application deteriorates. Therefore, in this embodiment, for example, as shown in FIG. 1, two modules, a server-side proxy 30 and a client-side proxy 40, are installed at both ends of the WAN 14 that connects the intra-center LAN 12 and the intra-office LAN 16, and the server 20 and the client When 50 communicates via LAN 12, proxy 30, WAN 14, proxy 40, and LAN 16, the communication data amount is reduced by performing fingerprint compression (FP compression) and differential compression, which will be described later, between the proxies. Eliminate bottlenecks. Furthermore, in these proxies, as will be described in detail later, a necessary amount of cache is reduced by holding a part of data necessary for compression as difference information.
[0031]
The server 20, the server-side proxy 30, the client-side proxy 40, and the client 50 of this embodiment all have software (server program, server-side proxy program, client-side proxy program, client program) on the computer. It can be realized by operating it. In this case, software having a desired function, such as OS, driver software, packet communication software, encryption software, etc., or hardware such as a communication interface device, external storage device, input / output device, etc., is added to the computer as necessary. Mounted or connected. In this case, it is preferable to use a graphical user interface (GUI) for inputting information from a user or an administrator or presenting information to the user.
[0032]
On the client 50 used by the user in order to use the service, a program such as a Web browser operates according to the purpose. A user uses a service by, for example, issuing a request message to a server that provides a desired service such as information transfer or order reception from a Web browser, receiving a reply message, or repeating this appropriately. . Instead of general-purpose software such as a Web browser, other software such as dedicated software for using a specific service may be used. Further, the client may be not a general-purpose computer but a mobile phone terminal having an Internet function, for example.
[0033]
A predetermined server program operates on the server 20 and provides services specific to the server site to the user of the client 50.
[0034]
As shown in FIG. 1, the server-side proxy 30 may be installed and implemented so as to be connected to both the server center LAN 12 and the WAN 14 and operate as a transparent proxy. Further, as shown in FIG. 2, it may be installed on the LAN 12 in the server center. Moreover, you may implement so that the function of the server side proxy 30 may be incorporated in the server 20 like FIG.
[0035]
Similarly, as shown in FIG. 1, the client side proxy 40 may be installed and connected to both the intra-user office LAN 16 and the WAN 14 so as to operate as a transparent proxy. Further, as shown in FIG. 2, it may be installed on the LAN 16 in the user office. Further, as shown in FIG. 3, the function of the client side proxy 40 may be incorporated in a browser or the like that operates on the client 50. Alternatively, the client-side proxy 40 for personal use may be operated on the client 50 that operates such as a browser.
[0036]
Note that the server-side proxy 30 and the client-side proxy 40 may have the same form as shown in FIGS. 1 to 3 or different forms.
[0037]
In the following, an outline of the fingerprint cache, FP compression and differential compression using this, and data management will be described.
[0038]
In the following description, when one direction is taken as an example of the data transfer direction between the server-side proxy 30 and the client-side proxy 40, data is transferred from the server-side proxy 30 to the client-side proxy 40. As will be described by way of example, the same can be done when transferring data from the client side proxy 40 to the server side proxy 30.
[0039]
The server side proxy 30 and the client side proxy 40 of this embodiment have a cache mechanism called a fingerprint cache. The fingerprint cache records and manages data exchanged by the HTTP protocol by a name called a fingerprint (FP).
[0040]
As shown in FIG. 4, the fingerprint is a short numerical value determined by a predetermined calculation method (hash function in FIG. 4) from the contents of data exchanged by the HTTP protocol (content in FIG. 4). is there. This numerical value may be a variable length, but a fixed-length numerical value is easier to handle from the viewpoint of ease of processing.
[0041]
As a method for calculating the fingerprint, a well-known hash function such as MD-5 or SHA-1 can be used. These hash functions are used for digital signatures on data, and when given data is converted into a 128-bit number for MD-5 and a 160-bit number for SHA-1. can do. The feature of these hash functions is that if two data X1 and X2 are given and the data X1 and the data X2 are the same, the hash value calculated for the data X1 and the hash value calculated for the data X2 However, when two different data A and B are given, the hash value calculated for data A and the hash value calculated for data B are different with a very high probability. (In principle, the hash values calculated for two different data A and B may be the same, but the probability is small enough to be ignored in practice).
[0042]
As shown in FIG. 5, the fingerprint cache (260) of the server-side proxy 30 and the client-side proxy 40 is a data body (261) exchanged by the HTTP protocol in the past or data generated by differential compression described later. Using the fingerprint value (262) obtained by calculating (264) from the exchanged data as a name, differentially compressed data indicating whether the recorded data is a data body or data generated by differential compression It is recorded and managed together with an identifier (hereinafter referred to as a compressed identifier) (263). In the following, the identifier = 0 indicates that “the recorded data is a data body”, and the identifier = 1 indicates that “the recorded data is data generated by differential compression”. The description will be made with reference to an example, but the present invention is not limited to this.
[0043]
For example, when transferring data from the server-side proxy 30 to the client-side proxy 40 using the HTTP protocol, the server-side proxy 30 calculates a fingerprint of the data, and the data corresponding to the fingerprint is stored in the fingerprint cache. If it is included and the compression identifier is 0, the data existing in the cache is not data generated by differential compression, but the data (data having the same contents) has been transferred in the past. The corresponding fingerprint value is transferred without transferring. The client-side proxy 40 that has received the fingerprint can reproduce the data to be transferred by taking out the data corresponding to the fingerprint value from the fingerprint cache.
[0044]
Also, for example, in the above case, if the data corresponding to the calculated fingerprint is in the fingerprint cache and the compression identifier is 1, the server-side proxy 30 generates the data existing in the cache by differential compression. It can be seen that differential compression has succeeded in the past with respect to the received data (data having the same content as the received data) and that differential information has been transferred, and the corresponding finger is transferred without transferring the data. Transfer print values. The client-side proxy 40 that has received the fingerprint can reproduce the data to be transferred by extracting the data corresponding to the fingerprint value from the fingerprint cache and decompressing the differentially compressed data. .
[0045]
If it is the same data (data body or data generated by differential compression) sent in the past by such a method (that is, compression of data by fingerprint → transfer of fingerprint → decompression of data by fingerprint) Since only the fingerprint value needs to be sent, the amount of data flowing through the network can be greatly reduced.
[0046]
In the above case, the data corresponding to the calculated fingerprint is not stored in the fingerprint cache. -By registering in the cache, the fingerprint can be transferred by FP compression from the next time.
[0047]
For explanation, when transferring data between the server-side proxy 30 and the client-side proxy 40, the fingerprint cache is used to replace the message body data with the fingerprint to compress the transfer information amount. This is called print compression (FP compression).
[0048]
Subsequently, differential compression will be described.
[0049]
In the above FP compression, the identity of the data is determined at high speed using the fingerprint for the data, and the same data as the data registered in the fingerprint cache is not transferred between the proxies. Reduces network load by transferring prints. However, FP compression cannot be applied to data that is different from data registered in the fingerprint cache, even if most of the data has the same contents. Therefore, in this embodiment, even when FP compression is not possible, one or more data registered in the fingerprint cache is used as reference data to transfer the reference data fingerprint, difference information for the reference data, or the like. By expressing transfer data with information for restoring power data, the transfer data is expressed with a small amount of information, and the transfer data amount is reduced as much as possible. That is, the data in the fingerprint cache is used as a dictionary so that data that can be extracted from the dictionary is not sent.
[0050]
For the sake of explanation, when transferring data between the server-side proxy 30 and the client-side proxy 40, the amount of transfer information is compressed by replacing the message body data with the fingerprint of the reference data using a fingerprint cache. This is called differential compression. In addition, transfer data generated by this differential compression is referred to as differential compressed data.
[0051]
Note that differential compression using such a fingerprint does not require that the relationship between the data to be transferred and the reference data be determined in advance. No The advantage is also obtained. That is, in the conventional differential transfer, it is necessary to decide in advance what the difference is to be taken in both sides. Therefore, when the differential transfer is actually used in the Web system, the data of this URL is the difference based on this data. Therefore, it is impossible to effectively operate arbitrary data because it is necessary to register a rule such as On the other hand, this differential compression method takes the difference in the fingerprint cache data as reference data, so that the effect of compressing the data by the difference can be obtained even if the base of the difference is not determined in advance. Obtainable.
[0052]
Next, the data management method of this embodiment will be described.
[0053]
In this embodiment, when the differential compression is successful for the reply data received from the server 20 in the server side proxy 30, the server side proxy 30 transmits the differential compression data generated by the differential compression to the client side proxy 40. In the own fingerprint cache 260, the fingerprint acquired from the reply data and the differentially compressed data are recorded and managed. At this time, the compression identifier is set to 1 to indicate that the recorded / managed data is differentially compressed data.
[0054]
In the client side proxy 40, when the reply data received from the server side proxy 30 is differentially compressed data, the reply data received by the server side proxy 30 is generated from the differentially compressed data and transmitted to the client 50. In its own fingerprint cache 260, the fingerprint for the reply data generated from the differentially compressed data and the differentially compressed data received from the server-side proxy 30 are recorded and managed. At this time, the compression identifier is set to 1 to indicate that the recorded / managed data is differentially compressed data.
[0055]
On the other hand, when the reply data received from the server 20 in the server side proxy 30 is the first data and both FP compression and differential compression have failed, the server side proxy 30 sends the reply data to the client side proxy 40. Is recorded and managed in the fingerprint cache 260 of its own and the fingerprint acquired from the reply data and the reply data are recorded and managed. At this time, the compression identifier is set to 0 to indicate that the recorded / managed data is not differentially compressed data.
[0056]
When the client-side proxy 40 detects that the reply data received from the server-side proxy 30 is data that has not been subjected to either FP compression or differential compression, the reply data is transmitted to the client 50. In the own fingerprint cache 260, the fingerprint acquired from the reply data and the reply data are recorded and managed. At this time, the compression identifier is set to 0 to indicate that the recorded / managed data is not differentially compressed data.
[0057]
As a result, only the differentially compressed data is stored in the fingerprint cache 260 of both the server-side proxy 30 and the client-side proxy 40 for the data that has been successfully differentially compressed.
[0058]
By the way, in an application such as a Web-based ASP, data that is mostly the same content but different in only a part is often used. For example, there are many forms of data that have the same information written in many fields but differ only in part. Further, for example, there are Web pages that differ only in date or time, and those that differ only in the total access count counter value. In such a case, transfer and storage using differentially compressed data are particularly effective, and transfer and storage using the data body can be effectively reduced.
[0059]
Note that all the messages may be applied to the FP compression between the proxies 30 and 40. For example, a message that satisfies a predetermined condition is excluded from the FP compression application (always FP compression). May be transferred). The predetermined condition is, for example, that predetermined information (for example, information indicating a GET method and information indicating a request) is described in a message header, data to be transferred is null or very Various variations such as short size are conceivable.
[0060]
Further, all messages subject to application of FP compression may be subject to application of differential compression, but among messages subject to application of FP compression, messages that satisfy a predetermined condition are excluded from the subject of application of differential compression. (In this case, the condition to which the differential compression is applied is a condition in which another condition is weighted to the condition to which the FP compression is applied). For example, the upper limit value U2 of the data size not subjected to differential compression is set larger than the upper limit value U1 of the data size not subjected to FP compression (the lower limit L2 of the data size to be differentially compressed is set lower than the lower limit value L1 of the data size to be FP compressed). It is determined by the data size whether it is an application target of FP compression, or whether it is an application target of differential compression, but data other than HTML or XML is an application target of the application target of FP compression. Various variations are conceivable, such as a method in which data is excluded (differential compression is performed only on HTML or XML data). In the following description, a case where all the messages subject to application of FP compression are subject to application of differential compression will be described as an example.
[0061]
In this embodiment, the message subject to FP compression and differential compression is a message in which data is FP-compressed (however, the compression identifier may be 0 or 1), and a message in which data is differentially compressed. Or a message whose data is not compressed is transferred between the server-side proxy 30 and the client-side proxy 40.
[0062]
When there is a message that is a target for FP compression and not a target for differential compression, the message is either a message in which data is FP compressed (compression identifier = 0) or a message in which data is not compressed. Transfer is performed between the server-side proxy 30 and the client-side proxy 40.
[0063]
When there is a message that is not subject to FP compression, the message is transferred between the server side proxy 30 and the client side proxy 40 as a message whose data is not compressed.
[0064]
Here, a differential compression method will be described.
[0065]
There are various methods of differential compression using a fingerprint cache such as the following.
One of the data registered in the fingerprint cache is used as reference data. Between the proxies, a fingerprint value corresponding to the reference data and information indicating a difference between the transfer data and the reference data are transferred.
In the above method, the data registered in the fingerprint cache is used for all or part of the difference from the reference data. Alternatively, all or part of the transfer data is expressed by combining data (entire or part) registered in the fingerprint cache. For example, any number of data registered in the fingerprint cache is used as reference data. Between the proxies, the fingerprint value corresponding to the reference data, the information indicating the portion of the reference data used for restoring the transfer data, and the information indicating the method for restoring the transfer data based on that portion are transferred. To do.
[0066]
Below, an example of the expression method of differential compression data is shown.
[0067]
FIG. 6 shows three types of instructions for expressing data to be transferred. The differentially compressed data transferred instead of the data is composed of a sequence of these instructions.
[0068]
(A) is an instruction for defining the reference data with a number when referring to the data in the fingerprint cache. The first byte 8n (n = 0 in the example of FIG. 6) is an instruction identifier. Among the instruction identifiers, n indicates that data designated by the fingerprint is handled as nth reference data. 16 bytes starting from the second byte indicate a fingerprint value for the reference data. In this example, 80 to 8F can handle reference data from 0th reference data to 15th maximum. The maximum number of reference data that can be handled can be more or less depending on the implementation.
[0069]
(B) represents an instruction to copy partial data from the reference data defined in (a). 9n in the first byte (n = 0 in the example of FIG. 6) is an instruction identifier. Among the instruction identifiers, n indicates that the nth reference data defined by the instruction (a) is used. The 4 bytes from the 2nd byte indicate the offset position in the nth reference data, and the 4 bytes from the 6th byte indicate the data length. These indicate that data of a specified length from the specified offset position in the nth reference data should be copied (as a constituent part of the transfer data) in accordance with the position (order) of this instruction sequence.
[0070]
(C) is an instruction for directly specifying data. A0 in the first byte is an instruction identifier. Four bytes from the second byte indicate the length of the data. The number of bytes specified by the length follows the 6th byte. Then, they indicate that the data after the 6th byte is copied (as a constituent part of the transfer data) in accordance with the position (order) of this instruction sequence.
[0071]
The differentially compressed data composed of such a sequence of instructions can be decompressed by simply copying and connecting the data in the order in which they are instructed while referring to the fingerprint cache.
[0072]
Next, an example of differential compression according to the above method is shown.
[0073]
Assume that data as shown in FIG. 7 is stored in the fingerprint cache as fingerprints 5E83... B6.
[0074]
At this time, if data as shown in FIG. 8 is given, differential compression can be performed as shown in FIG. 9 using the data in FIG. 7 as reference data.
[0075]
That is, the data in FIG. 8 is merely changed from “TOYO” in FIG. 7 to “OSAKA” as compared with the data in FIG. Therefore, as shown in FIG. 9, first, it is instructed to define the data of fingerprint “5E83... B6” as the 0th reference data at the 0th byte, and from the 0th byte of the 0th reference data to the 17th byte. Instruct to copy bytes, then instruct to copy 5 characters of “OSAKA” at 26th byte, and finally copy 51 bytes from 65th byte of 0th reference data at 36th byte Is instructing.
[0076]
If decompression is performed according to this instruction, the data in FIG. 8 can be reproduced.
[0077]
In this example, only one reference data is used, but a plurality of reference data can be used.
[0078]
Next, another example of differential compression according to the above method will be shown.
[0079]
Assume that data as shown in FIG. 10 is in the fingerprint cache as fingerprints 82F3... 38 and data as shown in FIG. 11 is in the fingerprint cache as fingerprints A20D. At this time, if data as shown in FIG. 12 is given, differential compression can be performed as shown in FIG. 13 using the data in FIG. 10 and the data in FIG. 11 as reference data.
[0080]
In FIG. 13, first, it is instructed to define the data of the fingerprint “82F3... 38” as 0th reference data at the 0th byte, and 53 bytes from the 0th byte of the 0th reference data are copied at the 17th byte. And instructing to define the fingerprint “A20D... CB” data as the first reference data at the 26th byte, and copying the 55th byte from the 96th byte of the first reference data at the 43rd byte. Is instructing.
[0081]
If decompression is performed according to this instruction, the data in FIG. 12 can be reproduced.
[0082]
In the above-described method, the method of specifying the portion to be used of the reference data and directly specifying the data to be used (method 1) is used. Instead, the portion of the reference data that is not used (directly). A method (method 2) in which an instruction of a part to be replaced with instruction data and direct designation of data to be inserted into the unused part is also possible.
[0083]
It is also possible to use method 1 and method 2 in combination.
[0084]
Next, an inter-proxy message format (for a message to which FP compression is applied) when data is transferred between the server-side proxy 30 and the client-side proxy 40 will be described with reference to FIGS. .
[0085]
When data is transferred between the server-side proxy 30 and the client-side proxy 40, the message to which FP compression is applied includes a message in which data is FP-compressed and replaced with a fingerprint (message at the time of FP compression), FP compressed Rena However, there are a message in which differentially compressed data is mounted (message at the time of differential compression) and a message in which data that is not subjected to FP compression or differential compression is mounted (message at the time of non-compression). In the case of a configuration in which not all messages are subject to FP compression, there are messages that are not subject to FP compression in addition to these three messages.
[0086]
In the proxy on the transmission side of both proxies, the data is deleted and the fingerprint is added in the message at the time of FP compression, the data is deleted in the message at the time of differential compression, and the data such as the fingerprint of the reference data is restored. Information is added. Data is not deleted for uncompressed messages and messages not subject to compression.
[0087]
Of the two proxies, the receiving proxy needs to be able to identify the above three or four types of messages. When a message is received at the time of FP compression, the fingerprint is returned to the data (however, if the compression identifier is 0, the data body is retained, but if the compression identifier is 1, the difference compression is used) Since the generated data is retained, the data is restored based on this data), and the data is restored when a message is received during differential compression. When a message is received at the time of differential compression or non-compression, the fingerprint cache is registered.
[0088]
FIG. 14 shows an example of the message format. (A) is a message at the time of non-compression, (b) is a message at the time of differential compression, and (c) is a message at the time of FP compression.
[0089]
In (a), data is placed in the message body, in (b) information for restoring the data is placed in the message body instead of data, and in (c), a finger is substituted in the message body instead of data. A print (FP) is placed.
[0090]
Also, in this example, identification information for identifying the type of message is described (in the compression proxy) in the message header, and whether or not FP compression is performed based on this identification information (in the decompression proxy). Identify (for example, 01 means no compression, 10 means differential compression, 11 means FP compression). The identification information may be special information used between proxies, or may be information that uses or uses a field originally present in a normal HTTP message header.
[0091]
Note that, in a configuration in which a message that is not subject to FP compression may exist, a message of a message that is not subject to FP compression, as shown in FIG. 14D, in the proxy on the compression side (transmission side). -The identification information (for example, 00) may be included in the header.
[0092]
Further, in a configuration in which a message that is not subject to FP compression may exist, the message on the decompression side (reception side) proxy is a message that is not subject to FP compression due to some information included in the message header. When it is possible to determine whether or not the message body is null and the message is determined not to be subject to FP compression, the FP compression is not applicable. It is also possible not to include identification information in the message header of the message.
[0093]
In the example of FIGS. 14A and 14B, the fingerprint for the data is not included in the message at the time of non-compression or differential compression. However, the fingerprint may be included in the message header. Good (or a fingerprint for the data may be included in the message body). FIGS. 15A to 15C show an example in which a fingerprint is included in the message header at the time of non-compression or differential compression, and a fingerprint is included in the message body at the time of FP compression. In this way, when the fingerprint cache is registered for the data on the decompression side, it is possible to save time and effort for obtaining the fingerprint from the data again by using the fingerprint.
[0094]
In any of the above cases, as shown in FIG. 16, in the message at the time of FP compression, a fingerprint (FP) may be included in the message header, and the message body may be null.
[0095]
In addition to the above, various message formats are possible.
[0096]
For example, in FIG. 17, the uncompressed flag is described in the header of the message at the time of (a) uncompressed and the fingerprint is not described, and the differential compression flag is described in the header of the message at the time of differential compression of (b). Also, the fingerprint is not described, the fingerprint is described in the header of the message at the time of FP compression in (c), and the body is null.
[0097]
In this case, it is possible to identify that the message is a message at the time of FP compression by detecting that the message body is null or the fingerprint is described in the header of the message. it can. Further, a message at the time of non-compression and a message at the time of differential compression can be identified by detecting that a non-compression flag or a differential compression flag is described in the header.
[0098]
In the case of a configuration in which a message that is not subject to FP compression may exist, in the proxy on the compression side (transmission side), as shown in (a) and (b) of FIG. If (b) is not null (if the data is null), a non-target flag may be included in the message header of a message that is not subject to FP compression.
[0099]
Further, in a configuration in which a message that is not subject to FP compression may exist, the message on the decompression side (reception side) proxy is a message that is not subject to FP compression due to some information included in the message header. 18 (c) and 18 (d), (c) indicates that the data is not empty (null), and (d) indicates that the data is empty (null). ) It is also possible not to include a non-target flag in the message header of a message that is not subject to FP compression.
[0100]
In the following, the present embodiment will be described in detail focusing on the case where the reply data is first FP compressed / decompressed when the reply message is transferred from the server side proxy 30 to the client side proxy 40.
[0101]
FIG. 19 shows a configuration example of the server-side proxy 30 in this embodiment, and FIG. Ru The structural example of the client side proxy 40 is shown. 19 and 20 mainly show the configuration when data is transferred from the server-side proxy 30 to the client-side proxy 40. FIG.
[0102]
The server-side proxy 30 in FIG. 19 is a receiving unit 31 that performs processing for receiving a transfer message from the intra-server center LAN 12 or the wide area network 14, and FP compression and differential compression for the data included in the transfer message. The processing unit 32, the transmission unit 33 that performs processing for transmitting a transfer message to the LAN 12 in the server center or the wide area network 14, and the fingerprint and the original data or differentially compressed data are stored in association with each other. A fingerprint cache (FP cache) 34 is provided. In addition, the processing unit 32 performs a search or registration for the fingerprint compression determination unit (FP compression determination unit) 321 and the FP cache 34 for determining whether or not the data included in the transfer message is to be compressed. Fingerprint cache management unit (FP cache management unit) 322, fingerprint compression processing unit (FP compression processing unit) 323 for performing processing such as replacing the data included in the transfer message with the corresponding fingerprint, and transfer message Includes a differential compression processing unit 324 for performing processing such as replacing the data included in the data with differential compression data.
[0103]
20 includes a receiving unit 41 that performs processing for receiving a transfer message from the user office LAN 16 or the wide area network 14, and a processing unit 42 that performs FP decompression on data included in the transfer message. A transmission unit 43 that performs processing for transmitting a transfer message to the user office LAN 16 or the wide area network 14, and a fingerprint for storing the fingerprint and the original data or differentially compressed data in association with each other A cache (FP cache) 44 is provided. The processing unit 42 also includes a fingerprint compression determination unit (FP compression determination unit) 421 for determining whether or not the data included in the transfer message is to be compressed, and whether or not the transfer message is subjected to FP compression and differential compression. In order to perform processing such as a fingerprint cache management unit (FP cache management unit) 422 for performing search and registration for the FP cache 44, and decompressing original data from the fingerprint included in the FP compressed transfer message. A fingerprint decompression processing unit (FP decompression processing unit) 423, and a differential decompression processing unit 424 for performing processing such as decompressing original data from differentially compressed data included in a differentially compressed transfer message.
[0104]
Note that the compression-side FP compression determination unit 321 and the decompression-side FP compression determination unit 421 check whether the message satisfies a predetermined condition as described above, thereby converting the data included in the message into the FP. It is determined whether or not compression is to be applied (when all messages are to be applied to FP compression, the compression side FP compression determination unit 321 and the corresponding part of the procedure example shown later and the decompression side FP compression determination are determined. The corresponding determination part of the part 421 and the corresponding part of the procedure example shown later are unnecessary). In addition, the decompression-side FP compression determination unit 421 determines whether or not the data subjected to FP compression application is FP compressed. In the following, a description will be mainly given of a case where a message to be applied to FP compression is transferred (when it is determined that FP compression is to be applied or all messages are to be applied to FP compression).
[0105]
FIG. 21 and FIG. 22 show an example of the processing procedure of the server side proxy 30 when the reply message is transferred from the server side proxy 30 to the client side proxy 40. 21 and 22 describe the processing when one reply message is received, but the processing illustrated in FIGS. 21 and 22 is actually performed for all reply messages received by the server-side proxy 30. I do.
[0106]
The server side proxy 30 receives a reply message from the server 20 by the receiving unit 31 (step S1).
[0107]
The FP compression determination unit 321 determines whether or not the reply data of the reply message is for FP compression (S2). If it is determined that the reply data is not subject to FP compression (S2), the received reply message is transferred from the transmission unit 33 to the client side proxy 40 (S13).
[0108]
If it is determined in step S2 that the reply data of the reply message is for FP compression, the FP cache management unit 322 calculates a fingerprint value of the reply data (S3), The FP cache 34 is searched using the print value as a key (S4).
[0109]
If the set of the fingerprint value and the corresponding data (data body or differentially compressed data) is registered in the FP cache 34 (S5), the FP compression processing unit 323 receives the received reply. The message is sent to the client side proxy 40 from the transmission unit 33 in the FP compression format using the fingerprint value (S6).
[0110]
On the other hand, as a result of the search in step S4, if the set of the fingerprint value and the corresponding data is not registered in the FP cache 34 (S5), the differential compression processing unit 324 performs differential compression. (S7), it is determined whether or not the differential compression is successful (S8).
[0111]
Whether or not the differential compression has succeeded is determined if, for example, the data amount r0 of the original data targeted for differential compression is compared with the data amount r1 of the differential compression data, and if r0−r1> d is satisfied, Judge that Here, the constant d is a predetermined integer of 0 or more.
[0112]
If it is determined in step S8 that the differential compression is successful, the following two operations are performed (either 1-1 or 1-2 may be performed first or in parallel).
(1-1) In the differential compression processing unit 324, the received reply message is converted into a format at the time of differential compression (using the fingerprint value as necessary), and the client side proxy 40 is sent from the transmission unit 33. (S9).
(1-2) The FP cache management unit 322 associates the fingerprint value with the differentially compressed data generated in step S7 (using the fingerprint value as a key) and registers it in the FP cache 34. (S10). At this time, since the data to be registered is differentially compressed data, the compression identifier is set to 1.
[0113]
In the above (1-2), all the differentially compressed data is registered in the FP cache. However, (i) the size of the differentially compressed data is almost the same as the size of the data that is the basis of the differentially compressed data. If not, store the original data of the differential compressed data instead of the differential compressed data. (Ii) When the size of the differential compressed data is very small compared to the size of the data of the differential compressed data It is also possible to maintain the registration in the FP cache of the data used as reference data when creating the differential compressed data without saving the differential compressed data and without saving the original data of the differential compressed data. . There are various other variations. This also applies to the case where differential compressed data is registered in the FP cache in other processing procedure examples described later (in the following description, all differential compressed data is registered in the FP cache as well). The case is assumed to be an example).
[0114]
On the other hand, if it is determined in step S8 that the differential compression has failed, the following two operations are performed (either 2-1 or 2-2 may be performed first or in parallel). .
(2-1) The differential compression processing unit 324 (or the FP compression processing unit 323) converts the received reply message into the uncompressed format from the transmission unit 33 to the client side proxy 40 (S11). .
(2-2) The FP cache management unit 322 associates the fingerprint value with the reply message (using the fingerprint value as a key) and registers it in the FP cache 34 (S12). At this time, since the data to be registered is not differential compression data, the compression identifier is set to 0.
[0115]
Next, FIG. 23 to FIG. 25 show an example of the processing procedure of the client side proxy 40 when the reply message is transferred from the server side proxy 30 to the client side proxy 40. 23 to 25 describe the processing when one request message is received, but the processing illustrated in FIGS. 23 to 25 is actually performed for all the request messages received by the client-side proxy 40. I do.
[0116]
The client side proxy 40 receives a reply message from the server side proxy 30 by the receiving unit 41 (step S21).
[0117]
The FP compression determination unit 421 checks and determines whether the reply data of the reply message is for FP compression (S22). If it is determined that the reply data is not subject to FP compression (S22), the received reply message is transferred from the transmitter 43 to the client 50 (S38).
[0118]
If it is determined in step S22 that the reply data of the reply message is for FP compression, the FP compression determination unit 421 further checks whether or not the reply data is FP compressed (S23). ).
[0119]
If it is determined in step S23 that the reply data of the reply message is FP compressed, the FP cache management unit 422 obtains the fingerprint value of the reply data (S24), The FP cache 44 is searched using the value as a key (S25).
[0120]
Next, whether or not the data obtained as a result of the search in step S25 is differentially compressed data is examined and determined by the FP cache management unit 422 using the compression identifier of the data (S26).
[0121]
If it is determined in step S26 that the data is differentially compressed data (in this example, the compression identifier is 1), the differential decompression processing unit 424 searches for and acquires reference data in the FP cache management unit 422. After that, the differential compressed data is decompressed to restore the original reply data (S27), the restored reply data is added to the reply message, and when using special information between proxies, the received reply message is used. The information is deleted, and the reply message is transmitted from the transmission unit 43 to the client 50 (S28).
[0122]
If it is determined in step S26 that the data is not differentially compressed data (in this example, the compression identifier is 0), the FP decompression processing unit 423 retrieves the received reply message from the FP cache 44. In addition, when data corresponding to the fingerprint value is added and special information is used between proxies, the information is deleted and then transmitted from the transmission unit 43 to the client 50 (S29).
[0123]
On the other hand, if it is determined in step S23 that the reply data of the reply message is not FP compressed, the FP compression determination unit 421 further determines whether or not the reply data is differentially compressed (S30). ).
[0124]
If it is determined in step S30 that differential compression has been performed, the following two operations are performed (either 1-1 or 1-2 may be performed first or in parallel).
(1-1) In the differential decompression processing unit 424, the differential reply data is decompressed (after retrieving and acquiring the reference data by the FP cache management unit 422) to restore the original reply data (S31). When reply data is added to the reply message and special information is used between proxies, the information is deleted from the received reply message, and the reply message is transmitted from the transmission unit 43 to the client 50 (S32).
(1-2) The FP cache management unit 422 obtains the fingerprint value of the reply data (S33) and associates the fingerprint value with the differentially compressed data (using the fingerprint value as a key). And is registered in the FP cache 44 (S34). At this time, in registration in the FP cache 44, since the reply message is differentially compressed data, the compression identifier is set to 1.
[0125]
On the other hand, if it is determined in step S30 that differential compression has not been performed, the following two operations are performed (either 2-1 and 2-2 may be performed first or in parallel). .
(2-1) When the difference decompression processing unit 424 (or the FP decompression processing unit 423) uses special information between proxies, the information is deleted from the received reply message, and then is transmitted from the transmission unit 43. It transmits to the client 50 (S35).
(2-2) The FP cache management unit 422 obtains a fingerprint value of the reply data (S36), associates the fingerprint value with the reply message (using the fingerprint value as a key). And register in the FP cache 44 (S37). At this time, in registration in the FP cache 44, the reply message is not differentially compressed data, so the compression identifier is set to zero.
[0126]
In steps S24 / S33 / S36, when a fingerprint is described in the message, a method for obtaining the fingerprint from the message, and when no fingerprint is described in the message, the reply data is used. There is a method of calculating a fingerprint value by a hash function or the like. Even when the fingerprint is described in the message, a method of calculating the fingerprint value based on the reply data is also possible. Steps S24 / S33 / S36 may be performed at other timing before using the fingerprint. Moreover, you may perform all or one part of judgment of step S22, S23, S26, S30 simultaneously.
[0127]
Next, a procedure for performing differential compression will be described.
[0128]
FIG. 26 shows an example of the differential compression procedure. This is a procedure example in the case of using the three types of designation described above.
[0129]
Here, in order to perform differential compression, it is assumed that a history table in which recently accessed data is sequentially arranged among the data stored in the fingerprint cache is used. This history table does not necessarily need to record the fingerprints of all data stored in the fingerprint cache. For example, a predetermined number may be recorded (in this case, the number is determined in advance assuming a number effective for use as reference data for differential compression, for example).
Of course, one or a plurality of criteria other than the order of recent access may be used as the criteria for recording in the history table, and one or more other criteria may be used in addition to the criteria of the order of recent access. May be.
[0130]
There is also a method in which the history table is configured integrally with the fingerprint cache.
[0131]
(Step S231)
First, the working copy buffer and instruction buffer are emptied.
[0132]
Based on the contents copied to the copy buffer, the instructions in FIGS. 6A to 6C are created and written to the instruction buffer. Finally, the sequence of instructions written in the instruction buffer becomes differentially compressed data.
[0133]
(Step S232)
Data to be subjected to differential compression is handled as a character string, and a pointer pointing to the character string is prepared. First, the pointer is set to point to the first character in the character string.
[0134]
In the following, loop processing is executed until it is determined in step S241 that the pointer has reached the last character in the character string.
[0135]
(Step S233)
The data corresponding to the fingerprints recorded in the history table are taken out in order from the newest one, and the data that matches more than a predetermined length from the location indicated by the pointer in the data to be differentially compressed is the reference data Take out as.
[0136]
There are various reference data and determination methods. For example, a method of examining data corresponding to fingerprints recorded in the history table in the latest order, and first extracting data that matches a predetermined length or more as reference data, fingerprints recorded in the history table There is a method in which all of the data corresponding to the above are examined, and the data having the longest matching length (provided that the matching is longer than a predetermined length) is taken out as reference data.
[0137]
(Step S234)
If reference data is found, the process proceeds to step S237. If the reference data is not found, the process proceeds to step S235.
[0138]
(Step S237)
When the reference data is found in step S234, if the copy buffer is not empty, a copy instruction by direct designation in FIG. 6C corresponding to the character string in the copy buffer is created and designated. Write to buffer. The copy buffer is emptied.
If the copy buffer is empty, nothing is done.
[0139]
(Step S238)
If the reference data definition of FIG. 6A corresponding to the reference data found in step S233 has not yet been written to the instruction buffer, the reference data definition instruction is written to the instruction buffer.
[0140]
(Step S239)
A copy instruction for the matched character string from the reference data is written in the instruction buffer as a copy instruction in FIG.
[0141]
(Step S240)
Advance the pointer by the length of the character string that matches the reference data.
[0142]
(Step S235)
On the other hand, when the reference data is found in step S234, the character pointed to by the pointer is entered in the copy buffer.
[0143]
(Step S236)
Advance the pointer by one character.
[0144]
(Step S241)
If the pointer has not reached the end of the data to be subjected to differential compression (if there is a data portion that has not been processed), the process returns to step S233. If the pointer has come to the end of the data to be subjected to differential compression (if no unprocessed data portion remains), the processing loop is exited and the process proceeds to step S242.
[0145]
(Step S242)
If the copy buffer is not empty, a copy instruction by direct designation in FIG. 6C corresponding to the character string in the copy buffer is created and written to the instruction buffer. The copy buffer is emptied.
If the copy buffer is empty, nothing is done.
[0146]
The contents of the instruction buffer at this time become differentially compressed data.
[0147]
In addition, as described above, after the differential compression is actually performed in this way, the data size as a result of the differential compression is smaller than the data size before the execution (or a constant standard (constant Confirm that it is smaller than the compression amount). If the data size is not reduced by differential compression (or if it does not decrease beyond a certain standard (a certain amount of compression)), it is better not to perform differential compression, so the data is transferred as it is.
[0148]
By the way, in the differential compression process described above, the process of selecting the latest data having a character string that matches the predetermined length or more in step S233 is considered to be the process that takes the longest time. To speed up this process, a hash table can be used. Extract all the character strings of a predetermined length from the data to be entered in the history table, calculate the hash value (for example, the sum of all the character codes), and register it in the hash table. Keep it. This hash table is overwritten with the latest data if there is data having the same hash value.
Using this hash table, the hash value of a character string of a predetermined length is obtained from the current pointer position of the data to be differentially compressed, and the data obtained by subtracting the hash table using the hash value is the step It becomes the first candidate of data to be selected in S233. Here, since the same hash value may be generated even in different character strings, whether or not they are really the same is confirmed by actually comparing the character strings. Search for candidates.
[0149]
Another method for speeding up the processing in step S233 is a method of performing comparison processing in units of rows. In the procedure of FIG. 26, the comparison processing is performed in units of characters. However, for all data in the history table, a hash value column for each row in the data is calculated. First, a column of hash values of each row is calculated from data to be differentially compressed. Thereafter, the procedure is the same as that in FIG. 26, but the comparison is performed not in units of characters but in units of row hash values. Since this method uses a line as a unit, the number of comparisons can be reduced as compared with a character unit. However, in order to compare by hash value, there may be the same hash value in different rows, so after finally judging that the hash values are the same, whether or not they are really the same by actually comparing the rows It is desirable to judge.
Of course, the above-described hash technique can be combined for such a configuration in which the comparison is performed in units of rows. In this case, a plurality of consecutive rows may be combined and registered in the hash table in units of the minimum number of rows that is the same as or longer than a predetermined length.
[0150]
Next, FIG. 27 shows another example of the differential compression procedure. This uses one of the data registered in the fingerprint cache as reference data, and between the proxies, the fingerprint value corresponding to the reference data, and information indicating the difference between the transfer data and the reference data It is an example of a procedure when transferring.
[0151]
Again, the history table as described above is used.
[0152]
First, of the data corresponding to the fingerprints recorded in the history table, one satisfying a predetermined standard is selected as reference data (step S245).
[0153]
The predetermined standard is, for example, that the data amount of a portion that does not match the reference data among the data to be subjected to differential compression is equal to or less than a predetermined data amount, and the reference When the data is divided into a plurality of chunks by the unmatched portion, if the number of the divided chunks is equal to or less than a predetermined number, this can be selected as reference data.
[0154]
As a method for determining the reference data, as described above, for example, the data corresponding to the fingerprints recorded in the history table are checked in the latest order, and the data that matches the predetermined length or more at first As a reference data, to examine all of the data corresponding to the fingerprint recorded in the history table, to extract the data that did not match and the data with the least number of divisions as reference data, etc. There are various methods.
[0155]
If the reference data is found (S246), the reference data definition instruction (for example, (a) in FIG. 6) is written in the instruction buffer (S247).
[0156]
An instruction indicating a part of the reference data that has not been matched (a part to be replaced with the directly specified data) (for example, an instruction in the same format as in FIG. Writing to the buffer (S248).
[0157]
A direct designation instruction (for example, (c) in FIG. 6) indicating data to be inserted into the unmatched portion of the reference data is written to the instruction buffer (S249).
[0158]
If there are a plurality of replacement parts, step S248 and step S249 are executed accordingly.
[0159]
The contents of the instruction buffer at this time become differentially compressed data.
[0160]
On the other hand, if reference data is found (S246), differential compression is not performed.
[0161]
In the above description of FIG. 26 and FIG. 27, the history table is provided separately from the fingerprint cache. However, there is a method in which the history table is integrated with the fingerprint cache.
[0162]
In the above description, the history table is used. However, there is a method that does not use the history table. For example, the fingerprint cache is checked in a predetermined order (for example, in order of entry or at random) by a predetermined upper limit number, and the best one of them is used. There are various methods such as a method of checking reference data in a predetermined order (for example, in order of entry or randomly) and determining reference data when reference data satisfying a predetermined condition is first obtained.
[0163]
In addition to the fingerprint, the history table or fingerprint cache also stores the URL of the reply message that included the reply data that was the original when the fingerprint was registered, and stores the reference data. When searching, first, it is checked whether or not data having the same URL as the reply message including the reply data to be subjected to differential compression is registered in the history table or the fingerprint cache. Data having the same URL as the URL may be prioritized over other data to check whether it can be used as reference data.
[0164]
Of course, in addition to FIG. 26 and FIG. 27, various differential compression procedures are possible.
[0165]
In the case where the fingerprint cache is not used when the request message is transferred from the client side proxy 40 to the server side proxy 30, the server side proxy 30, as illustrated in FIG. A procedure of receiving a request message from the proxy 40 (step S60) and transmitting it to the server 20 (step S61) may be used. Similarly, as illustrated in FIG. 29, the client-side proxy 40 may receive a request message from the client 50 (step S62) and transmit it to the server-side proxy 30 (step S63).
[0166]
Hereinafter, the data transfer using the fingerprint cache will be described more specifically with reference to FIGS. 30 to 32.
[0167]
First, referring to FIG. 30, data that has not been registered in the fingerprint cache and that has not been successfully compressed is transferred from the server-side proxy 30 to the client-side proxy 40, and The operation when registering a print cache will be described.
[0168]
(1) It is assumed that the browser or the like on the client 50 issues a POST method request message to the server 20 with a URL “/A.cgi”, for example. First, a browser or the like is set so that a request message to the server 20 is sent to the client-side proxy 40.
[0169]
(2) The client-side proxy 40 that has received the request message from the client 50 transfers the request message to the server-side proxy 30.
[0170]
(3) The server-side proxy 30 that has received the request message transfers the request message to the server 20.
[0171]
(4) After processing the request message, the server 20 sends the reply message back to the server-side proxy 30.
[0172]
(5) Upon receiving the reply message, the server-side proxy 30 first calculates the fingerprint of the reply data included in the received reply message, and checks whether data having the fingerprint name is in the FP cache 34. Here, for the first time not entered, the data that was registered in the fingerprint cache is deleted or invalidated in the case of the configuration that may be deleted or invalidated afterwards. Therefore, the data is entered (registered) in the FP cache 34 with the fingerprint as a name. Here, it is assumed that the data amount cannot be reduced by differentially compressing the reply data. At this time, in the registration in the FP cache 34, since the data is not differentially compressed data, the compression identifier is set to 0.
[0173]
(6) The server side proxy 30 transfers the reply message carrying the data to the client side proxy 40. If the fingerprint value calculated from the reply data is sent in a reply header or the like, the trouble of calculating the fingerprint again by the client side proxy 40 can be saved.
[0174]
(7) Since the client side proxy 40 that has received the reply message is the first data, it registers the reply data in the FP cache 44. At this time, in the registration in the FP cache 44, since the reply data is not differentially compressed data, the compression identifier is set to 0. As described above, the fingerprint is calculated from the reply data, or the fingerprint put in the reply header or the like by the server side proxy is taken out and this is entered as a name.
[0175]
(8) The client-side proxy 40 deletes the information used only between the server-side proxy 30 and the client-side proxy 40 (such as a fingerprint value in the reply header). After that, a reply message is sent back to the client 50 (such as a browser operating on the client 50).
[0176]
The server-side proxy 30 may perform the fingerprint cache registration (5) described above after the operation (6). Further, in the client side proxy 40, the fingerprint cache registration of (7) may be performed after the operation of (8).
[0177]
Next, referring to FIG. 31, data that has not been registered in the fingerprint cache and has been subjected to differential compression is transferred from the server-side proxy 30 to the client-side proxy 40, and the fingerprint of the data is fingerprinted. -The operation when registering a cache will be described.
[0178]
(1) to (4) are the same as (1) to (4) in the operation described with reference to FIG.
[0179]
(5) Upon receiving the reply message, the server-side proxy 30 first calculates the fingerprint of the reply data included in the received reply message, and checks whether data having the fingerprint name is in the FP cache 34. Here, it is determined that the data is not included and is the first data. If the data to be compressed cannot be FP compressed, differential compression processing is performed. If the reply data can be differentially compressed to reduce the data amount, differential compression is performed. Here, it is assumed that when data having a fingerprint of “71F0... 73E6” is used as reference data, the reply data is differentially compressed to reduce the data amount. Reply data is replaced with differential compression data, and information indicating that differential compression has been performed is put in a reply header or the like.
[0180]
(6) The server-side proxy 30 puts (registers) the differentially compressed data generated in (5) into the FP cache 34 with the fingerprint calculated in (5) as a name. At this time, in the registration in the FP cache 34, since the data is differentially compressed data, the compression identifier is set to 1.
[0181]
(7) The server side proxy 30 transfers the reply message to the client side proxy 40.
[0182]
(8) The client-side proxy 40 sees the reply header and the like, knows that the reply data is differentially compressed, and decompresses the compression. At this time, first, the fingerprint “71F0... 73E6” of the reference data specified in the compressed data is extracted, and then the data corresponding to the fingerprint is extracted from the FP cache 44 and used. Then, the decompressed data is put in the reply data, and necessary headers such as the reply header content size are rewritten.
[0183]
(9) Since the client-side proxy 40 that has received the reply message is the first data, it registers the differentially compressed data received from the server-side proxy 30 in the FP cache 44. At this time, in the registration in the FP cache 44, since the data to be registered is differential compression data, the compression identifier is set to 1. In addition, in the case of a configuration in which information used only between the server-side proxy 30 and the client-side proxy 40 such as a fingerprint value exists in the reply header or the like, this is deleted. As described above, the fingerprint is calculated from the reply data (data obtained by decompressing the differentially compressed data), or the fingerprint put in the reply header or the like by the server side proxy is taken out, and this is used as the name. Put in.
[0184]
(10) The client-side proxy 40 sends a reply message back to the client 50 (such as a browser operating on the client 50).
[0185]
In (5) above, all unregistered data to be compressed is registered in the fingerprint cache, and all the data used as reference data in (6) above is the fingerprint cache. However, when the only data is used for the reference data and the data subject to differential compression is almost the same as the reference data used for the differential compression ( For example, if the amount of data that does not match is below the standard), (i) unregistered data to be compressed is not registered in the fingerprint cache, and the reference data maintains the fingerprint cache registration Or (ii) Register unregistered data to be compressed in the fingerprint cache and reference data as fingerprint To remove from the cache, it is also possible to methods such as. In addition, when the only data is used for the reference data, the data subject to differential compression and the reference data used for the differential compression have almost the same contents and the same URL. In this case, it is possible to register the unregistered data to be compressed in the fingerprint cache and delete the reference data from the fingerprint cache, assuming that the data has been updated. There are various other variations.
[0186]
Next, with reference to FIG. 32, an operation in the case where the data registered in the cache by performing the operation of FIG. 30 or 31 is transferred from the server side proxy 30 to the client side proxy 40 will be described.
[0187]
(1) to (4) are the same as (1) to (4) in the operation described with reference to FIG.
[0188]
(5) Upon receiving the reply message from the server 50, the server-side proxy 40 first calculates the fingerprint of the reply data included in the received reply message, and whether or not the data having the fingerprint name is in the FP cache 34 Check out. Since the fingerprint cache is registered here, the reply body data is replaced with the fingerprint (for example, by inserting the fingerprint value into the reply header or the like and emptying the reply body).
[0189]
(6) The server side proxy 30 transfers the reply message in which the reply body is replaced with the fingerprint to the client side proxy 40.
[0190]
(7) The client-side proxy 40 that has received the reply message detects that the reply data has been replaced with the fingerprint, and FP is performed using the specified fingerprint (for example, in the reply header as described above). The corresponding data is retrieved from the cache 44.
Here, when the compression identifier of the data is 0, since the data is not differentially compressed data, it is put in the reply body. On the other hand, when the compression identifier of the data is 1, since the data is differentially compressed data, the reply data that has been decompressed is restored and placed in the reply body. FIG. 32 illustrates the former case.
In addition, in the case of a configuration in which information used only between the server-side proxy 30 and the client-side proxy 40 such as a fingerprint value exists in the reply header or the like, this is deleted.
[0191]
(8) Then, the client side proxy 30 sends a reply message back to the client 50 (such as a browser operating on the client 50).
[0192]
By the way, since the fingerprint caches of the server side proxy 30 and the client side proxy 40 have an upper limit in capacity, garbage collection is performed according to a predetermined algorithm, for example, old data or data that is unlikely to be used is deleted. Is preferred.
[0193]
However, in this case, even if the FP cache 34 of the server-side proxy 30 is held, data that has already been deleted can be generated in the FP cache 44 of the client-side proxy 40, so see FIG. In (7) in the operation described above, the client-side proxy 40 tried to obtain data to replace the reply data from the FP cache 44 based on the fingerprint, but the fingerprint corresponding to the FP cache 44 There may be no data set. In such a case, for example, the client-side proxy 40 requests the server-side proxy 30 to send the specified fingerprint data, and the requested server-side proxy 30 receives the specified fingerprint. It is sufficient to provide a mechanism for fetching the data from the FP cache 34 and sending it back.
[0194]
Conversely, the FP cache 34 of the server side proxy 30 has already been deleted, but the FP cache 44 of the client side proxy 40 is still held. Have When such data exists, the fingerprint / data is set in the client-side proxy 40 in (7) in the operation described with reference to FIG. 30 and (9) in the operation described with reference to FIG. When registering in the FP cache 44, the fingerprint / reply data registered at that time may be overwritten.
[0195]
In (5) in the operation described with reference to FIG. 32, the server-side proxy 30 obtains the fingerprint of the reply data. If the fingerprint is in the FP cache 34, the same data as the ply data is Processing is performed assuming that the fingerprint and the fingerprint are contained in the FP cache 34. In practice, this method is sufficient if the same fingerprint is not generated from different data, but an error that occurs when the fingerprints of different data accidentally become the same value with a very small probability. There is also a way to get rid of. In this case, when the fingerprint obtained from the reply data is in the FP cache 34, the data in the FP cache 34 paired with the fingerprint is compared with the ply data, What is necessary is just to judge whether it is the same. At this time, if it is determined that data having the same fingerprint but different contents is registered, a method exemplified below can be considered.
・ The fingerprint is not used anymore (the data giving the fingerprint will not be cached)
Prioritize fingerprints / data registered earlier (other data giving the same fingerprint value as the fingerprint being registered will not be cached during the registration)
-Give priority to the fingerprint / data currently registered (the fingerprint / data being registered will be updated one after another by other data giving the same value fingerprint)
By the way, in the example described so far, when the reply data is transferred from the server-side proxy 30 to the client-side proxy 40, the fingerprint cache is used, and a set of a certain data and a fingerprint corresponding thereto is used as a fingerprint. The data is registered in the cache when the data is transferred from the server-side proxy 30 to the client-side proxy 40 for the first time. However, in the usage such as Web-based ASP, for example, data is first created in a user office or the like and registered in a server and accessed from a browser or the like in many cases. If the data is registered in the server (as the data main body or differentially compressed data) in the fingerprint cache of the client-side proxy and the server-side proxy, subsequent access can be speeded up. Therefore, if the reply data sent by the server is data originally transferred from the client to the server (however, in this case, request data), the registration request is sent to the original request data as the reply data. May be transferred from the client-side proxy 40 to the server-side proxy 30 for the first time (in this case, when the data becomes reply data and is transferred from the server-side proxy 30 to the client-side proxy 40 for the first time). Since the registration to the fingerprint cache has already been completed, even if it is the first transfer of reply data, the amount of transferred data can be reduced by using the fingerprint cache. it can).
[0196]
In the example described so far, when the reply data is transferred from the server side proxy 30 to the client side proxy 40, if the reply data is the same as the data registered in the fingerprint cache, The network traffic is reduced by transferring a corresponding fingerprint or differentially compressed data instead of the reply data. In the present invention, a request is sent from the client side proxy 40 to the server side proxy 30. The present invention can be further applied to the case of transferring data.
[0197]
In addition, when applying FP compression to both, it is also possible to apply differential compression only about either one.
[0198]
Further, it is also possible to apply FP compression and differential compression only when request data is transferred from the client side proxy 40 to the server side proxy 30.
[0199]
When FP compression and differential compression are applied to request data transfer from the client-side proxy 40 to the server-side proxy 30, the roles of the server-side proxy 30 and the client-side proxy 40 with respect to the reply data already described may be reversed. When the FP compression and differential compression are applied to both data transfers, the server-side proxy 30 further includes a fingerprint decompression processing unit and a differential decompression processing unit in the processing unit 32 in addition to the configuration of FIG. In addition to the configuration of FIG. 20, the proxy 40 may further include a fingerprint compression processing unit and a differential compression processing unit in the processing unit 42.
[0200]
In any proxy, the fingerprint compression processing unit and the fingerprint decompression processing unit may be combined into a fingerprint (FP) compression / decompression processing unit. Similarly, the differential compression processing unit and the differential decompression processing unit may be combined into a differential compression / decompression processing unit.
[0201]
The server-side proxy 30 and the client-side proxy 40 may be provided with a fingerprint cache for request data transfer independently of a fingerprint cache for reply data transfer, but the same for reply data transfer and quest data transfer. The fingerprint cache may be shared. In the case of the configuration using the above-described history table at the time of differential compression, the history table may be provided independently as well, or may be shared.
[0202]
FIG. 33 shows a configuration example of proxies (server-side proxy and client-side proxy) when the same fingerprint cache is shared for reply data transfer and quest data transfer.
[0203]
FIG. 34 and FIG. 35 show an example of a processing procedure of the client side proxy 40 when transferring a request message from the client side proxy 40 to the server side proxy 30. 34 and 35 are the same as in FIG. 21 and FIG. 22, the message to be transferred changes from the reply message to the request message, the message transmission source changes from the server to the client, the message transmission destination changes from the client to the server, and the server side proxy 30. And the operation of the client side proxy 40 are interchanged.
[0204]
36 to 38 show an example of a processing procedure of the server-side proxy 30 when a request message is transferred from the client-side proxy 40 to the server-side proxy 30.
[0205]
In FIGS. 36 to 38, the message to be transferred is changed from the reply message to the request message, the message transmission source is changed from the server to the client, the message transmission destination is changed from the client to the server, and the server side proxy 30 is changed. And the operation of the client side proxy 40 are interchanged.
[0206]
If the request data is replaced with the fingerprint or the differentially compressed data in this way, for example, when uploading the same file to the server many times, it is only necessary to send the fingerprint after the second time. Can reduce traffic.
[0207]
In the present embodiment, the request message transferred from the client-side process to the server-side proxy and the reply message transferred from the server-side proxy to the client-side process have been shown. If both the device that sends the request message and the device that sends the reply message, or the device that sends both the request message and the reply message are connected, of course, it is transferred from the client side process to the server side proxy. Target request messages and reply messages, and request messages and reply messages transferred from the server side proxy to the client side process. It is also such that the target only request message transmitted from the request message and the server side proxy is transferred to the side proxy to the client process.
[0208]
By the way, the description has so far been focused on the one-to-one communication between one server-side proxy and one client-side proxy. Of course, the scope of the present invention is not limited to the server-side proxy and the client-side proxy. The system is not limited to a one-to-one communication system, a server-side proxy and a client-side proxy communicate one-to-many, a server-side proxy and a client-side proxy communicate many-to-one, Alternatively, the present invention can be applied to a system in which a server-side proxy and a client-side proxy communicate many-to-many. For example, as shown in FIG. 39, a client-side proxy installed in a plurality of user offices, a personal proxy used by a mobile user, or the like can be implemented so as to share and use the server-side proxy.
[0209]
In the past, the entire data included in one message was subject to FP compression (target to be registered in the fingerprint cache). For example, the data included in one message is the data of a predetermined unit. In the case of being configured as a set, a configuration in which only a part of unit data included in one message is subject to FP compression (target to be registered in the fingerprint cache) is also possible.
[0210]
By the way, the URL specified in the request message issued by the client for the reply data that can be cached by the shared cache mechanism of the proxy in the server side proxy or the client side proxy (either or both) of this embodiment A reply data included in the reply message corresponding to the request message, a fingerprint corresponding to the reply data, and a request header such as the MIME type included in the reply header of the reply message. Information necessary for configuration and information such as a time stamp to be used for determining the validity period are cached in association with each other (may be stored in the fingerprint cache or information excluding reply data (URL) Fingerprint and may be other information) separately corresponding table holding provided), by a combination of this with fingerprint cache, it is possible to perform also an operation of the shared cache of the proxy server. For example, when the cache function is provided in the client-side proxy, when the reply data for the URL specified in the request message transmitted by the client is cached and the data is valid, the client-side proxy Can retrieve data corresponding to the URL from the fingerprint cache, create a reply message, and respond to the client.
[0211]
Whether or not this function is provided for each server-side proxy 30 and for each client-side proxy 40 can be determined.
[0212]
First, the client side proxy 40 provided with the above functions will be described.
[0213]
FIG. 40 shows a configuration example of the client side proxy 40 in this case. This client-side proxy 40 has a URL / fingerprint table (URL / FP table) 45 that holds correspondence between URLs accessed in the past and fingerprints of reply data in addition to the configuration / function of FIG. And a URL cache processing unit 427.
[0214]
In addition to the URL and fingerprint, the URL / FP table 45 also includes information such as the MIME type entered in the reply header when accessed with the URL, and a time stamp used to determine the validity period. Also record. The URL / FP table 45 records information necessary only when the conventional shared cache can be cached.
[0215]
FIG. 41 shows a processing procedure example of the client side proxy 40 when a reply message is transferred from the server side proxy 30 to the client side proxy 40.
[0216]
The processing procedure in this case is the same as the procedure in FIGS. 23 to 25 except that it is added after terminal 4 and terminal 6 and step S38 in the procedure in FIGS. 23 to 25. The terminal 4 and the terminal 6 of the procedure of FIGS. 23-25, and the part of the process procedure after step S38 are shown. Here, the description will focus on the part added to the procedure described with reference to FIGS.
[0217]
The client side proxy 40 transmits a reply message to the client 50 by the transmission unit 43 (steps S28, S29, S32, S35, or S38 in FIGS. 23 to 25), and then the URL cache processing unit 427 performs the reply message. Is determined whether it is a cache target or not (S39). If it is determined that the URL is to be cached, the URL cache processing unit 427 associates the URL, the fingerprint, the information necessary for constructing the reply header, etc. (using the URL as a key) with the URL / FP. Register in the table 45 (S40). If it is determined not to be cached, nothing is done.
[0218]
The determination in step S39 and the registration in the URL / FP table in step S40 may be performed between step S23 and step S28 or S29, or between step S23 and step S32 or step S35.
[0219]
Note that the method for determining whether or not the received reply message at the time of registration is a cache target may be the same as the conventional method at the time of registration (for example, GET method reply data and its header) Caches that do not contain information indicating that caching is not possible).
[0220]
Next, FIG. 42 and FIG. 43 show an example of a processing procedure regarding the operation of the shared cache of the proxy server in the client side proxy 40 when the request message received from the client 250 is transferred from the client side proxy 40 to the server side proxy 30. Show.
[0221]
The client side proxy 40 receives a request message from the client 50 by the receiving unit 41 (step S141).
[0222]
The URL cache processing unit 427 determines whether or not the reply message for the request message is a cache target (S142). It should be noted that the method for determining whether or not it is a cache target at the time of response may be the same as the conventional method at the time of response (for example, whether or not the reception request message is of the GET method).
[0223]
If the request data is determined not to be cached (S142), the received request message is transferred from the transmission unit 43 to the server side proxy 30 (S151).
[0224]
If it is determined in step S142 that the reply message for the request message is to be cached, the URL cache processing unit 427 further extracts the URL specified in the request message (S143), and the URL The URL / FP table 45 is searched using as a key (S144).
[0225]
If the fingerprint of the reply data corresponding to the URL is not cached (S145), the received request message is transferred from the transmission unit 43 to the server side proxy 30 (S151). At this time, if the time stamp of the currently held data is entered in the If-Modified-Since header of the request message and transferred to the server side proxy 30, the currently held data from the server side proxy 30 is valid. When the reply message is received, the process may be performed to go to step S147.
[0226]
Even if the fingerprint of the reply data corresponding to the URL is registered (S145), it is determined that the data is invalid based on the information for determining the validity period held together. If so (S146), the received request message is transferred from the transmitter 43 to the server-side proxy 30 (S151).
[0227]
On the other hand, the fingerprint of the reply data corresponding to the URL is registered (S145), and it is determined that the data is valid based on the information for determining the validity period held together. In step S146, the URL cache processing unit 427 obtains information necessary to construct reply data from the URL / FP table 45, and searches the FP cache 44 using the fingerprint of the reply data corresponding to the URL as a key. (S147).
[0228]
Next, the FP cache management unit 422 determines whether the data acquired in step S147 is differentially compressed data based on whether the compression identifier is 0 or 1 (S148). If the compression identifier is 0, it is determined not to be differential compressed data, and if it is 1, it is determined to be differential compressed data.
[0229]
When it is determined that the data is differentially compressed data (S148), the differential decompression processing unit 424 decompresses the data and generates reply data (S149). Then, a reply message is generated from the reply data and transferred from the transmission unit 43 to the client (S150).
[0230]
On the other hand, when it is determined that the data is not differentially compressed data (S148), the FP compression / decompression unit 423 generates reply data from the data and transfers it from the transmission unit 43 to the client (S150).
[0231]
Hereinafter, the operation of the shared cache will be described more specifically with reference to FIG. 44 (during response).
[0232]
(1) It is assumed that the browser or the like on the client 50 issues a GET method request message to the server 20 with a URL “/C.html”, for example.
[0233]
(2) When a request is received with a new URL, if the URL is listed in the URL / FP table 45, the validity period is determined in the same manner as in the conventional shared cache. The fingerprint to be obtained is obtained by drawing the URL / FP table 45, and the data having the name is taken out from the FP cache 44. Then, it is checked whether or not the data is differentially compressed data based on the compression identifier. In the example of FIG. 44, since the compression identifier is 1, it is differentially compressed data. Then, the differential compression is decompressed (in FIG. 44, since there is reference information to the data whose fingerprint is “71F0... 73E6” in the differential compression data, the already cached data is referred to and decompressed. This is used as reply data. Further, information necessary for constructing a reply header such as MIME type is extracted from the URL / FP table 45 to create a reply header.
[0234]
(3) The created reply message is sent back to the client 50 (such as a browser operating on the client 50).
[0235]
Even in the case of a request message having an If-Modified-Since header that requests to send data only when the cache contents have been updated after the specified time, it is first updated by referring to the URL / FP table. If it can be determined that there is no such message, a reply message can be created and returned. Otherwise, the server can rewrite the If-Modified-Since information and listen to it.
[0236]
Next, the server side proxy 30 provided with the cache function will be described.
[0237]
Although the cache function of the client-side proxy 40 has been described above, the server-side proxy 30 can be similarly implemented.
[0238]
In this case, the message transfer source client 50 and the message transfer destination server side proxy 30 for the client side proxy 40 are respectively connected to the client side proxy 40 (transfer source) and the server 20 (transfer destination) for the server side proxy 30. Thus, the configuration and procedure related to the cache are the same.
[0239]
FIG. 45 shows a configuration example of the server side proxy 30 in this case. This server-side proxy 30 has a URL / fingerprint table (URL / FP table) 35 that holds correspondence between URLs accessed in the past and fingerprints of reply data in addition to the configuration / function of FIG. And a URL cache processing unit 327.
[0240]
FIG. 46 shows an example of a processing procedure of the server side proxy 30 when a reply message is transferred from the server side proxy 30 to the client side proxy 40.
[0241]
The processing procedure in this case is the same as the procedure in FIGS. 21 and 22 except that it is added after step S6 and terminals 2 and S13 in the procedure in FIGS. 21 and 22, and FIG. And the part of the process procedure after step S6 and the terminal 2 and S13 of the procedure of FIG. 22 is shown. Here, the description will focus on parts that are different from the procedures described in FIGS.
[0242]
The server side proxy 30 transmits a reply message to the client side proxy 40 by the transmission unit 33 (steps S6, S9, S11, or S13 in FIGS. 21 and 22), and then the URL cache processing unit 327 performs the reply. It is determined whether or not the reply data of the message is to be cached (S14). If it is determined that the URL is to be cached, the URL cache processing unit 327 associates the URL, the fingerprint, the information necessary for constructing the reply header, etc. (using the URL as a key) with the URL / FP. It is registered in the table 35 (S15). If it is determined not to be cached, nothing is done.
[0243]
Of course, as described above, this procedure can be variously modified.
[0244]
FIG. 47 shows an example of a processing procedure related to the operation of the shared cache of the proxy server in the server side proxy 30 when the request message received from the client side proxy 40 is transferred from the server side proxy 30 to the server 20.
[0245]
The server-side proxy 30 receives a request message from the client-side proxy 40 by the receiving unit 31 (step S161).
[0246]
The URL cache processing unit 327 determines whether or not the reply message for the request message is a cache target (S162). Note that the method for determining whether or not the response target is a cache target may be the same as the conventional response method (for example, whether or not the reception request message is for the GET method).
[0247]
If the request data is determined not to be cached (S162), the received request message is transferred from the transmission unit 33 to the server 20 (S169).
[0248]
If it is determined in step S162 that the reply message for the request message is to be cached, the URL cache processing unit 327 further extracts the URL specified in the request message (S163), and the URL The URL / FP table 35 is searched using as a key (S164).
[0249]
If the fingerprint of the reply data corresponding to the URL is not cached (S165), the received request message is transferred from the transmission unit 33 to the server 20 (S169). At this time, the time stamp of the currently held data is written in the If-Modified-Since header of the request message and transferred to the server 20, and a reply message from the server 20 that the currently held data is valid is sent. If it is received, the process may go to step S167.
[0250]
Even if the fingerprint of the reply data corresponding to the URL is registered (S165), it is determined that the data is invalid based on the information for determining the validity period held together. If so (S166), the received request message is transferred from the transmission unit 33 to the server 20 (S169).
[0251]
On the other hand, the fingerprint of the reply data corresponding to the URL is registered (S165), and it is determined that the data is valid based on the information for determining the validity period held together. In step S166, the URL cache processing unit 327 obtains information necessary for constructing reply data from the URL / FP table 35, and searches the FP cache 34 using the fingerprint of the reply data corresponding to the URL as a key. (S167).
[0252]
The FP cache management unit 322 generates a reply message from the data acquired from the FP cache 34 using the fingerprint value in a format at the time of FP compression, and transmits it from the transmission unit 33 to the client side proxy 40 (S168). ).
[0253]
As described above, the configuration in which the server side proxy also has the URL / FP table table to perform the cache processing works effectively when one server side proxy is used by a plurality of client side proxies. In other words, if cacheable data requested by a client-side proxy is already being accessed by another client-side proxy, it is also cached by the server-side proxy. Complete.
[0254]
As described above, the configuration in which the server side proxy also has the URL / FP table table to perform the cache processing works effectively when one server side proxy is used by a plurality of client side proxies. In other words, if cacheable data requested by a client-side proxy is already being accessed by another client-side proxy, it is also cached by the server-side proxy. Complete.
[0255]
In the above description, the URL / FP table table is provided separately from the fingerprint cache. However, the URL / FP table table may be integrated with the fingerprint cache.
[0256]
In any proxy, the fingerprint compression processing unit and the fingerprint decompression processing unit may be combined into a fingerprint (FP) compression / decompression processing unit. In any proxy, the differential compression processing unit and the differential decompression processing unit may be combined into a differential compression / decompression processing unit. FIG. 48 shows a configuration example of the proxy (server side proxy, client side proxy) in this case.
[0257]
When the FP compression is applied only to the request data transfer, the FP compression is applied only to the reply data transfer, or the FP compression is applied to the request data transfer and the reply data transfer, the request message is The configuration related to the shared cache function for the reply data corresponding to the specified URL can be provided only in the client side proxy, only in the server side proxy, or in both proxies.
[0258]
The present embodiment may be implemented as a program for causing a computer to execute predetermined means (or for causing a computer to function as predetermined means, or for causing a computer to realize predetermined functions). It can also be implemented as a computer-readable recording medium in which the program is recorded.
[0259]
Note that the configuration illustrated in the embodiment of the present invention is an example, and is not intended to exclude other configurations, and a part of the illustrated configuration may be replaced with another or one of the illustrated configurations. Other configurations obtained by omitting a part, adding another function or element to the illustrated configuration, or combining them are also possible. Also, another configuration that is logically equivalent to the exemplified configuration, another configuration that includes a portion that is logically equivalent to the exemplified configuration, another configuration that is logically equivalent to the main part of the illustrated configuration, and the like are possible. is there. Further, another configuration that achieves the same or similar purpose as the illustrated configuration, another configuration that achieves the same or similar effect as the illustrated configuration, and the like are possible.
In addition, various variations of various components illustrated in the embodiment of the present invention can be implemented in appropriate combination.
Further, the embodiment of the present invention is an invention of an invention as an individual device, an invention of two or more related devices, an invention of the entire system, an invention of components within an individual device, or a method corresponding thereto. The invention includes inventions according to various viewpoints, stages, concepts, or categories.
Therefore, the present invention can be extracted from the contents disclosed in the embodiments of the present invention without being limited to the exemplified configuration.
[0260]
The present invention is not limited to the embodiment described above, and can be implemented with various modifications within the technical scope thereof.
[0261]
【The invention's effect】
According to the present invention, the correspondence between the data and its name is held between the data transfer devices, and the data having this correspondence is transferred by transferring the corresponding name instead of transferring the data body. The amount of data transferred between transfer devices can be reduced.
[Brief description of the drawings]
FIG. 1 is a diagram showing a configuration example of a network system according to an embodiment of the present invention.
FIG. 2 is a view showing another configuration example of the network system according to the embodiment;
FIG. 3 is a diagram showing still another configuration example of the network system according to the embodiment;
FIG. 4 is a diagram for explaining a fingerprint
FIG. 5 is a diagram for explaining a fingerprint cache according to the embodiment;
FIG. 6 is a diagram for explaining a differential compression instruction method used in the embodiment;
FIG. 7 is a diagram for explaining a specific example of differential compression according to the embodiment;
FIG. 8 is a diagram for explaining a specific example of differential compression according to the embodiment;
FIG. 9 is a diagram for explaining a specific example of differential compression according to the embodiment;
FIG. 10 is a diagram for explaining another specific example of differential compression according to the embodiment;
FIG. 11 is a diagram for explaining another specific example of differential compression according to the embodiment;
FIG. 12 is a diagram for explaining another specific example of differential compression according to the embodiment;
FIG. 13 is a diagram for explaining another specific example of differential compression according to the embodiment;
FIG. 14 is a view showing an example of a message format used in the embodiment.
FIG. 15 is a view showing another example of the message format used in the embodiment.
FIG. 16 is a view showing still another example of the message format used in the embodiment.
FIG. 17 is a view showing still another example of the message format used in the embodiment.
FIG. 18 is a diagram showing still another example of the message format used in the embodiment.
FIG. 19 is a view showing a configuration example of a server side proxy according to the embodiment;
FIG. 20 is a view showing a configuration example of a client side proxy according to the embodiment;
FIG. 21 is a flowchart showing a procedure example of the server side proxy according to the embodiment;
FIG. 22 is a flowchart showing a procedure example of the server-side proxy according to the embodiment;
FIG. 23 is a flowchart showing a procedure example of the client side proxy according to the embodiment;
FIG. 24 is a flowchart showing a procedure example of the client side proxy according to the embodiment;
FIG. 25 is a flowchart showing a procedure example of the client side proxy according to the embodiment;
FIG. 26 is a flowchart showing an example of a difference compression processing procedure according to the embodiment;
FIG. 27 is a flowchart showing another example of the difference compression processing procedure according to the embodiment;
FIG. 28 is a flowchart showing a procedure example of the server side proxy according to the embodiment;
FIG. 29 is a flowchart showing a procedure example of a client side proxy according to the embodiment;
FIG. 30 is a view for explaining data transfer between proxies according to the embodiment;
FIG. 31 is a view for explaining data transfer between proxies according to the embodiment;
FIG. 32 is a view for explaining data transfer between proxies according to the embodiment;
FIG. 33 is a diagram showing another configuration example of the proxy in the embodiment;
FIG. 34 is a flowchart showing another procedure example of the client side proxy according to the embodiment;
FIG. 35 is a flowchart showing another procedure example of the client side proxy according to the embodiment;
FIG. 36 is a flowchart showing another procedure example of the server-side proxy according to the embodiment;
FIG. 37 is a flowchart showing another procedure example of the server-side proxy according to the embodiment;
FIG. 38 is a flowchart showing another procedure example of the server-side proxy according to the embodiment;
FIG. 39 is a diagram showing still another configuration example of the network system according to the embodiment;
FIG. 40 is a view showing another configuration example of the client side proxy according to the embodiment;
FIG. 41 is a flowchart showing another procedure example of the client side proxy according to the embodiment;
FIG. 42 is a flowchart showing another procedure example of the client side proxy according to the embodiment;
FIG. 43 is a flowchart showing another procedure example of the client side proxy according to the embodiment;
44 is a view for explaining data transfer between the client and the client-side proxy according to the embodiment; FIG.
FIG. 45 is a view showing another configuration example of the server side proxy according to the embodiment;
FIG. 46 is a flowchart showing another procedure example of the server-side proxy according to the embodiment;
FIG. 47 is a flowchart showing another procedure example of the server-side proxy according to the embodiment;
FIG. 48 is a view showing still another configuration example of the proxy according to the embodiment;
FIG. 49 is a diagram for explaining a conventional computer network system;
[Explanation of symbols]
2 ... ASP server center
4. User office
12 ... ASP server center LAN
14 ... WAN
16 ... LAN in user office
20 ... Server device
30 ... Server-side proxy device
40 ... Client-side proxy device
31, 41 ... receiving unit
32, 42 ... processing section
33, 43 ... Transmitter
34, 44 ... Fingerprint cash
35, 45 ... URL / FP table
321, 421 ... FP compression determination unit
322, 422 ... Fingerprint cache management unit
323... FP compression processing unit
423 ... FP decompression processing part
324 ... Differential compression processing unit
424 ... Differential decompression processing unit
325, 425 ... FP compression / decompression processor
326, 426 ... differential compression / decompression processor
327, 427 ... URL cache processing unit
50: Client device

Claims (10)

In a data transfer device that relays communication between a server device and another data transfer device that relays data transmitted to the client device by the server device,
A first complete data was transmitted to another data transfer apparatus is received from the server apparatus, a first fingerprint generated based on the full data of the first, the first complete data Holding means for holding in association with first identification information indicating complete data ;
Receiving means for receiving a second complete data by the server device transmits the client apparatus as a destination,
Difference data that is a difference between the first complete data and the second complete data, a second fingerprint generated based on the second complete data, and the difference data is difference data Rutotomoni is held by the holding means in association with second identification information indicating, and the difference data and the second identification information, further comprising a transmission means for transmitting said client device as a destination A data transfer device. The transmission means includes
If the second fingerprint matches the first fingerprint, the second fingerprint and third identification information indicating that the second fingerprint is a fingerprint Sent to the client device,
If the second fingerprint does not match the first fingerprint, the difference data, the second fingerprint, and the second identification information are associated with each other and held in the holding unit. The data transfer apparatus according to claim 1, wherein the difference data and the second identification information are transmitted to the client apparatus. A determination unit configured to determine whether or not a difference in data amount obtained by subtracting the difference data from the second complete data is greater than a predetermined value when the second fingerprint does not match the first fingerprint; In addition,
The transmission means includes
If the second fingerprint matches the first fingerprint, the second fingerprint and third identification information indicating that the second fingerprint is a fingerprint Sent to the client device,
When the second fingerprint does not match the first fingerprint, and the determination means determines that the difference in data amount obtained by subtracting the difference data from the second complete data is greater than a predetermined value. In this case, the difference data, the second fingerprint, and the second identification information are associated with each other and held in the holding unit, and the difference data and the second identification information are Sent to the client device,
When the second fingerprint does not match the first fingerprint, and the difference in the data amount obtained by subtracting the difference data from the second complete data is not greater than a predetermined value. If it is determined, the holding means is associated with the second complete data, the second fingerprint, and the first identification information indicating that the second complete data is complete data. 2. The data transfer device according to claim 1, wherein the second complete data and the first identification information are transmitted to the client device. In another data transfer device that relays data transmitted from the server device to the client device and the data transfer device that relays communication with the client device,
First and complete data, said a first fingerprint first full data generated based, the first complete data complete data transmission received from another data transfer apparatus to the client apparatus holding means for holding in association with first identification information indicating that it,
The client device is destined to receive the differential data which is the first complete data and the difference between the second complete data and a second identification information indicating that said difference data is a difference data Receiving means;
The second complete data is generated using the first complete data and the difference data, the difference data, a second fingerprint generated based on the second complete data, and the first data Rutotomoni is held by the holding means in association with second identification information, the full data of the second data transfer apparatus characterized by comprising a transmitting means for transmitting the client device as a destination. The transmission means includes
When the receiving means receives the first identification information, the third complete data received accompanying the first identification information is transmitted to the client device,
The receiving means receives a third fingerprint and third identification information indicating that the third fingerprint is a fingerprint, and the third fingerprint is the first fingerprint. If it matches the fingerprint, send the first complete data associated with the first fingerprint to the client device;
When the receiving means receives the third fingerprint and the third identification information, and the third fingerprint matches the second fingerprint, the first fingerprint Transmitting the second complete data generated using the first complete data associated with the fingerprint and the differential data associated with the second fingerprint to the client device. The data transfer apparatus according to claim 4. The name, the data transfer apparatus according to any one of claims 1 to 5, characterized in that the data is a value obtained by applying a predetermined hash function. A data transfer method in a data transfer device that relays communication between a server device and another data transfer device that relays data transmitted by the server device to a client device,
A first complete data was transmitted to another data transfer apparatus is received from the server apparatus, a first fingerprint generated based on the full data of the first, the first complete data Holding the first identification information indicating complete data in association with the holding means ;
Receiving the second complete data sent by the server device to the client device;
Difference data that is a difference between the first complete data and the second complete data, a second fingerprint generated based on the second complete data, and the difference data is difference data is held by the holding means in association with second identification information indicating a Rutotomoni, and the difference data and the second identification information, and characterized by a step of transmitting said client device as a destination Data transfer device. A data transfer method in another data transfer device that relays data transmitted from a server device to a client device, and a data transfer device that relays communication with the client device,
First and complete data, said a first fingerprint first full data generated based, the first complete data complete data transmission received from another data transfer apparatus to the client apparatus a step of holding the first identification information holding means in association with each other to indicate that it,
Receiving differential data that is a difference between the first complete data and the second complete data, and second identification information indicating that the differential data is differential data, which is addressed to the client device; When,
Using said difference data and the first complete data to generate the second complete data, and the difference data, a second fingerprint generated based on the full data of the second, the third Rutotomoni is held by the holding means in association with second identification information, the full data of the second data transfer method characterized by a step of transmitting said client device as a destination. A program for causing a computer to function as a data transfer device that relays communication between a server device and another data transfer device that relays data transmitted by the server device to a client device,
A first complete data was transmitted to another data transfer apparatus is received from the server apparatus, a first fingerprint generated based on the full data of the first, the first complete data A function of associating with the first identification information indicating complete data in association with the storage device,
A receiving function for receiving the second complete data transmitted by the server device to the client device;
Difference data that is a difference between the first complete data and the second complete data, a second fingerprint generated based on the second complete data, and the difference data is difference data It is held in the storage device in association with second identification information indicating a Rutotomoni, realized and the difference data and the second identification information, and that functions to send to the computer the client device as a destination Program to let you. A program for causing a computer to function as another data transfer device that relays data transmitted from a server device to a client device, and a data transfer device that relays communication with the client device,
First and complete data, said a first fingerprint first full data generated based, the first complete data complete data transmission received from another data transfer apparatus to the client apparatus a function of holding the storage device in association with first identification information indicating that it,
The client device is destined to receive the differential data which is the first complete data and the difference between the second complete data and a second identification information indicating that said difference data is a difference data Function and
The second complete data is generated using the first complete data and the difference data, the difference data, a second fingerprint generated based on the second complete data, and the first data Rutotomoni is held in the storage device in association with second identification information, the full data of the second program for realizing the to that function transmits to the computer the client device as a destination.

Images