JP2004127125A - Search server device, search method, program, and recording medium - Google Patents

Abstract

Provided is a technique for selecting an optimal node in a distributed search system.
A node configuring a distributed search system includes: a metadata database that stores at least one search target data having metadata; a reception unit that receives a request for specifying metadata and stores the metadata; Searches for search target data corresponding to the metadata specified in the search request received by the reception unit, and is suitable for requesting a search when the search target data is not stored in the metadata database. A search unit that selects a node based on the search result history table and, if stored, a communication address and evaluation value data to a search request source; and a communication address and a communication address that are passed from the node selected by the search unit. Updates the search result history table based on the evaluation value and the metadata stored by the accepting means And a that updating means.
[Selection diagram] FIG.

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a technology for performing an efficient search in a distributed search system in which a plurality of search server devices are connected via a communication line.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, a distributed search system in which a plurality of search server devices are connected via a communication line and each search server device performs keyword search in a distributed manner has been known (for example, Non-Patent Document 1 and Non-Patent Document 2). ).
As an example of such a distributed search system, when there is no data corresponding to the search server device requested to search, the search server device selects another search server device and sends the selected search server device to the selected search server device. There is a distributed search system that repeats a process of requesting a search until applicable data is found or until another search server runs out. Such a distributed search system will be described with reference to FIG.
[0003]
Each of the nodes A, B, C, and D shown in FIG. 15 is a search server device, stores search target data representing a document to be searched, and receives search request data in which search conditions are specified. A search is performed by comparing a keyword specified by the search request data with a character string constituting a document represented by the search target data.
The node A stores a search result history table that stores search result history data indicating the history of past search results. The search result history data is data generated for each search performed at each node, has items such as a keyword, a node, and an evaluation value, and represents at least one keyword used for the search; It is composed of data indicating the node and data indicating an evaluation value calculated based on the result of the search. This evaluation value increases as the number of hit search target data items increases, and decreases as the number of hit search data items decreases.
Upon receiving the search request data specifying the two search keywords “X” and “Y”, the node A performs a keyword search on the stored search target data using the keywords “X” and “Y” ( Process P1). If no search target data is found in this search, the node A selects search result history data having data representing these search keywords in the keyword item from the search result history table. At this time, when there is a plurality of corresponding search result history data, the search result history data having the largest evaluation value indicated by the data of the evaluation value item is selected. Further, the node A sends the search request data for searching the search target data to which the two search keywords “X” and “Y” are specified to the node D indicated by the data of the node item of the search result history data thus selected. Is transmitted (process P2). The node D that has received the search request data performs a keyword search using the keywords “X” and “Y” on the stored search target data. When the search target data hits in this search, the node D transmits to the node A data indicating the number of hit search target data and search result data including the search target data (process P3). Node A updates the search result history table based on the received search result data (process P4).
[0004]
[Non-patent document 1]
NEROGRID P2PSEARCH [Philosophy 21-08-01 v03] Internet <URL: http: // www. neurogrid. net / php / philosophy. php>
[Non-patent document 2]
NeuroGrid P2P search Distributed search Internet <URL: http: // www. jnutella. org / jnudev / jws # 0522 / jws # sam. lzh>
[0005]
[Problems to be solved by the invention]
As described above, in the keyword search, all search target data representing a document including the search keyword is extracted. For example, when the search keyword is "Shibuya", the search target data is extracted for all documents including the character string "Shibuya" without distinction between "Shibuya" representing a person's name and "Shibuya" representing a place name. You. Since the search result history table is updated based on such a search result, the next time a keyword search in which “Shibuya” is specified as a search keyword is performed, the node to which the search request data is transmitted is a character string. A node that stores a large number of search target data including the column “Shibuya” is selected. This node selection is appropriate if the user requesting the search wants to extract the search target data including the character string “Shibuya”. However, usually, a user inputs a search keyword “Shibuya” assuming either a person name or a place name, and a user who wants to extract search target data including “Shibuya” as a person name is sent to a user. There is no point in providing search target data that includes "Shibuya" as a place name. That is, there is a problem that the node selected by the above-described node selection is not always appropriate.
[0006]
When a search in which a plurality of search keywords “Shibuya” and “Chinese restaurant” are specified is performed, “Shibuya” is included in the keyword item of the search result history data stored in the search result history table based on the search. And "Chinese restaurant". Therefore, next, when a search is performed in which the same search keywords “Shibuya” and “Chinese restaurant” are specified, the node of the transmission destination of the search request data can be selected using the search result history table. . However, some search keywords such as "Shibuya" and "French" are different from "Shibuya" and "Chinese restaurant", or "Shibuya" such as "Shibuya", "Chinese restaurant" and "luxury" If a new search keyword that is not included in “” and “Chinese restaurant” is added, the corresponding search result history data cannot be extracted from the search result history table. In other words, despite having performed a search for "Shibuya" and "Chinese restaurant" in the past, when selecting a node to which search request data is to be transmitted, the search result cannot be used, and as a result, There is a risk of selecting an inappropriate node.
The present invention has been made in view of such circumstances, and has as its object to provide a technique capable of selecting an optimum node storing search target data to be searched in a distributed search system. And
[0007]
[Means for Solving the Problems]
In order to solve the above problem, the present invention provides a method in which each of a plurality of search server devices stores search target data to be searched, executes a requested search, and does not store the corresponding search target data. In a search server device constituting a distributed search system that selects another search server device, requests the selected other search server device for the search, and obtains the result of the search in the other search server device, And data storage means for storing at least one search target data having metadata indicating the meaning of the character string and semantic information indicating the meaning of the character string, and receiving a request for specifying the metadata and storing the metadata Means, selection storage means for storing selection data for selecting a search server device, and meta data specified in a search request received by the reception means. Search for search target data corresponding to the search data, and when the search target data is not stored in the data storage means, select a search server device suitable for requesting the search based on the selected data. If stored, a search unit that passes identification data for uniquely identifying itself and evaluation value data indicating an evaluation value indicating a corresponding degree to a request source of the search, and is selected by the search unit. And updating means for updating the selection data stored in the selection storage means based on the identification data and the evaluation value passed about the search from the search server device and the metadata stored by the reception means. A search server device is provided.
[0008]
In a preferred aspect, the search server device is configured such that the selection data stored in the selection storage means is metadata, evaluation value data passed from a search server device that is a search request destination that specifies the metadata, and identification data. Data is associated with each request for the search, and the search unit, if the corresponding search target data is not stored in the data storage unit, stores the metadata in the metadata stored by the reception unit. An evaluation index, which is a value obtained by adding the evaluation values represented by the evaluation value data stored in the selection storage means in association with each other, is calculated for each identification data, and the search server corresponding to the identification data having the largest evaluation index is calculated. Select a device.
[0009]
In a preferred aspect, the search server device is configured such that the selection data stored in the selection storage means is metadata, evaluation value data passed from a search server device that is a search request destination that specifies the metadata, and identification data. Data is associated with each metadata, and the search unit associates the data with the metadata stored by the reception unit when the corresponding search target data is not stored in the data storage unit. Calculating an evaluation index, which is a value obtained by adding the evaluation value represented by the evaluation value data stored in the selection storage means, for each identification data, and the search server device having the evaluation index represented by the largest identification data Select
[0010]
In a preferred aspect, the search server device is configured such that the selection data stored in the selection storage means is metadata, evaluation value data passed from a search server device that is a search request destination that specifies the metadata, and identification data. And the data is associated with each metadata. If the data to be searched is not stored in the data storage means, the search means specifies the search request in the search request received by the reception means. According to the importance of the metadata, the evaluation index is a value obtained by increasing and decreasing the evaluation value represented by the evaluation value data stored in the selection storage means in association with the metadata and identifying the evaluation data. The search server device is calculated every time, and the search server device represented by the identification data having the largest evaluation index is selected.
[0011]
In a preferred aspect, the search server device is configured such that the selection data stored in the selection storage means is metadata, evaluation value data passed from a search server device that is a search request destination that specifies the metadata, and identification data. Data is associated with each search request or each metadata, and the search unit is stored by the reception unit when the corresponding search target data is not stored in the data storage unit. An evaluation index, which is the largest value among the evaluation values represented by the evaluation value data stored in the selection storage means in association with the metadata, is calculated for each identification data, and the identification data having the largest evaluation index is calculated. Select the search server device corresponding to.
[0012]
In a preferred aspect, the search server device is configured such that the selection data stored in the selection storage means is metadata, evaluation value data passed from a search server device that is a search request destination that specifies the metadata, and identification data. Data is associated with each search request or each metadata, and the search unit is stored by the reception unit when the corresponding search target data is not stored in the data storage unit. An evaluation index, which is a value obtained by adding an absolute value of a value obtained by subtracting a predetermined threshold value from an evaluation value represented by the evaluation value data stored in the selection storage means in association with the metadata, for each identification data. And the search server device corresponding to the identification data having the smallest evaluation index is selected.
[0013]
In a preferred embodiment, the search server device identifies the selection data stored in the selection storage unit with metadata and evaluation value data passed from a search request destination search request device that specifies the metadata. Data and threshold data indicating an optimum evaluation value are associated with each other for each search request or each metadata, and the search unit is configured to store the corresponding search target data in the data storage unit. The absolute value of the value obtained by subtracting the threshold value corresponding to the evaluation value from the evaluation value represented by the evaluation value data stored in the selection storage means in association with the metadata stored by the reception means Is calculated for each piece of identification data, and a search server device corresponding to the identification data with the smallest evaluation index is selected.
[0014]
In a preferred embodiment, the search server device identifies the selection data stored in the selection storage unit with metadata and evaluation value data passed from a search request destination search request device that specifies the metadata. Data is associated with each search request or each metadata, and the search unit is stored by the reception unit when the corresponding search target data is not stored in the data storage unit. Calculating an evaluation index for each identification data by using a predetermined function based on the evaluation value represented by the evaluation value data stored in the selection storage means in association with the metadata, Select the search server device corresponding to the identification data with the largest value.
[0015]
Further, according to the present invention, a request in which metadata indicating a character string and semantic information indicating a meaning of the character string is received, a receiving step of storing the metadata, and a request received by the receiving step A search step of searching for data to be searched corresponding to the metadata specified in the search request from data storage means for storing at least one piece of search data having metadata; If it is not stored in the search server device, a search server device suitable for requesting the search is selected based on the selection data stored in the selection storage means for storing the selection data for selecting the search server device. If the data to be searched is stored in the data storage means, the identification data for uniquely identifying itself and the corresponding degree A selection step of passing evaluation value data representing an evaluation value indicating a search request source to the search request source; and the selection storage means based on the identification data and the evaluation value passed about the search from the search server device selected in the selection step And an update step of updating the selection data stored in the search method.
[0016]
As a preferred mode, in the search method, the selection data stored in the selection storage unit includes metadata and evaluation value data and identification data passed from a search server device that requests search for the metadata. In the selection step, the evaluation value represented by the evaluation value data stored in the selection storage means in association with the metadata stored in the reception step is configured to be associated with each search request. An evaluation index that is the added value is calculated for each piece of identification data, and a search server device represented by the identification data having the largest evaluation index is selected.
[0017]
As a preferred mode, in the search method, the selection data stored in the selection storage unit includes metadata and evaluation value data and identification data passed from a search server device that requests search for the metadata. In the selection step, an evaluation value represented by the evaluation value data stored in the selection storage means is added in association with the metadata stored in the reception step in the selection step. An evaluation index that is a value is calculated for each piece of identification data, and a search server device represented by the identification data having the largest evaluation index is selected.
[0018]
As a preferred mode, in the search method, the selection data stored in the selection storage unit includes metadata and evaluation value data and identification data passed from a search server device that requests search for the metadata. An evaluation value stored in the selection storage unit in association with the metadata in the selection step, in accordance with the importance of the metadata stored in the reception step, in the selection step; An evaluation index, which is a value added after increasing or decreasing the evaluation value represented by the data, is calculated for each piece of identification data, and a search server device represented by the identification data having the largest evaluation index is selected.
[0019]
As a preferred mode, in the search method, the selection data stored in the selection storage unit includes metadata and evaluation value data and identification data passed from a search server device that requests search for the metadata. When the search target data is not stored in the data storage means in the search step, the metadata stored in the reception step is stored in association with each search request or each metadata. The evaluation index that is the largest value among the evaluation values represented by the evaluation value data stored in the selection storage means is calculated for each piece of identification data, and the evaluation index corresponds to the largest identification data. Select a search server device.
[0020]
As a preferred mode, in the search method, the selection data stored in the selection storage unit includes metadata and evaluation value data and identification data passed from a search server device that requests search for the metadata. When the search target data is not stored in the data storage means in the search step, the metadata stored in the reception step is stored in association with each search request or each metadata. An evaluation index, which is a value obtained by adding an absolute value of a value obtained by subtracting a predetermined threshold value from an evaluation value represented by the evaluation value data stored in the selection storage means in association with And selecting a search server device corresponding to the identification data having the smallest evaluation index.
[0021]
As a preferred mode, the search method is characterized in that the selection data stored in the selection storage unit is configured to include the metadata and the evaluation value data and the identification data passed from the search server device that has requested the search specifying the metadata. Threshold value data representing an appropriate evaluation value is associated with each search request or each metadata, and in the search step, if the corresponding search target data is not stored in the data storage unit, The absolute value of the value obtained by subtracting the threshold value corresponding to the evaluation value from the evaluation value represented by the evaluation value data stored in the selection storage means in association with the metadata stored in the reception step is added. An evaluation index that is a value is calculated for each piece of identification data, and a search server device corresponding to the identification data with the smallest evaluation index is selected.
[0022]
As a preferred mode, in the search method, the selection data stored in the selection storage unit includes metadata, evaluation value data and identification data passed from a search server device that requests search for the metadata. When the search target data is not stored in the data storage means in the search step, the metadata stored in the reception step is stored in association with each search request or each metadata. Based on the evaluation value represented by the evaluation value data stored in the selection storage means in association with the evaluation index, an evaluation index is calculated for each piece of identification data using a predetermined function, and the evaluation index is the largest. Select a search server device corresponding to the identification data.
[0023]
Further, according to the present invention, a computer device has a data storage means for storing at least one search target data having metadata representing a character string and semantic information indicating a meaning of the character string, and the metadata is specified. Receiving means for receiving the request, storing the metadata, selecting storage means for storing selection data for selecting a search server device, and storing the selected data in the metadata specified in the search request received by the receiving means. Search for the relevant search target data, and if the relevant search target data is not stored in the data storage means, select a search server device suitable for requesting the search based on the selection data, and store the search server device. Is passed, the identification data for uniquely identifying itself and the evaluation value data indicating the evaluation value indicating the degree of the corresponding are passed to the search request source. Search means; selection data stored in the selection storage means based on identification data and an evaluation value passed about the search from the search server device selected by the search means and metadata stored by the reception means And a program for realizing an updating means for updating the information.
[0024]
According to this invention, the search server device receives the request in which the metadata is specified, stores the metadata, and stores the search target data corresponding to the metadata specified in the search request received by the receiving unit. If the relevant search target data is not stored in the data storage means, a search server device suitable for requesting the search is selected based on the selection data, and if the search server device is stored, The identification data for uniquely identifying itself and the evaluation value data indicating the evaluation value indicating the degree of the correspondingness are passed to the search request source. Further, the search server device is stored in the selection storage device based on the identification data and the evaluation value passed about the search from the search server device selected by the search device and the metadata stored by the reception device. Update selected data.
[0025]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, a first embodiment and a second embodiment of the present invention will be described with reference to the drawings. In both embodiments, metadata described below is used for the search. Note that the same reference numerals are given to common parts in each drawing.
[0026]
1. First embodiment
1-1. Constitution
FIG. 1 is a block diagram illustrating a configuration of a distributed search system 1 according to the first embodiment. The distributed search system 1 includes a communication network 20 such as the Internet, and nodes 30A, 30B, 30C, and 30D connected to the communication network 20. Since the hardware configuration of each node is common, hereinafter, when there is no need to distinguish each node, it is described as “node 30”. The node 30 is a search server device in which a general WWW (World Wide Web) server device has a function of performing a search process using a metadata database and a search result history table described later. The user terminal 10 is a general computer device, starts a WWW browser, sends a request message according to HTTP to the communication network 20, and receives a response message to the message.
[0027]
1-1-1. Node configuration
FIG. 2 is a block diagram illustrating a hardware configuration of the node 30. As shown in the figure, the node 30 is different from a general WWW server device in that the hard disk device 36 has communication addresses of all the nodes constituting the present system 1, a metadata database 361 described later, and a search result history. The only difference is that the table 362 and the search program 363 are stored, and that the CPU 31 reads out and executes the search program 363 stored in the hard disk device 36. Hereinafter, a function realized in the node 30 by the CPU 31 executing the search program 363 will be described mainly with reference to FIG.
[0028]
The metadata database 361 in this figure stores a large number of search target data created using XML (extensible Markup Language). Each search target data includes at least one piece of metadata. The metadata includes keyword data representing a character string (for example, “Shibuya”) serving as a search keyword and semantic information (for example, characters) indicating what the character string means. Column “place name”).
[0029]
The search request receiving unit 401 shown in FIG. 3 receives search request data addressed to the own node 30 transmitted from the device requesting the search via the communication network 20, and the request content holding unit 406 responds to the data. The search request data is stored until the search result data is returned. The search request data includes search condition data indicating search conditions and data indicating a communication address of a request source device, and the search condition data includes at least one piece of metadata. Hereinafter, the metadata forming the search condition data is referred to as search metadata.
The determining unit 402 extracts from the metadata database 361 search target data including all search metadata included in the search request data (hereinafter, referred to as search request data) received by the search request receiving unit 401. I do. This search process is realized by performing a process of comparing character strings while distinguishing between keyword data and semantic data between all search metadata and all search target data.
When at least one search target data is extracted by this trial, the determination unit 402 instructs the search result transmission unit 403 to return the search result data indicating the search result to the device that transmitted the search request data. The search result data includes the communication address of the request source device, a URL (Uniform Resource Locator) indicating the location of the extracted search target data, the number of extracted search target data, and the communication address of the own node 30.
Note that the search result transmitting unit 403 operates as instructed by the control unit 402.
In addition, when the search target data is not extracted in the search, the determination unit 402 requests the history reference unit 404 to select a node most suitable for the search request represented by the search request data. Further, when a node is selected by the history reference unit 404, the control unit 402 instructs the search request transmission unit 405 to transmit search request data to the node. However, when the communication address of the requesting device in the search request data is different from the communication address of any of the nodes 30, the determination unit 402 determines the communication address of the requesting device in the search request data as the communication address of the own node 30. After rewriting the address, the above instruction is performed. The search result transmitting unit 403 operates according to this instruction. When the node is not selected by the history reference unit 404, the determination unit 402 replaces the empty search result data including the communication address of the request source device and the communication address of the own node 30 with the search request data. The search result transmission unit 403 is instructed to return a reply to the transmission source device.
When selecting the above-described optimal node, the history reference unit 404 refers to the search result history table 362. A specific data configuration example of the search result history table 362 is as conceptually shown in FIG. As shown in this figure, the search result history table 362 stores search result history data created when a search result receiving unit 407 described later receives search result data from another node. The search result history data is data in which metadata forming search condition data is associated with bibliographic data that bibliographically represents a result of performing a search using the search condition data at the other node. The bibliographic data is composed of node data for identifying the node that created the search result data, and evaluation value data indicating the number of search target data hit in the search.
When requested by the determination unit 402, the history reference unit 404 expresses the evaluation value represented by the evaluation value data associated with the metadata that matches the search metadata in the search request data as node data. An evaluation index is calculated by adding each node, and a node having the highest calculated evaluation index is selected as an optimal node. However, if there is no metadata that matches the search metadata in the search request data in the search result history table 362, the node is selected according to a predetermined priority. The priority is set to be lower in the order of the nodes 30A, 30B, 30C, and 30D. Also, selection of a node having a higher priority than the own node or the own node is prohibited.
The search result receiving unit 407 receives the search result data for the search request data from the node selected by the history reference unit 404, and compares the communication address of the request source device in the search result data with the communication address of the own node 30. I do. If they do not match, the search result receiving unit 407 instructs the search result transmitting unit 403 to return the search result data to the device that transmitted the search request data, and the request content holding unit 406 stores the stored search request data. Discard the data. Conversely, when the two match, the history update unit 408 creates search result history data based on the search result data and the search request data, and additionally stores the search result history data in the search result history table 362. The search result transmitting unit 403 is instructed to return the data to the transmission source device of the search request data, and the request content holding unit 406 discards the search request data.
[0030]
1-2. motion
Next, the operation of the first embodiment will be described.
[0031]
1-2-1. Node selection processing
First, the node selection processing executed by the CPU 301 in the node 30A will be described with reference to FIGS.
It is assumed that the search result history table 362 previously stores search result history data shown in FIG. 5 created based on past search results.
The metadata database 361 of the node 30A includes ((“semantic information” = “place name”, “character string” = “Shinjuku”) ”and (“ semantic information ”=“ restaurant ”,“ character string ”=“ French It is assumed that search target data having the metadata “)) is not stored.
The metadata database 361 of the node 30B includes ((“semantic information” = “place name”, “character string” = “Shinjuku”) and (“semantic information” = “restaurant”, “character string” = “French It is assumed that 70 search target data having the metadata “)) are stored.
[0032]
First, the user operates the user terminal 10 to operate two search metadata (here, the metadata in which the semantic information “place name” is associated with the character string “Shinjuku”) and the semantic information “restaurant” in the character string “French” Is associated with the metadata). As a result, the search request data including the search condition data including the metadata is transmitted to the node 30A via the communication network 20.
[0033]
As illustrated in FIG. 4, the search request receiving unit 401 of the node 30A receives search request data. (Step S501). Then, the search request receiving unit 401 transmits the search request data to the request content holding unit 406. The request content holding unit 406 stores the received search request data (Step S502).
The search request receiving unit 401 transmits search request data to the determination unit 402.
The determination unit 402 that has received the search request data performs a search process of extracting search target data having search condition data from the metadata database 361 (step S503). That is, the determination unit 402 performs a process of comparing the character strings while distinguishing the character strings from the semantic information between all the search metadata and all the search target data.
Here, since the search target data having the search condition data is not stored in the metadata database 361, the search target data is not extracted (Step S503; No). The determination unit 402 transmits the search condition data to the history reference unit 404, and requests that the optimum node be selected. When requested by the determination unit 402, the history reference unit 404 selects an optimal node based on the search result history data stored in the search result history table 362. (Step S504).
[0034]
Here, a method in which the history reference unit 404 selects the optimum node from the search result history table 362 will be described with reference to FIG.
The history reference unit 404 includes an evaluation value of metadata (“semantic information” = “place name”, “character string” = “Shinjuku”) and metadata (“semantic information” = “restaurant”, “character string” = “ An evaluation index is calculated by adding the evaluation value of “French”) for each node, and a node having the largest calculated evaluation index is selected as an optimal node (processing P11, processing P12).
For example, in the node 30B shown in FIG. 5, two pieces of metadata (“semantic information” = “place name”, “character string” = “Shinjuku”) are stored, and the evaluation values of these pieces of metadata are “ 50 "and" 60 ". Further, since the metadata of “(semantic information) =“ restaurant ”,“ character string ”=“ French ”) is not stored, the evaluation index of the node 30B is 50 + 60 = 110.
Similarly, for the node 30C, one piece of metadata (“semantic information” = “place name”, “character string” = “Shinjuku”) is stored, and its evaluation value is “20”. Also, one piece of metadata (“semantic information” = “restaurant”, “character string” = “French”) is also stored, and its evaluation value is “20”. Therefore, the evaluation index of the node 30C is 20 + 20 = 40.
For the node 30D, the metadata (“semantic information” = “place name”, “character string” = “Shinjuku”) is also ((semantic information) = “restaurant”, “character string” = “French”). Since no metadata is stored, the evaluation index of the node 30D is “0”.
[0035]
As described above, since the node 30B is the largest evaluation index, the history reference unit 404 selects the node 30B as the optimal node, and transmits information that the node 30B is the optimal node to the search request transmitting unit 405. .
The search request transmitting unit 405 transmits the search request data including the search condition data and the data indicating the communication address of the requesting node 30A to the node 30B based on the received information (step S505 in FIG. 4 and FIG. 5). Process P13).
[0036]
1-2-2. History update processing
Next, an operation of updating the search result history table 362 in the node 30B that has received the search request data from the node 30A will be described with reference to FIGS.
[0037]
As shown in FIG. 4, the search request receiving unit 401 of the node B receives search request data. (Step S501). Then, the search request receiving unit 401 transmits the search request data to the request content holding unit 406. The request content holding unit 406 stores the received search request data (Step S502).
Next, the search request receiving unit 401 transmits the search request data to the determining unit 402. The determination unit 402 that has received the search request data performs a search process of extracting search target data having all search metadata from the metadata database 361 (step S503).
Here, since 70 pieces of search target data having search condition data are stored in the metadata database 361, 70 pieces of search target data are extracted (step S503; Yes). The determination unit 402 instructs the search result transmitting unit 403 to return search result data indicating the search result to the node 30A that has transmitted the search request data. The search result transmitting unit 403 includes the communication address of the requesting node 30A, a URL (Uniform Resource Locator) indicating the location of the extracted search target data, the number of extracted search target data “70”, and the The search result data including the communication address is transmitted to the requesting node 30A (step S506).
[0038]
Next, an operation in the node 30A will be described with reference to FIGS.
First, the search result receiving unit 407 receives the search result data transmitted from the search result transmitting unit 403 of the node 30B (step S601 in FIG. 6, processing P21 in FIG. 7). Then, the search result receiving unit 407 compares the communication address of the requesting node 30A included in the search result data with the communication address of the own node 30A. Since both match, the history updating unit 408 searches the search condition data ((“semantic information” = “place name”, “character string” = “Shinjuku”) stored in the request content holding unit 406) and ( “Semantic information” = “restaurant”, “character string” = “French”)), node data indicating the node “30B” to which the search result data was transmitted, and search condition data included in the search result data The search result history data is created from the data indicating the number of extracted “70” (step S602 in FIG. 6). Then, the history update unit 408 additionally stores the search result history data in the search result history table 362 as shown in the updated search result history table 362 in FIG. 7 (step S603 in FIG. 6, processing P22 in FIG. 7). ).
[0039]
Then, the history update unit 408 instructs the search result transmitting unit 403 to return the search result data to the user terminal 10 that has transmitted the search request data, and the request content holding unit 406 discards the search request data.
The search result transmitting unit 403 transmits the search result data to the user terminal 10.
[0040]
As described above, the search using the metadata is performed, and the result of the search performed in the past is stored in the search result history table 362, so that in the subsequent search, the node 30 stores the search result history table 362 in the search result history table 362. By utilizing, it is possible to select the optimum node where the search condition data exists. Therefore, the time from when the user issues a search instruction to when the search result is displayed on the user terminal 10 is reduced, and the user can perform an efficient search.
Also, for a search that combines a plurality of metadata that has not been performed in the past, an evaluation index is calculated by combining evaluation values corresponding to a single piece of metadata stored in the search result history table 362. Therefore, even if the search is a combination of a plurality of metadata, the optimum node can be selected.
[0041]
2. Second embodiment
Next, a second embodiment of the present invention will be described.
In the first embodiment, when the search condition data is composed of a plurality of search metadata, the number of cases where the search target data having the search condition data is stored in the metadata database 361 is used as the evaluation value. In the second embodiment, the number of cases in which search target data having search metadata is stored is used as an evaluation value. When an optimum node is selected, the evaluation value of search metadata specified by a user is weighted. Thus, an evaluation index can be calculated.
[0042]
2-1. Constitution
FIG. 8 is a block diagram illustrating a configuration of the distributed search system 2 according to the second embodiment. The distributed search system 1 includes a communication network 20 such as the Internet, and nodes 50A, 50B, 50C, and 50D connected to the communication network 20. Since the hardware configuration of each node is common, hereinafter, when there is no need to distinguish each node, it is described as “node 50”. The node 50 is a search server device in which a general WWW (World Wide Web) server device has a function of performing a search process using a metadata database and a search result history table described later. The configuration of the user terminal 10 is the same as in the first embodiment.
[0043]
2-1-1. Node configuration
FIG. 9 is a block diagram illustrating a hardware configuration of the node 50. As shown in the figure, the difference between the node 50 and the node 30 in the first embodiment is the data structure of the search result history data stored in the search result history table 562 and the functions of the search program 563. The functions realized in the node 50 by the CPU 31 executing the search program 563 will be described below mainly on the part different from the first embodiment with reference to FIG.
[0044]
The search request receiving unit 401 shown in FIG. 10 receives search request data addressed to the own node 50 transmitted from the device requesting the search via the communication network 20, and the request content holding unit 406 responds to the data. The search request data is stored until the search result data is returned. The search request data includes search condition data indicating search conditions, importance data for specifying weighted metadata to be searched, and data indicating a communication address of a request source device. The search condition data includes at least one search metadata.
When selecting the optimum node, the history reference unit 604 refers to the search result history table 562. A specific data configuration example of the search result history table 562 is as conceptually shown in FIG. As shown in this figure, the search result history table 562 stores search result history data created when a search result receiving unit 407 described later receives search result data from another node. The search result history data is data in which metadata forming search condition data is associated with bibliographic data that bibliographically represents a result of performing a search using the search condition data at the other node. The bibliographic data is composed of node data for identifying the node that created the search result data, and evaluation value data indicating the number of search metadata hit in the search.
When requested by the determination unit 402, the history reference unit 604 weights the evaluation value represented by the evaluation value data associated with the metadata that matches the search metadata in the search request data with the corresponding importance data. After that, the evaluation index is calculated by adding each of the nodes represented by the node data, and the node having the highest calculated evaluation index is selected as the optimal node.
[0045]
The configuration other than the configuration described above is the same as that of the first embodiment, and thus the duplicated description will be omitted.
[0046]
2-2. motion
2-2-1. Node selection processing
First, the node selection process executed by the CPU 301 in the node 50A will be described with reference to FIGS.
It is assumed that the search result history table 562 of the node 50A stores search result history data shown in FIG. 12 created based on past search results.
Also, the metadata database 361 of the node 50A contains (“semantic information” = “place name”, “character string” = “Shibuya”) and (“semantic information” = “restaurant”, “character string” = “Italian” It is assumed that search target data having the metadata of ()) is stored.
The metadata database 361 of the node 50D includes (“semantic information” = “place name”, “character string” = “Shibuya”) and (“semantic information” = “restaurant”, “character string” = “Italian” )) Is stored, and 20 pieces of search target data having metadata (“semantic information” = “place name”, “character string” = “Shibuya”) are stored. It is assumed that ten pieces of search target data having metadata (“semantic information” = “restaurant”, “character string” = “Italian”) are stored.
[0047]
First, as in the first embodiment, the user operates the user terminal 10 and operates two search metadata (here, metadata in which the character string “Shibuya” is associated with the semantic information “place name”; It is assumed that the metadata having the column “Italian” and the semantic information “restaurant” associated thereto and the importance data representing the semantic information “place name” to be searched by weighting are input. As a result, the search request data including the search condition data and the importance data composed of the search metadata is transmitted to the node 50A via the communication network 20.
[0048]
As shown in FIG. 11, the search request receiving unit 401 of the node 50A receives search request data from the user terminal 10. (Step S701). Then, the search request receiving unit 401 transmits the search request data to the request content holding unit 406. The request content holding unit 406 stores the received search request data (Step 702). The search request receiving unit 401 transmits search request data to the determination unit 402.
The determination unit 402 that has received the search request data performs a search process of extracting search target data having search condition data from the metadata database 361 (step S703). That is, the determination unit 402 performs a process of comparing a character string while distinguishing the character string from the semantic information between all the search metadata and all the search target data.
Here, since the search target data having the search condition data does not exist in the metadata database 361, the search target data is not extracted (Step S703; No). The determination unit 402 transmits the search condition data and the importance data to the history reference unit 604, and requests that the optimum node be selected. The history reference unit 604 selects an optimal node based on the search result history data stored in the search result history table 562. (Step S704).
[0049]
Here, a method in which the history reference unit 604 selects an optimal node will be specifically described with reference to FIG.
Here, since the semantic information to be weighted is “place name”, the history reference unit 604 stores the metadata (“semantic information” = “place name”, “character string” = “Shibuya”) of the metadata of the search result history table 562. In order to weight the evaluation value, for example, the evaluation value is doubled. Then, the history reference unit 604 adds the doubled evaluation value and the evaluation value of the metadata (“semantic information” = “restaurant”, “character string” = “Italian”) to obtain the search condition data. An evaluation index for ((“semantic information” = “place name”, “character string” = “Shibuya”) and (“semantic information” = “restaurant”, “character string” = “Italian”)) is obtained. The history reference unit 604 obtains the evaluation index for each node in this way, and selects the node having the largest evaluation index as the optimal node. By calculating the evaluation index in this manner, a node that stores a large amount of search target data having metadata (“semantic information” = “place name” and “character string” = “Shibuya”) in the metadata database 361 is selected. (Processes P31 and P32).
[0050]
For example, as shown in FIG. 12, for the node 50B, the evaluation value of the metadata of (“semantic information” = “place name”, “character string” = “Shibuya”) is “50” and (“meaning information”). Since the evaluation value of the metadata of “information” = “restaurant” and “character string” = “Italian”) is “10”, the evaluation index of the node 50B is 2 × 50 + 10 = 110.
Similarly, for the node 50C, the evaluation value of the metadata “(semantic information) =“ place name ”,“ character string ”=“ Shibuya ”) is“ 20 ”, and (“ semantic information ”=“ restaurant ”). , “Character string” = “Italian”), the evaluation value of the metadata is “20”, and the evaluation index of the node 50B is 2 × 20 + 20 = 60.
For the node 50D, the evaluation value of the metadata of “(semantic information) =“ place name ”,“ character string ”=“ Shibuya ”) is“ 80 ”, and“ (semantic information) = “restaurant”, “character” Since the evaluation value of the metadata of “column” = “Italian”) is “60”, the evaluation index of the node 50B is 2 × 80 + 60 = 220.
[0051]
As described above, since the evaluation index of the node 50D is the largest, the history referring unit 604 selects the node 50D as the optimal node, and transmits information that the node 50D is the optimal node to the search request transmitting unit 405. .
The search request transmitting unit 405 transmits search request data including search condition data, importance data, and data indicating the communication address of the requesting node 50A to the node 50D based on the received information (step S705 in FIG. 11). , Processing P33 in FIG. 12).
[0052]
2-2-2. History update processing
Next, an operation of updating the search result history table 562 in the node 50D that has received the search request data from the node 50A will be described.
[0053]
As illustrated in FIG. 11, the search request receiving unit 401 of the node 50D receives search request data. (Step S701). Then, the search request receiving unit 401 transmits the search request data to the request content holding unit 406. The request content holding unit 406 stores the received search request data (step S702).
Next, the search request receiving unit 401 transmits the search request data to the determining unit 402. The determination unit 402 that has received the search request data performs a search process of extracting search target data having search condition data from the metadata database 361 (step S703).
Here, since the search target data having the search condition data is stored in the metadata database 361, the search target data is extracted (Step S703; Yes). The determination unit 402 also determines, from the metadata database 361, 20 pieces of search target data having metadata that match search metadata (“semantic information” = “place name”, “character string” = “Shibuya”), Ten pieces of search target data having metadata that match the search metadata (“semantic information” = “restaurant”, “character string” = “Italian”) are extracted. The determining unit 402 instructs the search result transmitting unit 403 to return search result data indicating the search result to the node 50A that has transmitted the search request data. The search result transmitting unit 403 includes a communication address of the requesting node 50A, a URL (Uniform Resource Locator) indicating the location of the extracted search target data, the number of search target data having the extracted search metadata, and a node 50D. Is transmitted to the requesting node 50A (step S706).
[0054]
Next, an update process for updating the search result history table 562 in the node 50A will be described with reference to the flowcharts shown in FIGS.
First, the history updating unit 608 of the node 50A receives the search result data transmitted from the search result transmitting unit 403 of the node 50D (Step S801 in FIG. 13 and Process P41 in FIG. 14). Then, the search result receiving unit 407 compares the communication address of the requesting node 50A included in the search result data with the communication address of the own node 50A. Since both match, the history update unit 608 determines that the search metadata (“semantic information” = “place name”, “character string” = “Shibuya”) stored in the request content holding unit 406 and (“ (Semantic information) = “restaurant”, “character string” = “Italian”). Then, the history update unit 608 converts the node data indicating the node “50D” from which the search result data was transmitted, the metadata (“semantic information” = “place name”, “character string” = “Shibuya”) and the search result data Is associated with the data indicating the number of “20 cases” in which the metadata included in the search result data is extracted. Similarly, the node data indicating the node “50D” to which the search result data is transmitted and the metadata (“semantic information”) = "Restaurant", "character string" = "Italian") and data indicating the number of cases where the metadata included in the search result data is extracted is "10", and search result history data is created ( Step S802 in FIG. 13). Then, the history update unit 608 updates the search result history table 562 as shown in FIG. 14 (step S603 in FIG. 13, processing P42 in FIG. 14).
Thereby, as shown in FIG. 14, the data of (“node” = “50D”, “semantic information” = “place name”, “character string” = “Shibuya”, “evaluation value” = “80”) , “Evaluation value” = “20”, (“node” = “50D”, “semantic information” = “restaurant”, “character string” = “Italian”, “evaluation value” = “60”) Certain data is updated to “evaluation value” = “10”.
[0055]
Then, the history updating unit 608 instructs the search result transmitting unit 403 to return the search result data to the user terminal 10 that has transmitted the search request data, and the request content holding unit 406 discards the search request data.
The search result transmitting unit 403 transmits the search result data to the user terminal 10.
[0056]
As described above, by retaining the evaluation value for each single metadata in the search result history table 562, when selecting the optimal node in a search combining a plurality of metadata, a single metadata Since the evaluation index can be calculated by combining the evaluation values, it is possible to calculate an evaluation index that is useful for selecting a more optimal node.
In addition, since the evaluation value of the metadata that the user wants to focus on the search can be weighted, it is possible to calculate the evaluation index according to the user's needs and select the optimal node. .
[0057]
3. Modified example
As described above, the embodiments of the present invention have been described, but the present invention can be embodied in various other forms without departing from the main features. The embodiments described above are merely examples of one aspect of the present invention, and the scope of the present invention is as set forth in the appended claims, and includes modifications and changes that fall within the equivalent scope of the appended claims. Are all included within the scope of the present invention. Note that, for example, the following modifications can be considered.
[0058]
(1) In the above embodiment, as a method of calculating an evaluation index for selecting an optimal node, a method of simply adding an evaluation value and a method of weighting an evaluation value of metadata specified by a user have been described. The method for calculating the evaluation index is not limited to this.
For example, (a) a method is conceivable in which the maximum evaluation value among the nodes is directly used as an evaluation index. Specifically, in the first embodiment, for the node 30B shown in FIG. 5, the evaluation value of the metadata of “(semantic information) =“ place name ”,“ character string ”=“ Shinjuku ”) is“ 50 ”. And “60”, and the metadata (“semantic information” = “restaurant”, “character string” = “French”) is not stored, so the evaluation index of the node 30B is the maximum evaluation value “60”. ". Similarly, for the node 30C, the evaluation value of the metadata of (“semantic information” = “place name”, “character string” = “Shinjuku”) is “20”, and (“semantic information” = “restaurant”). , “Character string” = “French”) has an evaluation value of “20”. Therefore, the evaluation index of the node 30C is “20”. For the node 30D, the metadata (“semantic information” = “place name”, “character string” = “Shinjuku”) is also ((semantic information) = “restaurant”, “character string” = “French”). Since no metadata is stored, the evaluation index of the node 30D is “0”. As described above, the maximum evaluation index is “60”, and the node 30B is selected.
In addition to the method of specifying the metadata to be weighted by the user, a method of calculating the evaluation index by using a specific calculation formula or function and adding the metadata or the node to be weighted can be considered.
For example, (b) a threshold value that is an optimum evaluation value is determined in advance for each node or for all nodes, and the absolute value of the value obtained by subtracting the evaluation value corresponding to the search metadata from the threshold value is determined. , For each node, and selecting a node having the smallest value. For example, if it is determined that the threshold value of the node 30B is “60”, the threshold value of the node 30C is “20”, and the threshold value of the node 30D is “30”, the evaluation index of the node 30B is | 50-60 | + | 60-60 | = 10, | 20-20 | + | 20-20 | = 0 for the node 30C, and | 30-0 | + | 30-0 | = 60 for the node 30D. . As described above, the minimum evaluation index is “0”, and the node 30C is selected.
Also, the evaluation index may be calculated using (c) an arbitrary function, for example, a function for performing fuzzy integration using semantic information of metadata as a fuzzy measure.
In the above-described embodiment, the number of cases where the search condition data or the search target data having the metadata that matches the search metadata is stored in the metadata database 361 is set as the evaluation value. The evaluation value may be a value obtained by multiplying the number of search target data by a number less than 1.
Further, in the second embodiment, weighting is performed by doubling the evaluation value. However, the present invention is not limited to this, and an arbitrary multiple optimal for weighting can be selected. Further, a multiple of the weight may be specified from the user terminal 10 according to the importance of the metadata to be searched by the user.
[0059]
(2) In the above-described embodiment, metadata created using XML is used as search target data. However, any metadata in which a character string is associated with information indicating the attribute of the character string is searched. The embodiment may be modified so as to be target data.
[0060]
(3) In the above embodiment, a communication address is used as information for identifying a node included in search request data or search result data. However, the present invention is not limited to this, and any information that can uniquely identify a node is used. May be used.
[0061]
(4) In the above embodiment, the embodiment in which the AND search is performed by combining a plurality of search metadata is described. However, it goes without saying that the OR search may be performed. In the case of the OR search, for example, in step S503 of the first embodiment, the determination unit 402 does not extract the search target data having the search condition data, but extracts at least one piece of search metadata forming the search condition data. A search process is performed to extract the search target data from the metadata database 361. The evaluation value is the number of search target data having at least one search metadata.
[0062]
(5) In the above embodiment, the prohibition is provided to avoid the infinite loop. However, the prohibition is not required when the infinite loop is allowed. Further, an infinite loop may be avoided by another known method.
[0063]
(6) In the above embodiment, the same search is prevented from being performed repeatedly at one node. However, the present invention is not limited to this. The same search is performed repeatedly at one node. Is also good.
[0064]
【The invention's effect】
As described above, according to the present invention, in a distributed search system, an optimal node can be selected and an efficient search can be performed.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a configuration of a distributed search system according to a first embodiment of the present invention.
FIG. 2 is a block diagram showing a hardware configuration of a node according to the embodiment.
FIG. 3 is a functional block diagram for explaining functions realized by the nodes when the CPU according to the embodiment executes a search program;
FIG. 4 is a flowchart illustrating an operation of a node selection process in the node according to the embodiment;
FIG. 5 is a diagram illustrating an operation of a node selection process in the node according to the embodiment;
FIG. 6 is a flowchart illustrating an operation of a history update process in the node according to the embodiment.
FIG. 7 is a diagram illustrating an operation of a history update process in the node according to the embodiment.
FIG. 8 is a block diagram illustrating a configuration of a distributed search system according to a second embodiment of the present invention.
FIG. 9 is a block diagram showing a hardware configuration of a node according to the embodiment.
FIG. 10 is a functional block diagram for explaining functions realized by the nodes when the CPU according to the embodiment executes a search program;
FIG. 11 is a flowchart illustrating an operation of a node selection process in the node according to the embodiment;
FIG. 12 is a diagram for explaining an operation of a node selection process in the node according to the embodiment.
FIG. 13 is a flowchart illustrating an operation of a history update process in the node according to the embodiment.
FIG. 14 is a diagram illustrating an operation of a history update process in the node according to the embodiment. .
FIG. 15 is a diagram for explaining operations of a conventional node selection process and a history update process.
[Explanation of symbols]
1, 2 ... distributed search system, 10 ... user terminal, 20 ... communication network, 30, 50 ... node, 31 ... CPU, 36, 56 ... hard disk drive, 361 ... metadata database, 362 , 562... Search result history table, 363, 563... Search program, 401... Search request receiving unit, 402. ... Search request transmission, 406... Request content holding unit, 407... Search result receiving unit, 408, 608.

Claims (18)

Each of the plurality of search server devices stores the search target data to be searched, executes the requested search, and when the corresponding search target data is not stored, selects and selects another search server device. A search server device that constitutes a distributed search system that requests the search server device to perform the search and obtains the search result in the other search server device.
Data storage means for storing at least one search target data having metadata representing a character string and semantic information indicating a meaning of the character string,
Receiving means for receiving a request in which metadata is specified, and storing the metadata;
Selection storage means for storing selection data for selecting a search server device;
Searching for search target data corresponding to the metadata specified in the search request received by the receiving unit, and requesting the search if the relevant search target data is not stored in the data storage unit. A suitable search server device is selected based on the selection data, and if stored, identification data for uniquely identifying itself and evaluation value data indicating an evaluation value indicating the degree of the corresponding search server device are stored. Search means to be passed to the requester;
Update for updating the selection data stored in the selection storage means based on the identification data and the evaluation value passed about the search from the search server device selected by the search means and the metadata stored by the reception means And a search server device. In the selection data stored in the selection storage means, metadata, evaluation value data and identification data passed from a search server device that is a search request destination specifying the metadata are associated with each search request. Composed of
When the search target data is not stored in the data storage unit, the search unit associates the evaluation value data stored in the selection storage unit with the metadata stored by the reception unit. The search index according to claim 1, wherein an evaluation index, which is a value obtained by adding the evaluation values represented, is calculated for each piece of identification data, and a search server device corresponding to the identification data having the largest evaluation index is selected. Server device. The selection data stored in the selection storage unit is configured by associating metadata with evaluation value data and identification data passed from a search server device that is a search request destination specifying the metadata for each metadata. And
When the search target data is not stored in the data storage unit, the search unit associates the evaluation value data stored in the selection storage unit with the metadata stored by the reception unit. The evaluation index which is a value obtained by adding the evaluation value represented is calculated for each identification data, and a search server device represented by the identification data having the largest evaluation index is selected. Search server device. The selection data stored in the selection storage unit is configured by associating metadata with evaluation value data and identification data passed from a search server device that is a search request destination specifying the metadata for each metadata. And
The search means, if the relevant search target data is not stored in the data storage means, according to the importance of the metadata specified in the search request received by the reception means, the metadata An evaluation index, which is a value added after increasing or decreasing the evaluation value represented by the evaluation value data stored in the selection storage means in association with the identification data, is calculated for each identification data, and the evaluation index is the largest identification data. The search server device according to claim 1, wherein a search server device to be represented is selected. The selection data stored in the selection storage means is the metadata and the evaluation value data and the identification data passed from the search request destination search server device that specifies the metadata, for each search request or for each metadata. Is configured in correspondence with
When the search target data is not stored in the data storage unit, the search unit associates the evaluation value data stored in the selection storage unit with the metadata stored by the reception unit. The evaluation index that is the largest value among the evaluation values represented is calculated for each piece of identification data, and a search server device corresponding to the identification data having the largest evaluation index is selected. Search server device. The selection data stored in the selection storage means is the metadata and the evaluation value data and the identification data passed from the search request destination search server device that specifies the metadata, for each search request or for each metadata. Is configured in correspondence with
When the search target data is not stored in the data storage unit, the search unit associates the evaluation value data stored in the selection storage unit with the metadata stored by the reception unit. An evaluation index, which is a value obtained by adding an absolute value of a value obtained by subtracting a predetermined threshold value from an evaluation value represented, is calculated for each identification data, and the search server device corresponding to the identification data having the smallest evaluation index is selected. The search server device according to claim 1, wherein the search server device performs the search. The selection data stored in the selection storage means includes metadata, evaluation value data passed from a search server device that is a search request destination specifying the metadata, identification data, and threshold data representing an optimum evaluation value. It is configured in association with each search request or each metadata,
When the search target data is not stored in the data storage unit, the search unit associates the evaluation value data stored in the selection storage unit with the metadata stored by the reception unit. An evaluation index, which is a value obtained by adding an absolute value of a value obtained by subtracting the threshold value corresponding to the evaluation value from the expressed evaluation value, is calculated for each identification data, and the search server corresponding to the identification data with the smallest evaluation index is calculated. The search server device according to claim 1, wherein the search server device is selected. The selection data stored in the selection storage means includes metadata, evaluation value data and identification data passed from the search server device that is a search request destination specifying the metadata, for each search request or for each metadata. Is configured in correspondence with
When the search target data is not stored in the data storage unit, the search unit associates the evaluation value data stored in the selection storage unit with the metadata stored by the reception unit. An evaluation index is calculated for each piece of identification data using a predetermined function based on the expressed evaluation value, and a search server device corresponding to the identification data having the largest evaluation index is selected. Item 2. The search server device according to item 1. Receiving a request in which metadata representing a character string and semantic information indicating the meaning of the character string is specified, receiving a request for storing the metadata, and specifying in a search request received in the receiving step A search step of searching the search target data corresponding to the metadata from the data storage unit storing at least one search target data having the metadata;
If the relevant search target data is not stored in the data storage unit, a search server device suitable for requesting the search is stored in the selection storage unit storing selection data for selecting the search server device. If the data to be searched is selected based on the selected data and the corresponding search target data is stored in the data storage means, identification data for uniquely identifying itself and evaluation value data indicating an evaluation value indicating the degree of the selection. And a selection step of passing a search request to the search request source;
Updating the selection data stored in the selection storage means based on the identification data and the evaluation value passed about the search from the search server device selected in the selection step. . In the selection data stored in the selection storage means, metadata, evaluation value data and identification data passed from a search server device that is a search request destination specifying the metadata are associated with each search request. Composed of
In the selection step, an evaluation index, which is a value obtained by adding an evaluation value represented by evaluation value data stored in the selection storage means in association with the metadata stored in the reception step, is calculated for each identification data. 10. The search method according to claim 9, wherein a search server device represented by the identification data having the largest evaluation index is selected. The selection data stored in the selection storage unit is configured by associating metadata with evaluation value data and identification data passed from a search server device that is a search request destination specifying the metadata for each metadata. And
In the selection step, an evaluation index, which is a value obtained by adding an evaluation value represented by evaluation value data stored in the selection storage means in association with the metadata stored in the reception step, is calculated for each identification data. 10. The search method according to claim 9, wherein a search server device represented by the identification data having the largest evaluation index is selected. The selection data stored in the selection storage unit is configured by associating metadata with evaluation value data and identification data passed from a search server device that is a search request destination specifying the metadata for each metadata. And
In the selection step, after increasing or decreasing the evaluation value represented by the evaluation value data stored in the selection storage means in association with the metadata according to the importance of the metadata stored in the reception step The search method according to claim 9, wherein an evaluation index that is the added value is calculated for each piece of identification data, and a search server device represented by the identification data having the largest evaluation index is selected. The selection data stored in the selection storage means is the metadata and the evaluation value data and the identification data passed from the search request destination search server device that specifies the metadata, for each search request or for each metadata. Is configured in correspondence with
In the search step, if the relevant search target data is not stored in the data storage means, the evaluation value data stored in the selection storage means is associated with the metadata stored in the reception step. The evaluation index that is the largest value among the evaluation values represented is calculated for each piece of identification data, and a search server device corresponding to the identification data having the largest evaluation index is selected. retrieval method. The selection data stored in the selection storage means is the metadata and the evaluation value data and the identification data passed from the search request destination search server device that specifies the metadata, for each search request or for each metadata. Is configured in correspondence with
In the search step, if the relevant search target data is not stored in the data storage means, the evaluation value data stored in the selection storage means is associated with the metadata stored in the reception step. An evaluation index, which is a value obtained by adding an absolute value of a value obtained by subtracting a predetermined threshold value from an evaluation value represented, is calculated for each identification data, and the search server device corresponding to the identification data having the smallest evaluation index is selected. The search method according to claim 9, wherein the search is performed. The selection data stored in the selection storage means includes metadata, evaluation value data passed from a search server device that is a search request destination specifying the metadata, identification data, and threshold data representing an optimum evaluation value. It is configured in association with each search request or each metadata,
In the search step, if the relevant search target data is not stored in the data storage means, the evaluation value data stored in the selection storage means is associated with the metadata stored in the reception step. An evaluation index, which is a value obtained by adding an absolute value of a value obtained by subtracting the threshold value corresponding to the evaluation value from the expressed evaluation value, is calculated for each identification data, and the search server corresponding to the identification data with the smallest evaluation index is calculated. The search method according to claim 9, wherein a device is selected. The selection data stored in the selection storage means includes metadata, evaluation value data and identification data passed from the search server device that is a search request destination specifying the metadata, for each search request or for each metadata. Is configured in correspondence with
In the search step, if the relevant search target data is not stored in the data storage means, the evaluation value data stored in the selection storage means is associated with the metadata stored in the reception step. An evaluation index is calculated for each piece of identification data using a predetermined function based on the expressed evaluation value, and a search server device corresponding to the identification data having the largest evaluation index is selected. Item 9. The search method according to Item 9. Computer equipment,
Data storage means for storing at least one search target data having metadata representing a character string and semantic information indicating a meaning of the character string,
Receiving means for receiving a request in which metadata is specified, and storing the metadata;
Selection storage means for storing selection data for selecting a search server device;
Searching for search target data corresponding to the metadata specified in the search request received by the receiving unit, and requesting the search if the relevant search target data is not stored in the data storage unit. A suitable search server device is selected based on the selection data, and if stored, identification data for uniquely identifying itself and evaluation value data indicating an evaluation value indicating the degree of the corresponding search server device are stored. Search means to be passed to the requester;
Update for updating the selection data stored in the selection storage means based on the identification data and the evaluation value passed about the search from the search server device selected by the search means and the metadata stored by the reception means A program for realizing the means. A computer-readable recording medium recording the program according to claim 17.

Images