Integrated search server and integrated search method

• US 8,346,780 B2
• Filed: 04/16/2010
• Issued: 01/01/2013
• Est. Priority Date: 04/16/2010
• Status: Active Grant

• Current Assignee: Hitachi, Ltd.
• Original Assignee: Hitachi, Ltd.
• Inventors: Yokoi, Kazuhito, Ishii, Yohsuke
• Primary Examiner(s): AL HASHEMI, SANA A
• Application Number:
  US12/740,550
• Publication Number:
  US 20120047168A1
• Time in Patent Office: 991 Days
• Field of Search: 707/706, 707/708, 707/713, 707/727, 707/732, 707/748, 707/754, 707/755
• US Class Current: 707/748
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

In order to mitigate a load in communications among an integrated search server, a plurality of search users and a plurality of search servers in an integrated search system, in the present invention, when an integrated search is executed, first, information for narrowing down the search servers is obtained from each search server, and subsequently, the search servers are narrowed down based on this information, and search results are obtained from only the search servers which have been narrowed down. Thereby, there are advantageous effects of reducing memory usage of the integrated search server, and reducing unnecessary network traffic between the integrated search server and each search server.

Claims

1. An integrated search server which is coupled with a plurality of search clients, a plurality of search servers, and a plurality of file servers, the integrated search server comprising:

a processor, and a main storage device,

wherein the processor obtains digest information which is related to a search condition included in an integrated search result obtaining request received from one of the plurality of search clients, and which is for narrowing down search servers from which search results are obtained, from each of the plurality of search servers, and narrows down the search servers from which the search results are obtained, from the plurality of search servers, based on the obtained plural pieces of digest information and a predetermined search server narrow-down rule,

wherein the processor further obtains the search results corresponding to the integrated search result obtaining request, from the search servers which have been narrowed down, integrates the search results, and provides the integrated search result to the search client,

wherein each of the plural pieces of digest information is a score value of each document included in the search results provided based on the search condition by the plurality of search servers,

wherein the processor compares the score values for the plurality of search servers, and thereby narrows down the search servers from which the search results are obtained,

wherein the number of search results to be provided to the search client is set to k (an integer equal to or larger than

     1),

wherein each of the plural pieces of digest information is a maximum score value and a minimum score value of the documents included in the search results in each of the plurality of search servers, and

wherein the processor selects search servers (candidate search servers) having the number of search results, in which a sum of the search results is equal to or larger than the k, in descending order of the maximum score value, selects search servers (supplementary search servers) in which the maximum score value is larger than the minimum score value which is the smallest in the candidate search servers, from search servers other than the selected candidate search servers, narrows the servers from which the search results are obtained, down to the candidate search servers and the supplementary search servers, and provides k search results with top score values, from the search results provided from the search servers which have been narrowed down, to the search client.

3. The integrated search server according to claim 1, wherein each of the plural pieces of the digest information is information on the number of search results which are obtained by executing a search in a search index by using the search condition by the plurality of search servers,

the processor receives a plurality of the integrated search result obtaining requests from the plurality of search clients, and generates a plurality of digest information obtaining requests for obtaining the digest information corresponding to the plurality of integrated search result obtaining requests, from respective search conditions included in the plurality of integrated search result obtaining requests, and

the processor

transmits the respective plurality of digest information obtaining requests to the respective plurality of search servers,

receives a plurality of pieces of the digest information transmitted from the respective plurality of search servers in response to the plurality of digest information obtaining requests, uses each of the received plurality of pieces of the digest information to select each of a plurality of search server groups having the number of search results which is equal to or larger than a predetermined number,

transmits search result obtaining requests corresponding to the plurality of integrated search result obtaining requests, to the respective selected plurality of search server groups,

obtains a plurality of the search results from the respective selected plurality of search server groups, integrates the search results, and provides the integrated search results to the respective plurality of search clients.

4. The integrated search server according to claim 1, wherein

the number of search results to be provided to the search client is set to k (an integer equal to or larger than 1),

the number of top score values to be used as the digest information is defined in advance for each of the plurality of search servers, and

the processor sorts the score values obtained from each of the plurality of search servers in descending order, narrows the search servers from which the search results are obtained, down to all search servers including a predetermined number of top score values, and provides the predetermined number of the search results to the search client.

5. The integrated search server according to claim 1, wherein

each of the plural pieces of digest information is information on a language used in index information in each of the search servers, and

the processor collates a language used in a search keyword included in the search condition with the index information, and narrows the search servers from which the search results are obtained, down to search servers including all the languages used in the keyword.

6. The integrated search server according to claim 1, wherein

each of the plural pieces of digest information is information regarding an access right of the search client to each of the plurality of search servers, and

the processor collates access right information on the search client which is included in the search condition, with the digest information, and narrows the search servers from which the search results are obtained, down to search servers to which the access right exists.

7. The integrated search server according to claim 1, wherein

the processor temporarily stores the plural pieces of digest information in the main storage device, and uses the plural pieces of digest information to obtain search results corresponding to next and subsequent search requests including the same condition as the search condition, and

after at least one of the plurality of search servers has performed an index creation process for the file server which is associated therewith, the processor deletes the plural pieces of stored digest information.

2. An integrated search method in an integrated search system having a search client, a plurality of search servers, a plurality of file servers, and an integrated search server, wherein

the integrated search server receives an integrated search result obtaining request from the search client, and transmits a digest information obtaining request for requesting digest information, to each of the plurality of search servers,

each of the plurality of search servers receives the digest information obtaining request, and transmits the digest information which is related to a search condition included in the integrated search result obtaining request, and which is for narrowing down search servers from which search results are obtained, to the integrated search server,

the integrated search server obtains the digest information from each of the plurality of search servers, narrows down the search servers from which the search results are obtained, from the plurality of search servers, based on the obtained plural pieces of digest information and a predetermined search server narrow-down rule, and transmits a search result obtaining request including the search condition to the search servers which have been narrowed down,

the search servers which have been narrowed down transmit the search results to the integrated search server in response to the search result obtaining request,

the integrated search server integrates the search results obtained from each of the search servers which have been narrowed down, and provides the integrated search result to the search client,

wherein the number of search results to be provided to the search client is set to k (an integer equal to or larger than 1),

wherein the digest information is configured to have one of information on a score value of each document included in the search results provided based on the search condition by the plurality of search servers, information on a language used in index information in each of the search servers, and information regarding an access right of the search client to each of the plurality of search servers, or to have information including a combination of a plurality thereof,

wherein depending on the digest information, the integrated search server narrows the search servers from which the search results are obtained, down to search servers matching one condition or a combination of conditions, in search servers which provide top k score values, search servers including all languages used in a search keyword included in the search condition, and search servers to which the access right is included in the search client,

wherein each of the plural pieces of digest information is a maximum score value and a minimum score value of the documents included in the search results in each of the plurality of search servers, and

wherein the processor selects search servers (candidate search servers) having the number of search results, in which a sum of the search results is equal to or larger than the k, in descending order of the maximum score value, selects search servers (supplementary search servers) in which the maximum score value is larger than the minimum score value which is the smallest in the candidate search servers, from search servers other than the selected candidate search servers, narrows the servers from which the search results are obtained, down to the candidate search servers and the supplementary search servers, and provides k search results with top score values, from the search results provided from the search servers which have been narrowed down, to the search client.

8. The integrated search method according to claim 2, wherein

the integrated search server temporarily stores the digest information in a main storage device, and uses the digest information to obtain search results corresponding to next and subsequent search requests including the same condition as the search condition, and

if the plural pieces of digest information have been temporarily stored in the main storage device in the integrated search server, the file server manages a storage location of the digest information, and identification information on the search server which has transmitted the digest information.

9. The integrated search method according to claim 8, wherein

at least one of the plurality of search servers executes an index creation process for the file server which is associated with the search server,

if a search index has been rewritten by the index creation process, the file server transmits a cache deletion request for deleting the temporarily stored digest information from the main storage device, to the integrated search server, and

in response to the cache deletion request, the integrated search server deletes the temporarily stored digest information from the main storage device.