CN103309903A - Position search system and method based on cloud computing - Google Patents
Position search system and method based on cloud computing Download PDFInfo
- Publication number
- CN103309903A CN103309903A CN2012100703440A CN201210070344A CN103309903A CN 103309903 A CN103309903 A CN 103309903A CN 2012100703440 A CN2012100703440 A CN 2012100703440A CN 201210070344 A CN201210070344 A CN 201210070344A CN 103309903 A CN103309903 A CN 103309903A
- Authority
- CN
- China
- Prior art keywords
- search
- node
- location matches
- data
- user
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a position search system based on cloud computing, which is used for providing automatic search for dynamic or static information based on present positions, and adopts a real-time distributed computing framework. A module of the existing system server side is divided into 1-N computing node(s); the degree of parallelism in an original position matching module and a search service module is increased, so that a plurality of parallel computing nodes are further divided; the connection relationship among the computing nodes is determined so as to form a distributed computing topology; the distributed computing topology is deployed to a server cluster according to a strategy to execute continuously; and each computing node is parallelly executed through 1-M thread(s) or process(es). The invention also discloses a position search method based on cloud computing. Through using the position search system and the position search method, the problem of inadequate mass data processing capability in the prior art is solved, the instantaneity of position search can be further guaranteed, a plurality of categories of position search can be provided simultaneously, and thus better service can be provided for people.
Description
Technical field
The present invention relates to areas of information technology, relate to a kind of location finding system and method based on cloud computing especially.
Background technology
Along with the fast development of mobile Internet, location finding is permeating gradually and is affecting daily life.Location finding also is location-based search, is use location information in search, and then influences Search Results.Existing technology and research concentrate on mostly location finding result's ordering, location information index, to the geographical marking of information or webpage and from information or webpage aspect such as extract location information.In being the Chinese patent application file of 201210031400.X, application number or the patent No. disclosed a kind of system and method for the different location-based search of realization, this scheme is separated the coupling that circulates with positional information, if the location matches result is arranged, further search for processing again, thereby with repeatedly, utmost point light weight, the search of determining replaces single, complicated, uncertain search, and positional information is brought in constant renewal in, Search Results is pushed to the user by notification service, thereby make that user's Search Results is not disposable providing, but continuous correction provides immediately according to the change in location of user and information on services, can realize based on present position, to dynamic or static information, search has further developed location-based search technique automatically.But this scheme reaches ten million or during more than one hundred million other mass datas of level, processing power is not enough to some extent, causes instantaneity to descend to some extent the user.Be necessary it is improved, improve the processing power to mass data, thereby provide the location finding service for people better.
On the other hand, cloud computing (Cloud Computing) is starting the revolution again of scientific and technological industry.Cloud computing is the product that grid computing (Grid Computing), Distributed Calculation (Distributed Computing), parallel computation (Parallel Computing), effectiveness calculating (Utility Computing), the network storage (Network Storage Technologies), virtual (Virtualization), load balancing traditional computers such as (Load Balance) and network technical development merge.It is integrated into a system with supercomputing capability to the relatively low computational entity of a plurality of costs by network, and namely serves (IaaS), platform by infrastructure and namely serve (PaaS), software and namely serve the business model of innovations such as (SaaS) supercomputing capability is delivered in terminal user's hand.Be necessary location finding is combined with cloud computing, take full advantage of resource, thereby provide service for people better.
There is very big challenge but how above-mentioned location-based search system and method to take full advantage of cloud computing, it is distributed how to realize, how to calculate, how to increase degree of parallelism etc. in real time.
Summary of the invention
In view of this, the purpose of this invention is to provide a kind of location finding system and method based on cloud computing, it can take full advantage of cloud computing, solve existing location finding system to the problem of mass data processing scarce capacity, further ensure the instantaneity of location finding, provide the location finding service for people better.
For achieving the above object, the present invention adopts following technical scheme:
A kind of location finding system based on cloud computing is used for providing the automatic search to dynamic or static information based on present position, and it adopts real-time distributed computing architecture, and is realized by software, firmware, hardware or its combination.
As a further improvement on the present invention, described real-time distributed computing architecture is:
The module of existing system service end is split into 1 ~ N computing node, degree of parallelism in increase original position matching module, the search service module is further to be divided into a plurality of parallel computation nodes, determine the annexation between each computing node, form a Distributed Calculation topology; And according to strategy described Distributed Calculation topology is deployed on the server cluster continuously and carries out, each computing node is by 1 ~ M thread or process executed in parallel.
As a further improvement on the present invention, described real-time distributed calculating topology specifically comprises a following N computing node:
1 search operation input node, what be used for receiving the user opens, stops, upgrades the searching request data;
1 position input node, the position that is used for receiving the user is new data more;
1 matching range input node, the matching range that is used for receiving the user is new data more;
1 position pre-service node is used for positional data and carries out pre-service;
1 ~ N1 location matches and new node more are used for carrying out the location matches based on internal memory, adopt multithreading shared drive mode; Increase and delete the location matches of relative users; The position is upgraded and matching range upgrades;
1 ~ N2 location matches be merge node as a result, is used for by user ID merger location matches result data;
1 ~ N3 routine search node is used for pre-search, ordering, buffer memory and removing buffer memory; By user ID pre-search ranking results data and location matches result data are done to occur simultaneously and obtain search result data;
1 ~ N4 Search Results merge node is used for by user ID merge sort search result data;
1 Search Results notice pushes node, is used for pushing search result data and gives the user;
Wherein, search operation input node is connected to each routine search node, search operation input node is connected to position pre-service node, position input node is connected to position pre-service node, position pre-service node is connected to each location matches and new node more, matching range input node is connected to each location matches and new node more, location matches and more new node be connected to corresponding location matches merge node as a result, location matches merge node as a result is connected to corresponding routine search node, the routine search node is connected to corresponding Search Results merge node, and each Search Results merge node is connected to Search Results notice sending node; N is the integer greater than 9, and N1, N2, N3, N4 are the integer between 1 to N-9.
As a further improvement on the present invention, described 1 ~ N1 location matches and more new node divide according to classification for search and position subregion, each node adopts multithreading shared drive mode executing location coupling; Described 1 ~ N2 location matches merge node is as a result divided according to classification for search and position Da Qu; Described 1 ~ N3 routine search node is divided according to classification for search; Described 1 ~ N4 Search Results merge node divided according to the big class of search;
Wherein, location matches and the new node location matches merge node as a result that is connected to same classification for search and position Da Qu more, location matches merge node as a result is connected to the routine search node of same classification for search, and the routine search node is connected to the Search Results merge node of the big class of same search; The location matches data that will import into of merge node and Search Results merge node as a result adds timestamp and buffer memory, time of arrival threshold values namely by user ID and corresponding data recently or minimum time stab and carry out merger; Classification for search is to searching for the further segmentation of big class, and the position subregion is the further segmentation to position Da Qu.
As a further improvement on the present invention, described server cluster is that physical server cluster or cloud infrastructure are the Virtual Server Cluster on the service platform.
A kind of position search method based on cloud computing is used for providing the automatic search to dynamic or static information based on present position, and it is based on real-time distributed calculating.
As a further improvement on the present invention, described real-time distributed computing method are:
The step of existing system service end is split into 1 ~ N step, increase original position coupling, conventional pre-search and do degree of parallelism in the common factor step further to be divided into the step of a plurality of executed in parallel, data stream is carried out orientation, transfers order, the processing of filling a vacancy, carry out in the Distributed Calculation topology; And according to strategy described Distributed Calculation topology is deployed on the server cluster continuously and carries out, each computing node is by 1 ~ M thread or process executed in parallel.
As a further improvement on the present invention, the described step of carrying out in the Distributed Calculation topology specifically comprises:
A. search operation input node receives user's startup searching request data, is sent to corresponding routine search node and carries out pre-search, ordering and buffer memory; Simultaneously, be sent to the pre-service of position pre-service node executing location, the data after pre-service node in position will be handled again be sent to corresponding location matches and more new node increase relative users and carry out location matches based on internal memory at this node;
B. the position input node position new data more that receives relative users is sent to position pre-service node and carries out pre-service, the position data after pre-service node in position will be handled again be sent to corresponding location matches and more new node carry out the renewal of corresponding position;
C. the matching range input node matching range new data more that receives relative users, be sent to corresponding location matches and more new node carry out corresponding matching range renewal;
D. search operation input node receives the renewal searching request data of relative users, is sent to corresponding routine search node and re-executes pre-search, ordering and buffer memory;
E. location matches and more new node employing multithreading shared drive mode executing location coupling, with the location matches result data be sent to corresponding location matches as a result merge node carry out merger by user ID, location matches merge node as a result is sent to the result data of merger corresponding routine search node again and does common factor by user ID and the pre-search ranking results data of corresponding buffer memory, and the result data that the routine search node will be done common factor again is sent to corresponding Search Results merge node and carries out merger by user ID and obtain search result data;
F. the Search Results merge node is sent to the Search Results notice with search result data and pushes node, and the Search Results notice pushes node search result data is pushed to relative users;
G. search operation input node receives the service data that stops search of relative users, is sent to corresponding routine search node and removes corresponding pre-search buffer memory; Simultaneously, be sent to corresponding location matches and renewal knot removal relative users in the location matches of this node;
Wherein, step b, c, d are not necessary steps, and other step between step a ~ g there is no the sequencing on the logical level.
As a further improvement on the present invention, described step e specifically comprises step:
E1. the user data of participating in location matches is sent to corresponding location matches and new node more according to classification for search and position subregion, adopts multithreading shared drive mode executing location coupling to obtain the location matches result data;
E2. the location matches result data location matches merge node as a result that is sent to same classification for search and position Da Qu, add timestamp and buffer memory, time of arrival threshold values namely by user ID and corresponding data recently or minimum time stab and carry out the merger big zone position matching result data that obtain classifying;
E3. the routine search node that the big zone position matching result data of classifying are sent to same classification for search is done to occur simultaneously by user ID and the pre-search ranking results data of corresponding buffer memory and is obtained classifying big area searching result data;
E4. the big area searching result data of classifying is sent to the Search Results merge node of the big class of same search, add timestamp and buffer memory, time of arrival, threshold values namely obtained the big area searching result data of big class by the nearest of user ID and corresponding data or minimum time stamp execution merger;
Wherein, classification for search is to searching for the further segmentation of big class, and the position subregion is the further segmentation to position Da Qu.
As a further improvement on the present invention, described server cluster is that physical server cluster or cloud infrastructure are the Virtual Server Cluster on the service platform.
By technical scheme of the present invention as can be seen, use for reference distributed streaming and calculate thought, the location finding service end module of existing system is split into relatively independent computing node, determine the annexation between each computing node, to data stream carry out orientation, transfer order, processing such as fill a vacancy, carry out real-time distributed calculating and become possibility thereby make; Further use for reference Map/Reduce thought, the location matches service module of existing system is divided into a plurality of parallel computation nodes by classification for search and/or position subregion, the search service module of existing system is divided into a plurality of parallel computation nodes by classification for search or big class, and with each computing node by a plurality of threads or process executed in parallel, thereby take full advantage of the computation capability of server cluster, significantly improved the processing power to mass data.Thereby the present invention takes full advantage of cloud computing, has solved the problem of prior art to the mass data processing scarce capacity, has further ensured the instantaneity of location finding, provides the location finding service for people better.
Another beneficial effect of the present invention is, owing to divide and parallel computation by the big class of search, the automatic search service to dynamic or static information based on present position of a plurality of classifications can be provided for the user easily simultaneously, for example: user's location finding friend-making simultaneously, bank, purchase by group etc., form a location finding request list and be placed on client background, thereby bring more convenience for the user.
Another beneficial effect of the present invention is, strengthened the extensibility of location finding system, only needs further to increase computing node degree of parallelism and/or server cluster quantity, can further improve the mass data processing ability; Strengthened robustness and the fault-tolerance of location finding system, Single Point of Faliure or failure can not influence total system, thereby have further ensured the location finding service.
Description of drawings
Fig. 1 is the structural representation of first embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention;
Fig. 2 is the structural representation of second embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention;
Fig. 3 is the structural representation of the 3rd embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention;
Fig. 4 is the structural representation of the 4th embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention;
Fig. 5 is the structural representation of the 5th embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention;
Fig. 6 is the structural representation of the 6th embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention;
Fig. 7 is the structural representation of the 7th embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention;
Fig. 8 is the schematic flow sheet of first embodiment of the position search method based on cloud computing of the present invention.
Embodiment
In order to make the purpose, technical solutions and advantages of the present invention clearer, describe the present invention below in conjunction with the drawings and specific embodiments.
According to the location finding system based on cloud computing of the present invention, adopt distributed computing architecture, in the present embodiment, utilize the improved real-time distributed computing platform Storm that increases income.Storm provides one group of generic primitives for distributed real-time calculates, and can be used among " stream is handled " processing messages and more new database in real time; Also can be used to " calculating continuously ", data stream is done continuous-query, when calculating, just the result be exported to the user with the form of stream; Also can be used to " distributed RPC ", move expensive computing in parallel mode.The Storm cluster is made up of a host node and a plurality of working node.Host node has moved the finger daemon of " Nimbus " by name, is similar to the JobTracker of Hadoop, is used for allocation of codes, assigns a task and fault detect.Each working node has moved the finger daemon of " Supervisor " by name, is used for monitoring work, and beginning also stops the progress of work.Nimbus and Supervisor can both fail fast, and are stateless, and this makes them very healthy and strong, and both co-ordinations are finished by Zookeeper.The generic primitives of Storm comprises Stream, Tuple, Spout, Bolt, Task, Worker, Stream Grouping and Topology.Stream is processed data, is made up of Tuple stream, and Spout is data source, the Bolt deal with data, and Task is the thread that runs among Spout or the Bolt, Worker is the process of these threads of operation.Stream Grouping has stipulated that what Bolt receives as the input data.Data can Random assignment (ShuffleGrouping), distribute (FieldsGrouping), broadcasting (AllGrouping) according to field value, always issues a Task(GlobalGrouping), also can be indifferent to these data (NoneGrouping) or decide (DirectGrouping) by self-defined logic.Topology is Spout and the Bolt meshed network that is coupled together by Stream Grouping.
Carry out real-time distributed calculating at Strom, at first create Topology, be distributed to each working node by Nimbus then, on server cluster, carry out continuously.Described server cluster can be the physical server cluster, also can be that cloud infrastructure is namely served the Virtual Server Cluster on (IaaS) platform, the latter is because virtual with hardware resource, can sacrifice certain performance, what bring is low cost and high scalability, in the practice, accept or reject according to actual needs.
Fig. 1 is the structural representation of first embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention, and on Strom, the Distributed Calculation topology is Topology.
As shown in Figure 1, the Distributed Calculation topology based on the location finding system of cloud computing comprises:
1 search operation input node 11,1 position input node 12,1 matching range input node 13,1 position pre-service node 14,16 location matches and 1501 ~ 1516,8 location matches of new node, 161 ~ 168,4 routine search nodes 171 ~ 174 of merge node as a result more, 181 ~ 182,1 Search Results notice of 2 Search Results merge nodes pushes node 19;
Wherein, search operation input node 11 is connected to each routine search node 171～174, search operation input node 11 is connected to position pre-service node 14, position input node 12 is connected to position pre-service node 14, position pre-service node 14 is connected to each location matches and new node 1501～1516 more, matching range input node 13 is connected to each location matches and new node 1501～1516 more, location matches and the new node location matches merge node as a result that is connected to same classification for search and position Da Qu more, 1501～1502 are connected to 161,1503～1504 are connected to 162,1515～1516 are connected to 168, location matches merge node as a result is connected to the routine search node of same classification for search, 161～162 are connected to 171,163～164 are connected to 172,167～168 are connected to 174, the routine search node is connected to the Search Results merge node of the big class of same search, 171～172 are connected to 181,173～714 are connected to 182, and each Search Results merge node 181～182 is connected to the Search Results notice and pushes node 19;
X represents to divide by search category, and classification for search is to searching for the further segmentation of big class, be generally some relevant search categories, in the present embodiment, suppose to have two big classes of search, respectively comprising two classifications for search; Y represents that the opsition dependent zone is divided, the position subregion is the further segmentation to position Da Qu, is generally several continuous bands of position, in the present embodiment, supposes to have two position Da Qu, respectively comprises two position subregions.
Search operation input node 11, what be used for receiving the user opens, stops, upgrades the search operation data.Announcement service module through existing system transmits if open, stop, upgrade the search operation data and be, can with the announcement service module of existing system in location finding service interface module integrated, that monitors that communication connects opens, stops, upgrades search request message.On Storm, start and stop search input Spout11 adopts 4 Task; The user data that receives and produce comprises user ID, the big class of search and position Da Qu, search operation data, position data, matching range are upgraded in start and stop; Output Tuple1 comprises user ID, position Da Qu, search condition, start and stop search operation data; Output Tuple2 comprises user ID, the big class of search and position Da Qu, start and stop search operation data, position data, matching range; Output Tuple3 comprises user ID, position Da Qu, search condition, renewal search operation data.
Matching range input node 13, the matching range that is used for receiving the user is new data more.If matching range more new data is to transmit through the announcement service module of existing system, can with the announcement service module of existing system in location finding service interface module integrated, monitor the matching range updating message that communication connects.On Strom, matching range input Spout13 adopts 4 Task; The user data that receives and produce comprises more new data of user ID, the big class of search and position Da Qu, matching range; Output Tuple comprises more new data of user ID, matching range.
Position pre-service node 14 is used for positional data and carries out pre-service, comprises according to the user and permits positional data to store.Owing to be frequent operation, need to adopt a plurality of threads or process.On Strom, pre-service Bolt14 in position adopts 10 task; Input Tuple1 comprises user ID, the big class of search and position Da Qu, matching range, position data; Input Tuple2 comprises more new data of user ID, the big class of search and position Da Qu, position; Output Tuple comprises the position data after user ID, the processing.
Location matches and new node 1501 ~ 1516 more are used for carrying out the location matches based on internal memory; Increase and delete the location matches of relative users; Executing location is upgraded and matching range upgrades.Location matches and renewal are the cores of described system, also be to the highest part of execution speed requirement, be necessary on the basis based on internal memory operation of prior art, to be improved, use for reference Map/Reduce thought, be divided into a plurality of location matches and new node more, the location finding division of different classifications for search is come, and the location finding of same classification for search is divided into a plurality of positions subregion, to alleviate the load of individual node, and make each node executed in parallel, thereby improve described entire system execution speed, for example: life-food and drink-Beijing-Chaoyang District (X11Y11), life-food and drink-Beijing-Dongcheng District (X11Y12), life-food and drink-Shanghai-Huangpu District (X11Y21), life-food and drink-Shanghai-Xuhui District (X11Y22), life-shopping-Beijing-Chaoyang District (X12Y11), social activity-friend-making-Shanghai-Xuhui District (X22Y22).Further improve, in the location matches of same big class same position subregion, adopt J thread parallel to carry out, and shared drive, thread I is according to JR+I user's of strategy circulation execution location matches, wherein, R is integer, and J is the integer greater than 1, and I is the integer of 1～J.Further improve again, each location matches and more new node selection receives only this position subregion and adds location matches with the user who faces the position subregion mutually according to user's matching range, thereby further lighten the load.On Strom, be divided into 16 Bolt according to classification for search and position subregion; Each location matches and renewal Bolt adopt 8 Task; Input Tuple comprises the position data after user ID, the processing; Output Tuple comprises user ID, position subregion, classification district location matching result.
Location matches is merge node 161 ~ 168 as a result, be used for the location matches result of each position subregion of same classification for search is carried out merger by user ID, correspondingly, need to be divided into a plurality of location matches merge node as a result by the big zoning of classification for search and position, each node executed in parallel, for example: life-food and drink-Beijing (X11Y1), life-food and drink-Shanghai (X11Y2), life-shopping-Beijing (X12Y1) ... social activity-friend-making-Shanghai (X22Y2).Because handling the result of same user's location matches in once circulating, the location matches node of each position subregion of same classification for search might have or not have, the time of handling same user has successively, the order of handling different user also may be different, the disappearance that causes data stream, order changes, can take following dual mode to solve: to set a time threshold values, this threshold values can with single location matches and the more single cycle time correlation of single thread in the new node, when receiving certain user's classification district location matching result, namely add timestamp and buffer memory, when time of arrival threshold values namely carry out merger by what this user ID and this user respectively classified the district location matching result with nearest timestamp of current time, remove and used and expired time stamp data, the time threshold values of resetting; Perhaps, when receiving certain user's classification district location matching result for the first time, for this user sets a time threshold values, this threshold values can with single location matches and the more single cycle time correlation of single thread in the new node, and with this result and other classification district location matching result of this user who arrives subsequently add timestamp and buffer memory, when time of arrival threshold values namely stab and carry out merger by respectively the classify minimum time of district location matching result of this user ID and this user, remove the time stamp data that has used, this user's that resets time threshold values.On Strom, correspondingly, be divided into 8 Bolt according to the big zoning of classification for search and position; Each position matching result merger Bolt adopts 6 Task, adopts FieldsGrouping(ID with corresponding location matches with renewal Node B olt) be connected to determine the Task of reception; Input Tuple comprises user ID, position subregion, classification district location matching result; Output Tuple comprises user ID, the big zone position matching result of classification.
Search Results merge node 181 ~ 182, be used for the Search Results of each classification for search is carried out merger by user ID, in fact, a user only may be in a position Da Qu in a period of time, therefore needn't carry out merger by opsition dependent Da Qu, correspondingly, need to be divided into a plurality of Search Results merge nodes by the big class of search, each node executed in parallel, for example: life (X1), social (X2).Because the routine search node of each classification for search might have or not have same user's the result who does common factor, the time of handling same user has successively, the order of handling different user also may be different, and the reason of aforementioned location matched node, the disappearance that causes data stream, order changes, can take following dual mode to solve: to set a time threshold values, this threshold values can with single location matches and the more single cycle time correlation of single thread in the new node, when the big area searching of the classification that receives certain user as a result the time, namely add timestamp and buffer memory, when time of arrival threshold values namely carry out merger by what this user ID and this user respectively classified big area searching result with nearest timestamp of current time, remove and used and expired time stamp data, the time threshold values of resetting; Perhaps, when the big area searching of classification that for the first time receives certain user as a result the time, for this user sets a time threshold values, this threshold values can with single location matches and the more single cycle time correlation of single thread in the new node, and with this result and other big area searching result that classifies of this user who arrives subsequently add timestamp and buffer memory, when time of arrival threshold values namely stab and carry out merger by respectively classify big area searching result's minimum time of this user ID and this user, remove the time stamp data that has used, this user's that resets time threshold values.On Strom, correspondingly, be divided into 2 Bolt according to the big class of search; Each Search Results merger Bolt adopts 6 Task, adopts FieldsGrouping(ID with corresponding routine search Node B olt) be connected to determine the Task of reception; Input Tuple comprises user ID, the big area searching result of classification; Output Tuple comprises user ID, searches for big class, the big big area searching result of class.
Search Results notice pushes node 19, is used for the location finding service interface module of the announcement service module of existing system integratedly, pushes Search Results to the user and notifies by being connected with the relative client module communication of having set up in the existing system.Because this is to operate very frequently, can use a plurality of threads or process, to increase data throughout.On Strom, the Search Results notice pushes Bolt19 and adopts 20 Task; Input Tuple comprises user ID, the big big area searching result of class; Be pushed to user's data and comprise the big class of search, big class Search Results.
Fig. 2 is the structural representation of second embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention, it is a simple Distributed Calculation topology, only with location matches and more new node 251 ~ 252 opsition dependent Da Qu divide, location matches is distinguished in the zone that obviously isolates, to alleviate the load of individual node, and make each node executed in parallel, for example: Beijing (Y1), Shanghai (Y2), and do not divide by the big class of search and position subregion, therefore just do not need corresponding merge node yet, namely only carry out the search of a certain class, for example: make friends, other node is identical with first embodiment, repeats no more.It is applicable to the location finding in small-sized vertical field.
Fig. 3 is the structural representation of the 3rd embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention, it is another simple Distributed Calculation topology, only routine search node 361 ~ 362 is divided by the big class of search, to realize the location finding of a plurality of big classes simultaneously, for example: food and drink (X1), friend-making (X2), and do not divide by classification for search and position Da Qu, therefore just do not need corresponding merge node yet, other node is identical with first embodiment, repeats no more.It is applicable to the location finding in small-sized single zone.
Fig. 4 is the structural representation of the 4th embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention, wherein, location matches and new node 451 ~ 454 more, routine search node 461 ~ 462 is further divided by the big class of search on the basis of second embodiment, with location matches, pre-search is distinguished in visibly different search category with doing to occur simultaneously, to realize the location finding of the big class of a plurality of search simultaneously, and alleviate the load of individual node, make each node executed in parallel, for example: food and drink-Beijing (X1Y1), food and drink-Shanghai (X1Y2), friend-making-Beijing (X2Y1), friend-making-Shanghai (X2Y2), but do not divide by classification for search and position subregion, therefore just do not need corresponding merge node yet, other node is identical with first embodiment, repeats no more.It is applicable to small-sized location finding.
Fig. 5 is the structural representation of the 5th embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention, wherein, location matches and more new node 551 ~ 554 opsition dependent subregion on the basis of the 3rd embodiment divide, to alleviate the load of individual node, and make each node executed in parallel, for example: Beijing-Chaoyang District (Y11), Beijing-Dongcheng District (Y12), Shanghai-Huangpu District (Y21), Shanghai-Xuhui District (X1Y22), correspondingly, increase location matches 561 ~ 562 pairs of district location matching results execution of merge node as a result merger, other node is identical with first embodiment, repeats no more.It is applicable to medium-sized location finding.
Fig. 6 is the structural representation of the 6th embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention, wherein, location matches and more new node 651 ~ 658 on the basis of the 4th embodiment, further be divided into a plurality of positions subregion, further to alleviate the load of individual node, and make each node executed in parallel, for example: food and drink-Beijing-Chaoyang District (X1Y11), food and drink-Beijing-Dongcheng District (X1Y12), food and drink-Shanghai-Huangpu District (X1Y21), food and drink-Shanghai-Xuhui District (X1Y22), friend-making-Shanghai-Huangpu District (X2Y22), friend-making-Shanghai-Xuhui District (X2Y22), correspondingly, increase location matches 661 ~ 664 pairs of big class district location matching results execution merger of merge node as a result, other node is identical with first embodiment, repeats no more.It is applicable to medium-sized location finding.
Fig. 7 is the structural representation of the 7th embodiment of the Distributed Calculation topology of the location finding system based on cloud computing of the present invention, wherein, location matches and more new node 7501 ~ 7516 on the basis of the 6th embodiment, further divide by classification for search, further to alleviate the load of individual node, and make each node executed in parallel, for example: life-food and drink-Beijing-Chaoyang District (X11Y11), life-food and drink-Beijing-Dongcheng District (X11Y12), life-food and drink-Shanghai-Huangpu District (X11Y21), life-food and drink-Shanghai-Xuhui District (X11Y22), life-shopping-Beijing-Chaoyang District (X12Y11), social activity-friend-making-Shanghai-Xuhui District (X22Y22).Correspondingly, location matches 761 ~ 764 pairs of classification of merge node as a result district location matching result is carried out merger, and other node is identical with first embodiment, repeats no more.It is applicable to large-scale location finding.The difference of itself and first embodiment is, the former is merger-do common factor, and the latter is part merger-do common factor-merger, and obviously the latter's degree of parallelism is higher, thereby stronger to the processing power of mass data.
Certainly, the Distributed Calculation topology can also be according to the purposes of location finding system, scale, actual conditions etc., press further segmentation such as search category, the band of position, further increase and decrease, ordering, combination computing node, for thread or the process that computing node distributes varying number, exhaustive no longer one by one here; Distributed computing architecture can also utilize other real-time distributed computing platform or distributed streaming computing platform, for example: real-time Hadoop, S4, Streambase etc., or improved platform in addition, even software can also be converted into corresponding firmware, hardware or its combination and be realized, owing to be prior art, repeat no more here.
Fig. 8 is the schematic flow sheet of first embodiment of the position search method based on cloud computing of the present invention, and its first embodiment in the Distributed Calculation topology of the location finding system based on cloud computing of the present invention is carried out.Here be described in detail in conjunction with concrete example " user A wishes to find near the people of own 100 meters hobby tourism and near 200 meters cafe or supermarket at the street in Shanghai City ".
As shown in Figure 8, the position search method based on cloud computing comprises the steps:
Step 871 ~ 874 executed in parallel, the big class X1 of " life " search, each routine search node of the big class X2 of " social activity " search walk abreast user A are carried out conventional pre-search.Wherein, " food and drink " classification for search X11 routine search node carries out conventional pre-search to user A " K(cafe or supermarket) " according to " ON " in corresponding category index storehouse, search " Shanghai City " position Da Qu Y2 cafe ranking results collection " ID(A); RS(X11Y2) ", with its buffer memory, " shopping " classification for search X12 routine search node searching to the supermarket ranking results collection of " Shanghai City " position Da Qu Y2 " ID(A); RS(X12Y2) ", with its buffer memory, other routine search node searching of the big class X1 of " life " search result is empty; " friend-making " classification for search X22 routine search node carries out conventional pre-search to user A " people of K(hobby tourism) " according to " ON " in corresponding category index storehouse, search the hobby tourism of " Shanghai City " position Da Qu Y2 people's ranking results collection " ID(A); RS(X22Y2) ", with its buffer memory, other routine search node searching of the big class X2 of " social activity " search result is empty.
Simultaneously, step 84, position pre-service node is to L(T) carry out validity checking, coding, conversion, pre-service such as store according to user's permission, according to " Y2 " to " life " search big class X1, " social activity " search big class X2 " Shanghai City " position each location matches of Da Qu Y2 and more new node transmission " ID(A); PL(T), ON ".
Step 851 ~ 858 executed in parallel, " life " search big class X1, " social activity " search big class X2 " Shanghai City " position each location matches of Da Qu Y2 and more new node walk abreast in internal memory to circulate according to strategy and carry out location matches.Wherein, the X11Y21 location matches and more new node increase user A according to " ON " and participate in location matches at this node, it is user A that an one thread I is recycled to 8R+I, and wherein, R is integer, I is 1～8 integer, will " PL(T) D (100) " with this node in all users' " PL D (100) " mate, suppose has the result, the district location matching result collection that obtains classifying " ID(A); RL(X11Y21) ", to the merge node transmission as a result of X11Y2 location matches; The big class X1 of " life " search, " social activity " search big class X2 " Shanghai City " position other location matches of Da Qu Y2 and the more same operation of new node execution, obtain corresponding classification district location matching result collection, transmit to the big class X1 of " life " search, " social activity " search Da Qu Y2 relevant position, big class X2 " Shanghai City " position matching result merge node, if result set is sky then does not transmit.
Step 861 ~ 864 executed in parallel, the big class X1 of " life " search, each position matching result merge node of " social activity " search big class X2 " Shanghai City " position Da Qu Y2 walk abreast and carry out location matches merger as a result.Wherein, the X11Y2 location matches is merge node setting-up time threshold values as a result, when receive " ID(A); RL(X11Y21) ", namely add timestamp obtain " ID(A); RL(X11Y21TS1) ", " ID(A); RL(X11Y21TS2) ", and buffer memory, same other classification district location matching result collection of handling the user A that receives, when time of arrival threshold values namely to user A " RL(X11Y21NEAR(TS)); RL(X11Y22NEAR(TS)); ... " carry out merger obtain classifying big zone position matching result collection " ID(A); RML(X11Y2) ", transmit to X11Y1/Y2 routine search node, remove and used and expired time stamp data, the time threshold values of resetting; The big class X1 of " life " search, " social activity " search big class X2 " Shanghai City " position other location matches of Da Qu Y2 merge node are as a result carried out same operation, obtain the big zone position matching result of corresponding classification collection, transmit to the big class X1 of " life " search, the big class X2 corresponding conventional search node of " social activity " search, if result set is sky then does not transmit.
Step 871 ~ 874 executed in parallel, the big class X1 of " life " search, the parallel common factor of doing of each routine search node of the big class X2 of " social activity " search.
Wherein, " food and drink " classification for search X11 routine search node when receive " ID(A); RML(X11Y2) ", namely with buffer memory " ID(A); RS(X11Y2) " do common factor, supposing has the result, the big area searching result that obtains classifying " ID(A); RMSL(X11Y2) ", transmit to the big class X1 Search Results merge node of " life " search; The big class X1 of " life " search, other routine search node of the big class X2 of " social activity " search are carried out same operation, obtain the big area searching result of corresponding classification, transmit to the big class X1 of " life " search, the corresponding Search Results merge node of the big class X2 of " social activity " search, if result set is sky then does not transmit.
Step 881 ~ 882 executed in parallel, the big class X1 of " life " search, the big class X2 Search Results merge node of " social activity " search walk abreast and carry out the Search Results merger.The big class X1 Search Results merge node setting-up time threshold values of " life " search, when receive " ID(A); RMSL(X11Y2) ", namely add time stamp T obtain " ID(A); RMSL(X11Y2TS1) ", " ID(A); RMSL(X11Y2TS2) ", and buffer memory, same other big area searching result that classifies who handles the user A that receives, when time of arrival threshold values namely to user A " RMSL(X11Y2NEAR(TS)); RMSL(X12Y2NEAR(TS)); ... " carry out merger obtain the big area searching result of big class " ID(A); RMMSL(X1Y2) ", pushing node to the Search Results notice transmits, remove and used and expired time stamp data, the time threshold values of resetting; The big class X2 Search Results merge node of " social activity " search is carried out same operation, obtains the corresponding big big area searching result of class, pushes node to the Search Results notice and transmits.
Location matches is carried out continuously.Step 851 ~ 858 executed in parallel, " Shanghai City " position each location matches of Da Qu Y2 and more the new node circulation that walks abreast in internal memory according to strategy carry out location matches, wherein, the X11Y21 location matches and more in the new node thread I ' to be circulated again into 8R+I ' be user A, wherein, R is integer, I ' is 1～8 integer, will " PL(T) D (100) " again with this node in all users' " PL D (100) " mate, supposing has the result, the district location matching result collection " ID(A), RL ' is (X11Y21) " that obtains classifying, to the X11Y2 location matches as a result merge node transmit; " Shanghai City " position other location matches of Da Qu Y2 and the more same operation of new node execution.Up to step 89, Search Results notice pushes node and pushes Search Results again to user A and notify " RMMSL ' (X1) " and/or " RMMSL ' (X2) " by being connected with communication that user A client modules has been set up.Each step is the same therebetween, repeats no more.
Because user A is in walking about, the position is brought in constant renewal in.Step 82, input node in position receives the position updating request L(T ' of user A), to the position pretreatment module transmit " ID(A), L(T '), X1Y2, X2Y2, UPDATE ".Step 851 ~ 858 executed in parallel, " Shanghai City " position each location matches of Da Qu Y2 and more the new node circulation that walks abreast in internal memory according to strategy carry out location matches, wherein, X11Y21 location matches and more new node basis " UPDATE L " renewal " ID(A); L(T) " be " ID(A); L(T ') ", an one thread I " be recycled to 8R+I " be user A, wherein, R is integer; I " it is 1～8 integer, will " PL(T ') D (100) " with this node in all users' " PL D (100) " mate, suppose has the result, the district location matching result collection that obtains classifying " ID(A); RL ' ' (X11Y21) ", to the merge node transmission as a result of X11Y2 location matches; " Shanghai City " position other location matches of Da Qu Y2 and the more same operation of new node execution.Thereafter each step is the same, repeats no more.
User A becomes 200 meters to the people's scope that finds near own hobby tourism.Step 83, and the matching range update request of matching range input node reception user A " X2Y2 D(200) ", to " social activity " search big class X2 " Shanghai City " position each location matches of Da Qu Y2 and more new node transmission.Step 855 ~ 858 executed in parallel, " social activity " search big class X2 " Shanghai City " position each location matches of Da Qu Y2 and more new node walk abreast in internal memory to circulate according to strategy and carry out location matches, wherein, X22Y21 location matches and more new node basis " UPDATE D " renewal " ID(A); D(100) " be " ID(A); D(200) ", it is user A that an one thread I is recycled to 8R+I, wherein, R is integer, I is 1～8 integer, will " PL(T ') D(200) " with this node in all users' " PL D(200) " mate, suppose has the result, the district location matching result collection that obtains classifying " ID(A); RL ' (X22Y21) ", to the merge node transmission as a result of X21Y2 location matches; " social activity " search big class X1 " Shanghai City " position other location matches of Da Qu Y2 and the more same operation of new node execution.Thereafter each step is the same, repeats no more.
User A changes the people who finds near own 200 meters hobby tourisms into the people of hobby tourism and photography again.Step 81, and the search update request of search operation input node reception user A " ID(A), Y2, the people of the tourism of K(hobby and photography), UPDATE " and, to each routine search node transmission of the big class X2 of " social activity " search.Step 873 ~ 874 executed in parallel, each routine search node of the big class X2 of " social activity " search is parallel to carry out conventional pre-search again to user A, wherein, " friend-making " classification for search X22 routine search node according to " UPDATE " remove former " ID(A); RS(X22Y2) " buffer memory, in corresponding category index storehouse to " ID(A); the people of K(hobby tourism and photography) " carry out conventional pre-search, search hobby tourism and the photography of " Shanghai City " position Da Qu Y2 people's ranking results collection " ID(A); RS ' ' (X22Y2) ", with its buffer memory, other routine search node searching of the big class X2 of " social activity " search result is empty.Thereafter each step is the same, repeats no more.
User A stop search near 200 meters cafe or supermarket.Step 81, search operation input node receives the stopping search request of user A, to each routine search node transmission of the big class X1 of " life " search " ID(A); K(cafe or supermarket), OFF ", to the transmission of position pre-service node " ID(A); X1Y2, OFF ".Step 871 ~ 872 executed in parallel, the parallel user A buffer memory of removing of each routine search node of the big class X1 of " life " search, wherein, " OFF " removing of " food and drink " classification for search X11 routine search node basis " ID(A); RS(X11Y2) " buffer memory, " OFF " removing of " shopping " classification for search X12 routine search node basis " ID(A), RS(X12Y2) " buffer memory.Simultaneously, step 84, position pre-service node according to user permission to L(T ') store and wait processing, search for big class X1 " Shanghai City " position each location matches of Da Qu Y2 and more new node transmission " ID(A), OFF " according to " X1Y2 OFF " to " life ".Step 851 ~ 854 executed in parallel, " life " search big class X1 " Shanghai City " position each location matches of Da Qu Y2 and more new node walk abreast and to stop the location matches of user A, wherein, the X11Y21 location matches and more new node according to the location matches of " OFF " deletion user A at this node; " life " search big class X1 " Shanghai City " position other location matches of Da Qu Y2 and the more same operation of new node execution.Thereafter each step is the same, repeats no more.
So far, namely finish once the complete position search method based on cloud computing.
The method flow of carrying out at the method flow of carrying out based on other embodiment of the Distributed Calculation topology of the location finding system of cloud computing and at other distributed computing architecture is similar to the above, repeats no more here.
Should be appreciated that above-mentioned only is to the displaying of the present invention's spirit and principle, does not constitute improper restriction of the present invention; To those skilled in the art, can under the prerequisite of not paying creative work, be improved or conversion, and all these improvement or conversion all should be included within protection scope of the present invention.
Claims (10)
1. the location finding system based on cloud computing is used for providing the automatic search to dynamic or static information based on present position, it is characterized in that described system adopts real-time distributed computing architecture, and is realized by software, firmware, hardware or its combination.
2. the location finding system based on cloud computing according to claim 1 is characterized in that, described real-time distributed computing architecture is:
The module of existing system service end is split into 1 ~ N computing node, degree of parallelism in increase original position matching module, the search service module is further to be divided into a plurality of parallel computation nodes, determine the annexation between each computing node, form a Distributed Calculation topology; And according to strategy described Distributed Calculation topology is deployed on the server cluster continuously and carries out, each computing node is by 1 ~ M thread or process executed in parallel.
3. the location finding system based on cloud computing according to claim 2 is characterized in that, described real-time distributed calculating topology specifically comprises a following N computing node:
1 search operation input node, what be used for receiving the user opens, stops, upgrades the searching request data;
1 position input node, the position that is used for receiving the user is new data more;
1 matching range input node, the matching range that is used for receiving the user is new data more;
1 position pre-service node is used for positional data and carries out pre-service;
1 ~ N1 location matches and new node more are used for carrying out the location matches based on internal memory, adopt multithreading shared drive mode; Increase and delete the location matches of relative users; The position is upgraded and matching range upgrades;
1 ~ N2 location matches be merge node as a result, is used for by user ID merger location matches result data;
1 ~ N3 routine search node is used for pre-search, ordering, buffer memory and removing buffer memory; By user ID pre-search ranking results data and location matches result data are done to occur simultaneously and obtain search result data;
1 ~ N4 Search Results merge node is used for by user ID merge sort search result data;
1 Search Results notice pushes node, is used for pushing search result data and gives the user;
Wherein, search operation input node is connected to each routine search node, search operation input node is connected to position pre-service node, position input node is connected to position pre-service node, position pre-service node is connected to each location matches and new node more, matching range input node is connected to each location matches and new node more, location matches and more new node be connected to corresponding location matches merge node as a result, location matches merge node as a result is connected to corresponding routine search node, the routine search node is connected to corresponding Search Results merge node, and each Search Results merge node is connected to Search Results notice sending node; N is the integer greater than 9, and N1, N2, N3, N4 are the integer between 1 to N-9.
4. the location finding system based on cloud computing according to claim 3 is characterized in that, described 1 ~ N1 location matches and more new node divide according to classification for search and position subregion, each node adopts multithreading shared drive mode executing location coupling; Described 1 ~ N2 location matches merge node is as a result divided according to classification for search and position Da Qu; Described 1 ~ N3 routine search node is divided according to classification for search; Described 1 ~ N4 Search Results merge node divided according to the big class of search;
Wherein, location matches and the new node location matches merge node as a result that is connected to same classification for search and position Da Qu more, location matches merge node as a result is connected to the routine search node of same classification for search, and the routine search node is connected to the Search Results merge node of the big class of same search; The location matches data that will import into of merge node and Search Results merge node as a result adds timestamp and buffer memory, time of arrival threshold values namely by user ID and corresponding data recently or minimum time stab and carry out merger; Classification for search is to searching for the further segmentation of big class, and the position subregion is the further segmentation to position Da Qu.
5. the location finding system based on cloud computing according to claim 2 is characterized in that, described server cluster is that physical server cluster or cloud infrastructure are the Virtual Server Cluster on the service platform.
6. the position search method based on cloud computing is used for providing the automatic search to dynamic or static information based on present position, it is characterized in that described method is based on real-time distributed calculating.
7. the position search method based on cloud computing according to claim 6 is characterized in that, described real-time distributed computing method are:
The step of existing system service end is split into 1 ~ N step, increase original position coupling, conventional pre-search and do degree of parallelism in the common factor step further to be divided into the step of a plurality of executed in parallel, data stream is carried out orientation, transfers order, the processing of filling a vacancy, carry out in the Distributed Calculation topology; And according to strategy described Distributed Calculation topology is deployed on the server cluster continuously and carries out, each computing node is by 1 ~ M thread or process executed in parallel.
8. the position search method based on cloud computing according to claim 7 is characterized in that, the described step of carrying out in the Distributed Calculation topology specifically comprises:
A. search operation input node receives user's startup searching request data, is sent to corresponding routine search node and carries out pre-search, ordering and buffer memory; Simultaneously, be sent to the pre-service of position pre-service node executing location, the data after pre-service node in position will be handled again be sent to corresponding location matches and more new node increase relative users and carry out location matches based on internal memory at this node;
B. the position input node position new data more that receives relative users is sent to position pre-service node and carries out pre-service, the position data after pre-service node in position will be handled again be sent to corresponding location matches and more new node carry out the renewal of corresponding position;
C. the matching range input node matching range new data more that receives relative users, be sent to corresponding location matches and more new node carry out corresponding matching range renewal;
D. search operation input node receives the renewal searching request data of relative users, is sent to corresponding routine search node and re-executes pre-search, ordering and buffer memory;
E. location matches and more new node employing multithreading shared drive mode executing location coupling, with the location matches result data be sent to corresponding location matches as a result merge node carry out merger by user ID, location matches merge node as a result is sent to the result data of merger corresponding routine search node again and does common factor by user ID and the pre-search ranking results data of corresponding buffer memory, and the result data that the routine search node will be done common factor again is sent to corresponding Search Results merge node and carries out merger by user ID and obtain search result data;
F. the Search Results merge node is sent to the Search Results notice with search result data and pushes node, and the Search Results notice pushes node search result data is pushed to relative users;
G. search operation input node receives the service data that stops search of relative users, is sent to corresponding routine search node and removes corresponding pre-search buffer memory; Simultaneously, be sent to corresponding location matches and renewal knot removal relative users in the location matches of this node;
Wherein, step b, c, d are not necessary steps, and other step between step a ~ g there is no the sequencing on the logical level.
9. the position search method based on cloud computing according to claim 8 is characterized in that, described step e specifically comprises step:
E1. the user data of participating in location matches is sent to corresponding location matches and new node more according to classification for search and position subregion, adopts multithreading shared drive mode executing location coupling to obtain the location matches result data;
E2. the location matches result data location matches merge node as a result that is sent to same classification for search and position Da Qu, add timestamp and buffer memory, time of arrival threshold values namely by user ID and corresponding data recently or minimum time stab and carry out the merger big zone position matching result data that obtain classifying;
E3. the routine search node that the big zone position matching result data of classifying are sent to same classification for search is done to occur simultaneously by user ID and the pre-search ranking results data of corresponding buffer memory and is obtained classifying big area searching result data;
E4. the big area searching result data of classifying is sent to the Search Results merge node of the big class of same search, add timestamp and buffer memory, time of arrival, threshold values namely obtained the big area searching result data of big class by the nearest of user ID and corresponding data or minimum time stamp execution merger;
Wherein, classification for search is to searching for the further segmentation of big class, and the position subregion is the further segmentation to position Da Qu.
10. the location finding system based on cloud computing according to claim 7 is characterized in that, described server cluster is that physical server cluster or cloud infrastructure are the Virtual Server Cluster on the service platform.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100703440A CN103309903A (en) | 2012-03-16 | 2012-03-16 | Position search system and method based on cloud computing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100703440A CN103309903A (en) | 2012-03-16 | 2012-03-16 | Position search system and method based on cloud computing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103309903A true CN103309903A (en) | 2013-09-18 |
Family
ID=49135140
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012100703440A Pending CN103309903A (en) | 2012-03-16 | 2012-03-16 | Position search system and method based on cloud computing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103309903A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103699599A (en) * | 2013-12-13 | 2014-04-02 | 华中科技大学 | Message reliable processing guarantee method of real-time flow calculating frame based on Storm |
| CN103744777A (en) * | 2013-12-26 | 2014-04-23 | 浙江大学 | Detection method and purpose for detecting water content of tea by same |
| CN104916127A (en) * | 2014-03-13 | 2015-09-16 | 深圳市赛格导航科技股份有限公司 | Internet of vehicles distributed real-time traffic condition analysis method and system |
| CN105681308A (en) * | 2016-01-18 | 2016-06-15 | 中国石油大学(华东) | Attribute abstract system orienting towards real time big data platform Storm |
| CN105824618A (en) * | 2016-03-10 | 2016-08-03 | 浪潮软件集团有限公司 | Real-time message processing method for Storm |
| CN108776934A (en) * | 2018-05-15 | 2018-11-09 | 中国平安人寿保险股份有限公司 | Distributed data computational methods, device, computer equipment and readable storage medium storing program for executing |
| CN110196833A (en) * | 2018-03-22 | 2019-09-03 | 腾讯科技（深圳）有限公司 | Searching method, device, terminal and the storage medium of application program |
| CN112783922A (en) * | 2021-02-01 | 2021-05-11 | 广州海量数据库技术有限公司 | Query method and device based on relational database |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101821763A (en) * | 2007-10-19 | 2010-09-01 | 雅虎公司 | Method and apparatus for geographic specific search results including map-based display |
| CN102169507A (en) * | 2011-05-26 | 2011-08-31 | 厦门雅迅网络股份有限公司 | Distributed real-time search engine |
| CN102214221A (en) * | 2011-06-13 | 2011-10-12 | 苏州萃智新技术开发有限公司 | Location-based service body searching method |
| CN102291435A (en) * | 2011-07-15 | 2011-12-21 | 武汉大学 | Mobile information searching and knowledge discovery system based on geographic spatiotemporal data |
| CN102308184A (en) * | 2009-02-13 | 2012-01-04 | 雅虎公司 | Entity-based search results and clusters on maps |
| CN102375860A (en) * | 2010-08-27 | 2012-03-14 | 盛乐信息技术（上海）有限公司 | Cloud-based structured data search method |
-
2012
- 2012-03-16 CN CN2012100703440A patent/CN103309903A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101821763A (en) * | 2007-10-19 | 2010-09-01 | 雅虎公司 | Method and apparatus for geographic specific search results including map-based display |
| CN102308184A (en) * | 2009-02-13 | 2012-01-04 | 雅虎公司 | Entity-based search results and clusters on maps |
| CN102375860A (en) * | 2010-08-27 | 2012-03-14 | 盛乐信息技术（上海）有限公司 | Cloud-based structured data search method |
| CN102169507A (en) * | 2011-05-26 | 2011-08-31 | 厦门雅迅网络股份有限公司 | Distributed real-time search engine |
| CN102214221A (en) * | 2011-06-13 | 2011-10-12 | 苏州萃智新技术开发有限公司 | Location-based service body searching method |
| CN102291435A (en) * | 2011-07-15 | 2011-12-21 | 武汉大学 | Mobile information searching and knowledge discovery system based on geographic spatiotemporal data |
Non-Patent Citations (1)
| Title |
|---|
| 周一可: "云计算下MapReduce编程模型可用性的研究与优化", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103699599A (en) * | 2013-12-13 | 2014-04-02 | 华中科技大学 | Message reliable processing guarantee method of real-time flow calculating frame based on Storm |
| CN103699599B (en) * | 2013-12-13 | 2016-10-05 | 华中科技大学 | A kind of message reliable treatments support method based on Storm real-time streams Computational frame |
| CN103744777A (en) * | 2013-12-26 | 2014-04-23 | 浙江大学 | Detection method and purpose for detecting water content of tea by same |
| CN103744777B (en) * | 2013-12-26 | 2016-08-24 | 浙江大学 | Detection method and use this detection method detection water content of tea purposes |
| CN104916127A (en) * | 2014-03-13 | 2015-09-16 | 深圳市赛格导航科技股份有限公司 | Internet of vehicles distributed real-time traffic condition analysis method and system |
| CN105681308A (en) * | 2016-01-18 | 2016-06-15 | 中国石油大学(华东) | Attribute abstract system orienting towards real time big data platform Storm |
| CN105681308B (en) * | 2016-01-18 | 2019-05-17 | 青岛邃智信息科技有限公司 | A kind of attribute extraction system towards real-time big data platform Storm |
| CN105824618A (en) * | 2016-03-10 | 2016-08-03 | 浪潮软件集团有限公司 | Real-time message processing method for Storm |
| CN110196833A (en) * | 2018-03-22 | 2019-09-03 | 腾讯科技（深圳）有限公司 | Searching method, device, terminal and the storage medium of application program |
| CN108776934A (en) * | 2018-05-15 | 2018-11-09 | 中国平安人寿保险股份有限公司 | Distributed data computational methods, device, computer equipment and readable storage medium storing program for executing |
| CN112783922A (en) * | 2021-02-01 | 2021-05-11 | 广州海量数据库技术有限公司 | Query method and device based on relational database |
| CN112783922B (en) * | 2021-02-01 | 2022-02-25 | 广州海量数据库技术有限公司 | Query method and device based on relational database |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103309903A (en) | Position search system and method based on cloud computing | |
| US20210279098A1 (en) | File Operation Task Optimization | |
| Tang et al. | An offloading method using decentralized P2P-enabled mobile edge servers in edge computing | |
| CN102546402B (en) | Method for supporting distributed key-based processes and computing entity | |
| CN108469988A (en) | A kind of method for scheduling task based on isomery Hadoop clusters | |
| EP3053036B1 (en) | Dynamic shuffle reconfiguration | |
| Zhang et al. | A social community detection algorithm based on parallel grey label propagation | |
| CN103942308A (en) | Method and device for detecting large-scale social network communities | |
| CN105635283A (en) | Organization and management and using method and system for cloud manufacturing service | |
| US20220261392A1 (en) | Persistent Shuffle System | |
| CN104104621B (en) | A kind of virtual network resource dynamic self-adapting adjusting method based on Nonlinear Dimension Reduction | |
| CN102929961A (en) | Data processing method and device thereof based on building quick data staging channel | |
| CN104063501B (en) | copy balance method based on HDFS | |
| CN104809130A (en) | Method, equipment and system for data query | |
| CN104375882A (en) | Multistage nested data drive calculation method matched with high-performance computer structure | |
| US9794138B2 (en) | Computer system, method, and program | |
| Isoyama et al. | A scalable complex event processing system and evaluations of its performance | |
| CN104750718A (en) | Data information search method and data information search device | |
| Jangiti et al. | Scalable hybrid and ensemble heuristics for economic virtual resource allocation in cloud and fog cyber-physical systems | |
| US10572486B2 (en) | Data communication in a distributed data grid | |
| Rashid et al. | Edgestore: Towards an edge-based distributed storage system for emergency response | |
| WO2023093255A1 (en) | Clustering collection determining method and device, storage medium, and electronic device | |
| CN110505276A (en) | Object matching method, apparatus and system, electronic equipment and storage medium | |
| US11768814B2 (en) | Data transmissions between two databases | |
| CN109254844A (en) | A kind of triangle calculation method of Large Scale Graphs |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20130918 |
|
| RJ01 | Rejection of invention patent application after publication |