CN105447187B - Web search method and system - Google Patents
Web search method and system Download PDFInfo
- Publication number
- CN105447187B CN105447187B CN201510945454.0A CN201510945454A CN105447187B CN 105447187 B CN105447187 B CN 105447187B CN 201510945454 A CN201510945454 A CN 201510945454A CN 105447187 B CN105447187 B CN 105447187B
- Authority
- CN
- China
- Prior art keywords
- query word
- real
- search
- default
- node
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
- G06F16/972—Access to data in other repository systems, e.g. legacy data or dynamic Web page generation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention is on a kind of web search method and system, it is in the case where the real-time traffic of web page search system is increased to that the preset search result corresponding with target query word is not present in overload level and caching system, the retrieval node for choosing less number under relative normal discharge performs real-time search mission to target query word, the purpose for the load for reducing each retrieval node is reached by sacrificing the real-time search result quantity returned for each target query word, the response speed of whole system is improved.In addition, when real-time traffic reaches disaster tolerance rank, on the one hand logical too small amount of retrieval node performs real-time search mission to target query word, obtain real-time search result, on the other hand it is used as compensation by obtaining preset search result from caching system, both the load of retrieval node can be reduced, the integrality of search result can be ensured again.
Description
One, technical fields
The present invention relates to search engine optimization technical field, more particularly to a kind of web search method and system.
Two, background technologies
With the development and popularization of internet, increasing user carries out net by the browser in various terminal equipment
Page search, to obtain the information of oneself needs.Fig. 1 shows the Organization Chart of web page search system;The web page search system includes
Router (Router) 101, caching (Cache) system 102 and retrieval node (Searcher) array 103, wherein, retrieve node
(Searcher) comprising the common M*N retrieval node of M rows N row in array 103, router 101 is maintained and each retrieval node all the time
Link;In practical application, because Webpage search task amount is big, web page search system generally comprises multiple routers and corresponding
Retrieve node array.
The operation principle of above-mentioned web page search system is as follows:When router 101 receives the query word of client transmission,
Whether have the query word corresponding search result, if so, then directly by the search result of the caching if being inquired about in caching system 102
It is back to client;If caching system 102 does not cache the corresponding search result of the query word, in retrieval node array
At least one retrieval node (selecting at least N number of retrieval node altogether) is selected in 103 each row to search the query word
Rope, after the search result of the retrieval node is obtained, is on the one hand back to client by the search result, on the other hand should
Search result and corresponding query word are stored in the caching system, during to receive identical query word in next time, directly
Connect and corresponding search result is obtained from the caching system.
Based on above-mentioned operation principle, dashed forward in the flow (i.e. search mission amount to be dealt with the appointed time) of router
(for example, because the fortuitous events such as optical cable damage, power-off cause one or several routers to be stopped, other are just in the case of increasing
The router often worked shares the search mission for the router being stopped, then the router traffic of normal work is uprushed),
The search overhead of corresponding search node can also be significantly increased, even more than its upper loading limit, it is impossible to timely respond to so high
Search overhead, causes search speed slack-off, whole web page search system stability reduction or even machine of delaying.
The three, content of the invention
To overcome problem present in correlation technique, the present invention provides a kind of web search method and system.
First aspect present invention there is provided a kind of web search method, including:
Monitor the real-time traffic of web page search system;
When the real-time traffic is between default overload flow threshold and default disaster tolerance flow threshold, and, the webpage is searched
When query word default with target query word identical being not present in the caching system of cable system, according to the real-time traffic and normally
The ratio of flow determines to perform the retrieval node number Q of this search mission;Wherein, the default overload flow threshold is less than institute
State default disaster tolerance flow threshold;
Q retrieval node is selected in the retrieval node array of the web page search system, and triggers the Q retrieval selected
Node performs the real-time search mission to the target query word, obtains real-time search result.
With reference in a first aspect, in first aspect in the first feasible embodiment, methods described also includes:
The real-time search result is stored in the buffer unit of corresponding retrieval node.
The first feasible embodiment with reference to first aspect, in second of feasible embodiment of first aspect, institute
Stating method also includes:
The target query word and corresponding retrieval node number Q are stored in the caching system.
With reference in a first aspect, either second of the first feasible embodiment of first aspect or first aspect are feasible
Embodiment, it is described according to the real-time traffic and normal discharge in first aspect in the third feasible embodiment
Ratio determines the retrieval node number Q of this search mission of execution, including:
According to formulaCalculate the retrieval node number Q of this search mission;
Wherein, WnewThe real-time traffic is represented, W represents the normal discharge that the web page search system can be born, and N is represented
The columns of the retrieval node array.
With reference in a first aspect, either second of the first feasible embodiment of first aspect or first aspect are feasible
Embodiment, in the 4th kind of feasible embodiment of first aspect, the retrieval section of the selection web page search system
Q retrieval node in lattice array, including:
Q retrieval node needed for this search mission is selected by roller back-and-forth method.
With reference in a first aspect, in the 5th kind of feasible embodiment of first aspect, methods described also includes:
When the real-time traffic is more than the default disaster tolerance flow threshold, according to the real-time traffic and normal discharge
Ratio determines to perform the retrieval node number Q of this search mission, and is searched and target query word portion in the caching system
Divide the default query word of matching;
Q retrieval node in the retrieval node array of the web page search system is selected, and triggers the Q retrieval chosen
Node performs the real-time search mission to the target query word, obtains real-time search result;
The real-time search result and the corresponding preset search result of the default query word are merged, the target is obtained
The corresponding target search result of query word.
With reference to the 5th kind of feasible embodiment of first aspect, in the 6th kind of feasible embodiment of first aspect, institute
Stating method also includes:
According to the cryptographic Hash and its word segmentation result of each default query word stored in the caching system, in the caching
The inverted index of each participle is built in system, the corresponding cryptographic Hash of each participle row chain is obtained;
The default query word that lookup is matched with target query word part in the caching system, including:
Determine each corresponding target participle of target query word and the corresponding weighted value of each target participle;
The target participle as described in the sequential selection of weighted value from high to low, until the weighted value sum for the target participle chosen
Not less than default weight threshold;
At least one common factor cryptographic Hash is judged whether while positioned at the corresponding Kazakhstan of the multiple target participles chosen
Uncommon value is fallen in row chain;
If there is the common factor cryptographic Hash, then the corresponding default query word of the common factor cryptographic Hash is labeled as and target
The default query word of query word part matching.
With reference to the 6th kind of feasible embodiment of first aspect, in the 7th kind of feasible embodiment of first aspect, institute
Stating method also includes:
The target query word and corresponding retrieval node number Q, common factor cryptographic Hash are stored in the caching system.
The second aspect of the embodiment of the present invention there is provided a kind of web page search system, including:At flow monitoring unit and overload
Manage unit;
Wherein, the flow monitoring unit is used for, and monitors the real-time traffic of web page search system;
The overload processing unit is used for, and handles the real-time traffic in default overload flow threshold and default disaster tolerance flow
To the search mission of target query word when between threshold value;Wherein, the default overload flow threshold is less than the default disaster tolerance stream
Measure threshold value;
The overload processing unit includes:Overload node computing unit and overload node selecting unit;
Wherein, the overload node computing unit is used for, when the real-time traffic is in default overload flow threshold and presets
It is not present to preset with target query word identical between disaster tolerance flow threshold, and in the caching system of the web page search system and looks into
When asking word, determined to perform the retrieval node number Q of this search mission according to the ratio of the real-time traffic and normal discharge;
The overload node selecting unit is used for, and selects Q inspection in the retrieval node array of the web page search system
Socket point, and the Q retrieval node execution chosen is triggered to the real-time search mission of the target query word, searched in real time
As a result.
With reference to second aspect, in second aspect in the first feasible embodiment, the overload processing unit also includes:
The buffer unit of result first is overloaded, the caching list for the real-time search result to be stored in corresponding retrieval node
Member.
The first feasible embodiment with reference to second aspect, in second of feasible embodiment of second aspect, institute
Stating overload processing unit also includes:The buffer unit of result second is overloaded, for by the target query word and corresponding retrieval section
Point number Q is stored in the caching system.
With reference to second aspect, either second of the first feasible embodiment of second aspect or second aspect are feasible
Embodiment, in second aspect in the third feasible embodiment, it is described overload node computing unit be specifically configured to:
According to formulaCalculate the retrieval node number Q of this search mission;
Wherein, WnewThe real-time traffic is represented, W represents the normal discharge that the web page search system can be born, and N is represented
The columns of the retrieval node array.
With reference to second aspect, either second of the first feasible embodiment of second aspect or second aspect are feasible
Embodiment, in the 4th kind of feasible embodiment of second aspect, it is described overload node selecting unit be specifically configured to:
Q retrieval node needed for this search mission is selected by roller back-and-forth method.
With reference to second aspect, in the 5th kind of feasible embodiment of second aspect, the system also includes:Disaster tolerance processing
Unit, to the search mission of target query word during for handling the real-time traffic more than the default disaster tolerance flow threshold;
The disaster tolerance processing unit includes:Disaster tolerance node computing unit, query word matching unit, disaster tolerance node selecting unit
With disaster tolerance result combining unit;
Disaster tolerance node computing unit, for when the real-time traffic is more than the default disaster tolerance flow threshold, according to institute
The ratio for stating real-time traffic and normal discharge determines to perform the retrieval node number Q of this search mission;
Query word matching unit, for when the real-time traffic is more than the default disaster tolerance flow threshold, described slow
The default query word matched with target query word part is searched in deposit system;
Disaster tolerance node selecting unit, the Q retrieval section in retrieval node array for selecting the web page search system
Point, and the Q retrieval node execution chosen is triggered to the real-time search mission of the target query word, obtain search knot in real time
Really;
Disaster tolerance result combining unit, for by the real-time search result and the corresponding preset search of the default query word
As a result merge, obtain the corresponding target search result of the target query word.
With reference to the 5th kind of feasible embodiment of second aspect, in the 6th kind of feasible embodiment of second aspect, institute
Stating system also includes:
Index construct unit, for the cryptographic Hash according to each default query word stored in the caching system and its point
Word result, builds the inverted index of each participle in the caching system, obtains the corresponding cryptographic Hash of each participle row chain;
Accordingly, the query word matching unit includes:
Participle Weight Acquisition unit, for determining each corresponding target participle of target query word and each target participle pair
The weighted value answered;
Participle selecting unit, for the target participle as described in weighted value sequential selection from high to low, until the mesh chosen
The weighted value sum for marking participle is not less than default weight threshold;
Common factor judging unit, for judging whether at least one common factor cryptographic Hash while that is chosen described in is multiple
The corresponding cryptographic Hash of target participle is fallen in row chain, if there is the common factor cryptographic Hash, then the common factor cryptographic Hash is corresponding
Default query word is labeled as the default query word matched with target query word part.
With reference to the 6th kind of feasible embodiment of second aspect, in the 7th kind of feasible embodiment of second aspect, institute
Stating disaster tolerance processing unit also includes:
Disaster tolerance result cache unit, for by the target query word and corresponding retrieval node number Q, common factor cryptographic Hash
It is stored in the caching system.
From above technical scheme, the embodiment of the present application is increased to overload level in the real-time traffic of web page search system
And chosen in caching system in the absence of in the case of the preset search result corresponding with target query word with respect under normal discharge
The retrieval node of less number performs real-time search mission to target query word, is returned by sacrificing for each target query word
The real-time search result quantity returned reduces the purpose of each load for retrieving node to reach, improves the response speed of whole system
Degree.In addition, when real-time traffic reaches disaster tolerance rank, the present embodiment is on the one hand true according to the ratio of real-time traffic and normal discharge
Surely the retrieval node number Q of this search mission is performed, and selection Q retrieves node to target query in retrieval node array
Word performs real-time search mission, real-time search result is obtained, on the other hand also by obtaining preset search knot from caching system
Fruit is used as compensation;The search result that client is back under disaster tolerance rank is made up of two parts:Retrieve the real-time of node acquisition
Search result and the preset search result obtained from caching system, can both reduce the load of retrieval node, can ensure again
The integrality of search result.
It should be appreciated that the general description of the above and detailed description hereinafter are only exemplary and explanatory, not
The disclosure can be limited.
Four, are illustrated
Accompanying drawing herein is merged in specification and constitutes the part of this specification, shows the implementation for meeting the present invention
Example, and for explaining principle of the invention together with specification.
Fig. 1 is the Organization Chart of web page search system in correlation technique.
Fig. 2 is a kind of flow chart of web search method according to an exemplary embodiment.
Fig. 3 is the flow chart of another web search method according to an exemplary embodiment.
Fig. 4 is the flow chart of another web search method according to an exemplary embodiment.
Fig. 5 is the flow chart of another web search method according to an exemplary embodiment.
Fig. 6 is a kind of structured flowchart of web page search system according to an exemplary embodiment.
Fig. 7 is the structured flowchart of another web page search system according to an exemplary embodiment.
Fig. 8 is the structured flowchart of another web page search system according to an exemplary embodiment.
Five, embodiments
Here exemplary embodiment will be illustrated in detail, its example is illustrated in the accompanying drawings.Following description is related to
During accompanying drawing, unless otherwise indicated, the same numbers in different accompanying drawings represent same or analogous key element.Following exemplary embodiment
Described in embodiment do not represent and the consistent all embodiments of the present invention.On the contrary, they be only with it is such as appended
The example of the consistent apparatus and method of some aspects be described in detail in claims, the present invention.
Fig. 2 shows a kind of flow chart for web search method that the embodiment of the present application is provided.As shown in Fig. 2 this method
Comprise the following steps.
S11, the real-time traffic for monitoring web page search system.
S12, when the real-time traffic is between default overload flow threshold and default disaster tolerance flow threshold, and, the net
Page search system caching system in be not present with target query word identical preset query word when, according to the real-time traffic with
The ratio of normal discharge determines to perform the retrieval node number Q of this search mission.
Wherein, the default overload flow threshold is less than the default disaster tolerance flow threshold;, can with reference to practical situations
So that the default overload flow threshold is set as into maximum stream flow (the specific number that the web page search system normal work can bear
Depending on value is according to systematic function), the default disaster tolerance flow threshold is set as that the web page search system normal work can be held
2 times of the maximum stream flow received.
The flow of web page search system is divided into three by above-mentioned default overload flow threshold and default disaster tolerance flow threshold
Individual interval/rank, the risk that web page search system exists under different flow rank is different.Assuming that web page search system
Real-time traffic Wnew, it is W to preset overload flow threshold1, it is W to preset disaster tolerance flow threshold2If, then Wnew<W1, illustrate real-time traffic
The normal discharge scope that can be born in default level, i.e. system, under the default level, web page search system is using existing
Each search mission of technical finesse can meet application demand, if but W1<Wnew<W2, illustrate that real-time traffic is in overload level,
Under this traffic class, if web page search system still handles each search mission using prior art, its processing speed can be slack-off, if
Wnew>W2, illustrate that real-time traffic is in disaster tolerance rank, under this traffic class, appoint if still handling each search using prior art
Business, web page search system can have machine risk of delaying.Different processing methods are taken for different ranks.
S13, Q retrieval node is selected in the retrieval node array of the web page search system, and trigger Q selected
Node execution is retrieved to the real-time search mission of the target query word, real-time search result is obtained.
Above-mentioned real-time search result is the corresponding target search result of target query word, returns it into the target query word pair
The client answered, user are it is seen that desired search result.
Query word is preset with target query word identical in addition, if existing in the caching system, then directly returns to institute
State the corresponding preset search result of the default query word of cached in caching system described.This situation and prior art processing method class
Seemingly, here is omitted.
Above-mentioned technical proposal describes real-time traffic and is in overload level (i.e. W1<Wnew<W2) when web page search system institute
The searching method of use.System architecture with reference to shown in Fig. 1, when real-time traffic is in overload level, similarly to the prior art,
Router 101 is after the target query word from client is received, and whether be stored with phase in caching system 102 first
Same query word, if it is not, needing to retrieve in node array 103 by what M rows N row (M*N) retrieval node was constituted
Retrieve node and perform real-time search;In this case, searched in real time if the execution still chosen according to prior art searches for this
Retrieval node number it is identical under normal circumstances with flow, then because flow system flow is uprushed, average each retrieval node is identical
The search mission to be performed also can accordingly increase in time, that is, retrieve the load of node and accordingly increase.
In view of this, the present embodiment determines to perform this search mission according to the ratio of the real-time traffic and normal discharge
Retrieval node number Q, it is ensured that Q is less than selected retrieval node number in the case of normal discharge, so as to reduce each retrieval
The average load of node.
For example, it is assumed that totally 10 retrieval nodes in system, in the case of normal discharge, for each target query word, this 10
Individual retrieval node is required for performing real-time search respectively, returns to 10 real-time search results;Then according to prior art, work as real-time streams
When amount is in overload level, for x target query word, average each retrieval node is required for performing x search in real time, if but
Using the embodiment of the present application, reduce and perform the retrieval node number Q searched in real time for each target query word, such as take Q=5,
Then for x target query word, average each retrieval node need to only perform x/2 search in real time, although each target query word
The real-time search result number returned is reduced to 5, but the load of each retrieval node can be reduced into half, i.e. this implementation
Example is by sacrificing the part of each target query word search result in real time, to ensure the response speed of web page search system and steady
It is qualitative.
Under overload level, flow system flow is uprushed, if will still retrieve the real-time search result of node according to prior art
It is stored in caching system, caching system also is difficult to bear so high load, easily causes caching system collapse.Therefore, originally
Apply for embodiment under overload level, in the buffer unit that real-time search mission is stored in corresponding retrieval node oneself, that is, examine
Socket point 11 is searched for obtained real-time search result and is stored in the buffer unit of retrieval node 11, and the retrieval search of node 12 is obtained
Real-time search result be stored in retrieval node 12 buffer unit in, the rest may be inferred.It can be seen that, it is above-mentioned to incite somebody to action real under overload level
When search result be stored in the buffer unit of corresponding retrieval node, the storage burden of caching system can be mitigated, it is to avoid caching
System is collapsed because of load too high.
From above technical scheme, the embodiment of the present application is increased to overload level in the real-time traffic of web page search system
And chosen in caching system in the absence of in the case of the preset search result corresponding with target query word with respect under normal discharge
The retrieval node of less number performs real-time search mission to target query word, is returned by sacrificing for each target query word
The real-time search result quantity returned reduces the purpose of each load for retrieving node to reach, so as to improve the response of whole system
Speed.
Reference picture 3, in one feasible embodiment of the application, above-mentioned web search method can also include following step
Suddenly:
S14, the buffer unit that the real-time search result is stored in corresponding retrieval node.
Real-time search result is stored in the buffer unit of corresponding retrieval node by the present embodiment under overload level, can be with
When retrieval node receives the real-time search mission for identical target query word next time, directly from the caching list of oneself
Cached search result is taken out in member;When i.e. retrieval node performs real-time search mission in step s 13, oneself is searched for first
Buffer unit in whether be stored with target query word to corresponding real-time search result, if so, then directly return delay
The real-time search result deposited, if not provided, being searched in real time in internet again.It can be seen that, it is above-mentioned to incite somebody to action real under overload level
When search result be stored in the buffer unit of corresponding retrieval node, can further mitigate the load of retrieval node, improve whole
The response speed of individual system.
In one feasible embodiment of the application, above-mentioned steps S15 specifically can determine level-overload according to equation below
It is not lower to perform the retrieval node number Q searched in real time:Wherein, WnewThe real-time traffic is represented, W represents institute
The normal discharge that web page search system can be born is stated, N represents the columns of the retrieval node array.Because W and N are definite value,
And under overload level, Wnew>W, therefore according to above-mentioned formula, Q is certainly less than N, and Q is with real-time traffic WnewIncrease and reduce,
So as to avoid retrieving node overload.
In one feasible embodiment of the application, the inspection of the selection web page search system described in above-mentioned steps S16
Q retrieval node in socket lattice array, including:Q retrieval needed for this search mission is selected by roller back-and-forth method
Node.
The above-mentioned roller back-and-forth method i.e. progressive predetermined number on the basis of upper once selection result starts to select this institute
The retrieval node needed.For example, it is assumed that eight retrieval nodes that numbering is 1 to 8 are had in web search system, under overload level
Calculating obtains Q=5, then when retrieving node for first aim query selection, can select to number the retrieval section for being 1 to 5
Point performs the real-time search mission to first aim query word, when retrieving node for second target query selection, passs
Enter that 3 (i.e. 8-5) are individual, selected since the retrieval node that numbering is 4, i.e., it is final it is selected number the retrieval node for being 4 to 8, for
When 3rd target query selected ci poem selects retrieval node, continue progressive 3, selected since the retrieval node that numbering is 7, i.e., it is final
The retrieval node that selected numbering is 7,8 and 1 to 3.And for example, it is assumed that web page search system is altogether comprising 180 row retrieval nodes, Q=
120, then progressive 60 row, which are rolled, selects retrieval node, i.e., to first aim query word, respectively from each column in the 1st to 120 row
Middle one retrieval node of selection performs search mission, to second target query word, respectively from each column in the 61st to 180 row
One retrieval node of selection performs search mission, to the 3rd target query word, respectively from the 121st to 180 row and the 1st to 60
A retrieval node is selected to perform search mission in each column in row, the rest may be inferred.
The present embodiment selects to perform the retrieval node of real-time search mission every time by roller back-and-forth method, can avoid
Part retrieval node searching number of times is excessive, load too high, and another part retrieval node searching number of times is very few, and load is relatively low,
Will repeatedly real-time search mission mean allocation to each retrieval node.
Reference picture 4, in one feasible embodiment of the application, current-carrying capacity threshold was preset in real-time traffic less than described
It is worth (i.e. in default level), and, it is not present in the caching system of the web page search system pre- with target query word identical
If during query word, the web search method comprises the following steps:
S22, when real-time traffic be less than the default overload flow threshold, and, the caching system of the web page search system
In be not present with target query word identical preset query word when from retrieval node array (M rows N row) each row in select respectively
A retrieval node is selected, the selected N number of retrieval node of triggering performs the real-time search mission to target query word, obtained in real time
Search result.
That is, under default level, node number Q=N is retrieved, it is ensured that the integrality of search result.Specifically, can root
Come to select a conduct from M retrieval node in same row according to the cryptographic Hash (Hash, also known as hashed value) of target query word
The retrieval node of real-time search mission is performed in this row to target query word.
In addition, under default level, if existed and the target query word in the caching system of the web page search system
Identical presets query word, then directly returns to the corresponding preset search result of default query word.
Further, the web search method under default level also includes:
S23, the real-time search result and corresponding target query word be stored in caching system.
Query word and corresponding search result can be stored using distributed storage mode in caching system;Specifically
, in above-mentioned steps S23, corresponding real-time search result can be regard as key assignments using target query word as major key (key)
(value) it is stored in caching system.The features such as above-mentioned distributed storage mode has fast inquiry velocity, support high concurrent,
The response speed of caching system can be ensured.
Corresponding with above-mentioned steps S23 referring now still to Fig. 4, above-mentioned web search method can also be wrapped under overload level
Include following steps:
S15, the retrieval node number Q and corresponding target query word be stored in the caching system.
Corresponding to above-mentioned distributed storage mode, in step S15 storing process, target query word can be regard as master
Key, corresponding retrieval node number Q is used as key assignments.
It can be seen that, it is different from step S23, in step S15, i.e., under overload level, target query word is only stored in caching system
With corresponding Q values, without storing real-time search result, (according to step S14, real-time search result is stored in accordingly under overload level
In the buffer unit for retrieving node), so as to greatly reduce shared memory space in caching system, it is to avoid high flow capacity
Lower caching system collapses because storing load too high.When identical query word is received in next time, it can be looked into caching system
The corresponding Q values of the query word are found, and then pass through hash algorithm, you can learn that last time performs the real-time search times to the query word
Q retrieval node of business, so as to read the corresponding search knot of the query word in this Q buffer unit for retrieving node
Really;It is of course also possible to according to the actual requirements and the factor such as real-time flow data, it is determined whether still retrieve node using this Q, be
It is no to need several retrieval nodes of increase, to ensure the integrality of search result.Relative to recalculating Q values and again by retrieving
Node performs real-time search mission, and the present embodiment is realized jointly by above-mentioned steps S14 and S15 to be delayed to two grades of search result
Deposit, the search efficiency and response speed of system under overload level on the premise of caching system load is mitigated, can be improved.
Reference picture 5, in one feasible embodiment of the application, based on embodiment illustrated in fig. 2, is more than in real-time traffic
During default disaster tolerance flow threshold (i.e. in disaster tolerance rank), the web search method comprises the following steps:
S32, when the real-time traffic be more than the default disaster tolerance flow threshold when, according to the real-time traffic and normal stream
The ratio of amount determines to perform the retrieval node number Q of this search mission, and is searched and target query in the caching system
The default query word of word part matching;
Q retrieval node in S33, the retrieval node array of the selection web page search system, and trigger Q chosen
Node execution is retrieved to the real-time search mission of the target query word, real-time search result is obtained;
S34, the real-time search result and the corresponding preset search result of the default query word merged, obtain described
The corresponding target search result of target query word.
When reaching disaster tolerance rank, it is meant that flow system flow has reached that system can bear 3 times even 10 times of flow, is
System is likely to bear so high load so that machine of delaying, it is impossible to provide search service.For such case, the present embodiment exists
When receiving target query word, on the one hand determined to perform the inspection of this search mission according to the ratio of real-time traffic and normal discharge
Rope node number Q, and Q retrieval node of selection performs real-time search mission to target query word in retrieval node array, obtains
To real-time search result.Because real-time traffic is too high under disaster tolerance rank, therefore the Q values meeting very little finally determined, so that retrieval node
The real-time search result returned also can be seldom.Therefore, the present embodiment from caching system on the other hand also by obtaining default search
Hitch fruit is as compensation, i.e., the search result that client is back under disaster tolerance rank is made up of two parts:Retrieve what node was obtained
Real-time search result and the preset search result obtained from caching system, can both avoid retrieval node load too high, again may be used
To ensure the integrality of search result.
Specifically, in view of existing in caching system with the probability of the identical default query word of target query word very
Small, the present embodiment searches the default query word by the way of part is matched, i.e., searched and target query in caching system
The default query word of word part matching, and then read and return to these corresponding preset search results of similar default query word.
In one feasible embodiment of the application, it can be realized by inverted index under disaster tolerance rank in caching system
Part matched to target query word is searched.Wherein, the embodiment of the present application also comprises the following steps, with the structure in caching system
Build above-mentioned inverted index:
According to the cryptographic Hash and its word segmentation result of each default query word stored in the caching system, in the caching
The inverted index of each participle is built in system, the corresponding cryptographic Hash of each participle row chain is obtained.
In the present embodiment, with the increase of target query word and corresponding search result new in caching system, hold in real time
Row above-mentioned steps realize the structure and renewal to the inverted index.For example, the Hash of 3 query words in input-buffer system
Value and its word segmentation result are as shown in the table.
Query word | Cryptographic Hash | Word segmentation result |
Query1 | hash1 | A、B |
Query2 | Hash2 | A、C、E、D |
Query3 | Hash3 | B、C |
Upper table is considered as positive index, i.e., search corresponding participle (Term) according to cryptographic Hash, be converted to inverted index form,
The corresponding row chain of each participle is as follows:
A→hash1 hash2;
B→hash1 hash3;
C→hash2 hash3;
D→hash2;
E→hash2。
When there is new query word input-buffer system, directly on the basis of established inverted index by insertion or
Newly-increased mode is updated, for example, the query word Query4 of input-buffer system cryptographic Hash is Hash4, word segmentation result is
" C, D, F ", then be updated on the basis of the inverted index of five compositions of falling row chain of above-mentioned A to E, as a result as follows:
A→hash1 hash2;
B→hash1 hash3;
C→hash2 hash3 hash4;
D→hash2 hash4;
E→hash2;
F→hash4。
, can be according to participle lookup and related cryptographic Hash by above-mentioned inverted index, and then looking into for correlation can be determined
Ask word (each query word corresponds to a cryptographic Hash).
Based on above-mentioned inverted index, being searched and target query word part in caching system described in above-mentioned steps S32
The default query word matched somebody with somebody, specifically may include steps of:
Determine each corresponding target participle of target query word and the corresponding weighted value of each target participle;
The target participle as described in the sequential selection of weighted value from high to low, until the weighted value sum for the target participle chosen
Not less than default weight threshold;
At least one common factor cryptographic Hash is judged whether while positioned at the corresponding Kazakhstan of the multiple target participles chosen
Uncommon value is fallen in row chain;
If there is the common factor cryptographic Hash, then the corresponding default query word of the common factor cryptographic Hash is labeled as and target
The default query word of query word part matching.
For example, participle is carried out to some target query word can obtain tri- target participles of A, C, D, corresponding weighted value point
Not Wei WA, WC and WD, and, WA > WC > WD;Then added up by the order of weighted value from high to low, WA+WC result reaches default
Threshold value W ' (i.e. WA < W ' and WA+WC >=W ') is inquired about, then selected target participle is A and C;Obtain the corresponding row chains of A and C
" A → hash1 hash2 " and " C → hash2 hash3 hash4 ", and to two progress of falling row chain intersection operations, it is determined that in the presence of
One common factor cryptographic Hash hash2, the then corresponding default query word Query2 of common factor cryptographic Hash hash2 and target query word part
Match somebody with somebody;And then read common factor cryptographic Hash hash2 (presetting query word Query2) corresponding preset search result in caching system,
The real-time search result obtained with retrieval node is back to corresponding in the lump as the final search result of the target query word
Client.
Certainly, in particular cases, if there is no the common factor cryptographic Hash, then illustrate to be not present in caching system and target
The default query word of query word part matching, it is impossible to similar preset search result is obtained from caching system, is eventually returned to
The a small amount of real-time search result also just only obtained in the search result of client comprising retrieval node.
From above step, the present embodiment realizes the part in caching system to target query word by inverted index
Matched is searched, simple and easy to apply, so as to easily get the preset search knot that target query word is related from caching system
Really.
Corresponding to step S15 and step S23 described above, the Webpage searching result caching method that the present embodiment is provided exists
Also comprise the following steps under disaster tolerance rank:
The target query word and corresponding retrieval node number Q, common factor cryptographic Hash are stored in the caching system.
Based on above-mentioned steps, whole search result is not stored in caching system under disaster tolerance rank, and only stores target
Query word and corresponding retrieval node number Q, common factor cryptographic Hash, on the one hand can reduce the storage load of caching system, another
Aspect can also when system receives same target query word again, directly found in caching system corresponding Q values and
Common factor cryptographic Hash, and then corresponding real-time search result and the preset search result of quick obtaining, improve the sound of web page search system
Answer speed.
From above method embodiment, the real-time traffic of web page search system is divided into three ranks by the application, by low
It is followed successively by height:Default level, overload level and disaster tolerance rank;And different Webpage search sides are provided for different ranks
Method, it is ensured that the stability of a system and response speed under each rank, it is to avoid caching system or retrieval node overload, it is to avoid system is delayed
Machine.
In addition, the embodiment of the present application additionally provides a kind of computer-readable storage medium, for example, can be that ROM, arbitrary access are deposited
Reservoir (RAM), CD-ROM, tape, floppy disk and optical data storage devices etc.;Had program stored therein in the computer-readable storage medium, when
Program in the storage medium is in web page search system during corresponding computing device so that web page search system can be held
The part or all of step of web search method described in row above method embodiment.
A kind of structured flowchart for web page search system that Fig. 6 provides for the embodiment of the present application;The system performs list by multiple
Member composition, each execution unit is divided into the equipment such as router, caching system, retrieval node.Reference picture 6, constitutes the webpage
The execution unit of search system at least includes:Flow monitoring unit 100 and overload processing unit 200.
Wherein, the flow monitoring unit 100 is used for, and monitors the real-time traffic of web page search system, and according to default overload
Flow threshold and default disaster tolerance flow threshold determine the interval residing for the real-time traffic.Wherein, it is described to preset current-carrying capacity threshold
Value is less than the default disaster tolerance flow threshold.Relative to system shown in Figure 1 framework, the flow monitoring unit 100 can be arranged at
In router 101.
The overload processing unit 200 is used for, and handles the real-time traffic in default overload flow threshold and default disaster tolerance stream
To the search mission of target query word when between amount threshold value.
Specifically, the overload processing unit 200 includes:Overload node computing unit 201 and overload node selecting unit
202.Relative to system shown in Figure 1 framework, overload node computing unit 201 and overload node selecting unit 202 can be arranged at
In router 101,
The overload node computing unit 201 is used for, when the real-time traffic is in default overload flow threshold and default disaster tolerance
It is not present between flow threshold, and in the caching system of the web page search system and the default query word of target query word identical
When, determined to perform the retrieval node number Q of this search mission according to the ratio of the real-time traffic and normal discharge;
The overload node selecting unit 202 is used for, and selects Q inspection in the retrieval node array of the web page search system
Socket point, and the Q retrieval node execution chosen is triggered to the real-time search mission of the target query word, searched in real time
As a result.
From system above structure, real-time traffic liter of the embodiment of the present application the embodiment of the present application in web page search system
In the case of the preset search result corresponding with target query word up to is not present in overload level and caching system, phase is chosen
The retrieval node of less number performs real-time search mission to target query word under normal stream amount, and each mesh is directed to by sacrificing
The real-time search result quantity that mark query word is returned reduces the purpose of each load for retrieving node to reach, and then improves whole
The response speed of individual system.
In one feasible embodiment of the application, above-mentioned overload node computing unit 201 is specifically configured to:According to
FormulaCalculate the retrieval node number Q of this search mission;Wherein, WnewThe real-time traffic is represented, W is represented
The normal discharge that the web page search system can be born, N represents the columns of the retrieval node array.
In the application in another feasible embodiment, above-mentioned overload node selecting unit 202 is specifically configured to:It is logical
Q crossed needed for roller back-and-forth method selects this search mission retrieves node.
Reference picture 7, in one feasible embodiment of the application, above-mentioned web page search system also includes:Default treatment
Unit 300, to target query word during for handling real-time traffic less than default overload flow threshold (i.e. the default level)
Search mission.The default treatment unit 300 can include:Default node selecting unit 301 and acquiescence result cache unit 302.
The default node selecting unit 301 can be arranged in router 101, for being less than described preset when real-time traffic
Flow threshold is overloaded, and, it is not present in the caching system of the web page search system and the default inquiry of target query word identical
Select a retrieval node, triggering selected N number of retrieval during word respectively from each row of retrieval node array (M rows N row)
Node performs the real-time search mission to target query word, obtains real-time search result.
The acquiescence result cache unit 302 is arranged in caching system 102, for store the real-time search result and
Corresponding target query word.
Corresponding to above-mentioned acquiescence result cache unit 302, above-mentioned overload processing unit 200 can also include:Overload result
First buffer unit 203 and the second buffer unit of overload result 204.
Wherein, the first buffer unit of overload result 203 is arranged in each retrieval node, for storing corresponding retrieval node
The real-time search result obtained.
The second buffer unit of overload result 204 is arranged in caching system 102, for storing what is received under overload level
Target query word and corresponding retrieval node number Q.
It can be seen that, under overload level, deposited by the first buffer unit of overload result 203 being arranged in each retrieval node
The second buffer unit of overload result 204 in the corresponding search result in real time of storage, caching system stores target query words and right
The Q values answered, so as to greatly reduce shared memory space in caching system, it is to avoid caching system is because depositing under high flow capacity
Store up load too high and collapse.When identical query word is received in next time, it can be looked into overload the second buffer unit of result 204
The corresponding Q values of the query word are found, and then pass through hash algorithm, you can learn that last time performs the real-time search times to the query word
Q retrieval node of business, so as to read the corresponding search knot of the query word in this Q buffer unit for retrieving node
Really.Relative to recalculating Q values and performing real-time search mission by retrieval node again, the present embodiment is by being arranged at retrieval section
The buffer unit of overload result first and the buffer unit of overload result second being arranged in caching system realization pair jointly in point
The L2 cache of search result, on the premise of caching system load is mitigated, can improve the search effect of system under overload level
Rate and response speed.
Reference picture 8, in one feasible embodiment of the application, above-mentioned web page search system also includes:Disaster tolerance processing
Unit 400, to target query word during for handling real-time traffic more than default disaster tolerance flow threshold (i.e. the disaster tolerance rank)
Search mission.The disaster tolerance processing unit 400 can include:Disaster tolerance node computing unit 401, query word matching unit 402, disaster tolerance
Node selecting unit 403 and disaster tolerance result combining unit 404.
The disaster tolerance node computing unit 401 is arranged in router 101, described default for being more than in the real-time traffic
During disaster tolerance flow threshold, determined to perform the retrieval node of this search mission according to the ratio of the real-time traffic and normal discharge
Number Q;
The query word matching unit 402 is arranged in caching system 102, described default for being more than in the real-time traffic
During disaster tolerance flow threshold, the default query word matched with target query word part is searched in the caching system;
The disaster tolerance node selecting unit 403 is arranged in router 101, the retrieval for selecting the web page search system
Q retrieval node in node array, and trigger real-time search of the Q retrieval node execution chosen to the target query word
Task, obtains real-time search result;
The disaster tolerance result combining unit 404 is arranged in router 101, for by the real-time search result and described pre-
If the corresponding preset search result of query word merges, the corresponding target search result of the target query word is obtained.
From result above, when real-time traffic reaches disaster tolerance rank, the present embodiment on the one hand according to real-time traffic with
The ratio of normal discharge determines to perform the retrieval node number Q of this search mission, and Q inspection of selection in retrieval node array
Socket point performs real-time search mission to target query word, obtains real-time search result, on the other hand also by from caching system
Middle acquisition preset search result is used as compensation;The search result that client is back under disaster tolerance rank is made up of two parts:Inspection
The real-time search result of socket point acquisition and the preset search result obtained from caching system, can both reduce retrieval node
Load, can ensure the integrality of search result again.
In one feasible embodiment of the application, above-mentioned web page search system can also include:Index construct unit, if
It is placed in caching system 102, for the cryptographic Hash and its participle according to each default query word stored in the caching system
As a result, the inverted index of each participle is built in the caching system, the corresponding cryptographic Hash of each participle row chain is obtained.Phase
Answer, above-mentioned query word matching unit 402 can specifically include:
Participle Weight Acquisition unit, for determining each corresponding target participle of target query word and each target participle pair
The weighted value answered;
Participle selecting unit, for the target participle as described in weighted value sequential selection from high to low, until the mesh chosen
The weighted value sum for marking participle is not less than default weight threshold;
Common factor judging unit, for judging whether at least one common factor cryptographic Hash while that is chosen described in is multiple
The corresponding cryptographic Hash of target participle is fallen in row chain, if there is the common factor cryptographic Hash, then the common factor cryptographic Hash is corresponding
Default query word is labeled as the default query word matched with target query word part.
In one feasible embodiment of the application, corresponding to above-mentioned acquiescence result cache unit 302, overload result second
Buffer unit 203 and the second buffer unit of overload result 204, above-mentioned disaster tolerance processing unit 400 can also include:Disaster tolerance result is delayed
Memory cell;The disaster tolerance result cache unit is arranged in caching system 102, for when flow system flow is in disaster tolerance rank, depositing
Store up the target query word and corresponding retrieval node number Q, common factor cryptographic Hash.
It can be seen that, whole search result is not stored in caching system under disaster tolerance rank, and only pass through disaster tolerance result cache
Unit stores target query word and corresponding retrieval node number Q, common factor cryptographic Hash, on the one hand can reduce depositing for caching system
Storage load, on the other hand can also directly find when system receives same target query word again in caching system
Corresponding Q values and common factor cryptographic Hash, and then corresponding real-time search result and the preset search result of quick obtaining, improve webpage and search
The response speed of cable system.
Those skilled in the art will readily occur to its of the present invention after considering specification and putting into practice invention disclosed herein
Its embodiment.The application be intended to the present invention any modification, purposes or adaptations, these modifications, purposes or
Person's adaptations follow the general principle of the present invention and including the undocumented common knowledge in the art of the disclosure
Or conventional techniques.Description and embodiments are considered only as exemplary, and true scope and spirit of the invention are by following
Claim is pointed out.
It should be appreciated that the invention is not limited in the precision architecture for being described above and being shown in the drawings, and
And various modifications and changes can be being carried out without departing from the scope.The scope of the present invention is only limited by appended claim.
Claims (16)
1. a kind of web search method, it is characterised in that including:
Monitor the real-time traffic of web page search system;
When the real-time traffic is between default overload flow threshold and default disaster tolerance flow threshold, and, the Webpage search system
When query word default with target query word identical being not present in the caching system of system, it can be born according to the web page search system
Normal discharge and the real-time traffic ratio and product with the columns of the retrieval node array of the web page search system,
It is determined that performing the retrieval node number Q of this search mission;Wherein, the default overload flow threshold is less than the default disaster tolerance
Flow threshold;
Q retrieval node is selected in the retrieval node array of the web page search system, and triggers the Q retrieval node selected
The real-time search mission to the target query word is performed, real-time search result is obtained.
2. according to the method described in claim 1, it is characterised in that also include:
The real-time search result is stored in the buffer unit of corresponding retrieval node.
3. method according to claim 2, it is characterised in that also include:
The target query word and corresponding retrieval node number Q are stored in the caching system.
4. the method according to any one of claims 1 to 3, it is characterised in that can be born according to the web page search system
Normal discharge and the real-time traffic ratio and product with the columns of the retrieval node array of the web page search system,
It is determined that the retrieval node number Q of this search mission is performed, including:
According to formulaCalculate the retrieval node number Q of this search mission;
Wherein, WnewThe real-time traffic is represented, W represents the normal discharge that the web page search system can be born, and N represents described
Retrieve the columns of node array.
5. the method according to any one of claims 1 to 3, it is characterised in that the selection web page search system
Q retrieval node in node array is retrieved, including:
Q retrieval node needed for this search mission is selected by roller back-and-forth method.
6. according to the method described in claim 1, it is characterised in that also include:
When the real-time traffic is more than the default disaster tolerance flow threshold, according to the real-time traffic and the ratio of normal discharge
It is determined that performing the retrieval node number Q of this search mission, and searched and target query word part in the caching system
The default query word matched somebody with somebody;
Q retrieval node in the retrieval node array of the web page search system is selected, and triggers the Q retrieval node chosen
The real-time search mission to the target query word is performed, real-time search result is obtained;
The real-time search result and the corresponding preset search result of the default query word are merged, the target query is obtained
The corresponding target search result of word.
7. method according to claim 6, it is characterised in that also include:
According to the cryptographic Hash and its word segmentation result of each default query word stored in the caching system, in the caching system
The middle inverted index for building each participle, obtains the corresponding cryptographic Hash of each participle row chain;
The default query word that lookup is matched with target query word part in the caching system, including:
Determine each corresponding target participle of target query word and the corresponding weighted value of each target participle;
The target participle as described in the sequential selection of weighted value from high to low, until the weighted value sum for the target participle chosen is not small
In default weight threshold;
At least one common factor cryptographic Hash is judged whether while positioned at the corresponding cryptographic Hash of multiple target participles chosen
In row chain;
If there is the common factor cryptographic Hash, then the corresponding default query word of the common factor cryptographic Hash is labeled as and target query
The default query word of word part matching.
8. method according to claim 7, it is characterised in that also include:
The target query word and corresponding retrieval node number Q, common factor cryptographic Hash are stored in the caching system.
9. a kind of web page search system, it is characterised in that including:Flow monitoring unit and overload processing unit;
Wherein, the flow monitoring unit is used for, and monitors the real-time traffic of web page search system;
The overload processing unit is used for, and handles the real-time traffic in default overload flow threshold and default disaster tolerance flow threshold
Between when to the search mission of target query word;Wherein, the default overload flow threshold is less than the default disaster tolerance flow threshold
Value;
The overload processing unit includes:Overload node computing unit and overload node selecting unit;
Wherein, the overload node computing unit is used for, when the real-time traffic is in default overload flow threshold and default disaster tolerance
It is not present between flow threshold, and in the caching system of the web page search system and the default query word of target query word identical
When, the ratio of the normal discharge that can be born according to the web page search system and the real-time traffic and with the Webpage search system
The product of the columns of the retrieval node array of system, it is determined that performing the retrieval node number Q of this search mission;
The overload node selecting unit is used for, and selects Q retrieval section in the retrieval node array of the web page search system
Point, and the Q retrieval node execution chosen is triggered to the real-time search mission of the target query word, obtain search knot in real time
Really.
10. system according to claim 9, it is characterised in that the overload processing unit also includes:Overload result first
Buffer unit, is specifically configured to:
The real-time search result is stored in the buffer unit of corresponding retrieval node.
11. system according to claim 10, it is characterised in that the overload processing unit also includes:
The buffer unit of result second is overloaded, it is described for the target query word and corresponding retrieval node number Q to be stored in
In caching system.
12. the system according to any one of claim 9 to 11, it is characterised in that the overload node computing unit is specific
It is configured as:
According to formulaCalculate the retrieval node number Q of this search mission;
Wherein, WnewThe real-time traffic is represented, W represents the normal discharge that the web page search system can be born, and N represents described
Retrieve the columns of node array.
13. the system according to any one of claim 9 to 11, it is characterised in that the overload node selecting unit is specific
It is configured as:
Q retrieval node needed for this search mission is selected by roller back-and-forth method.
14. system according to claim 9, it is characterised in that also include:Disaster tolerance processing unit, for handling the reality
To the search mission of target query word when Shi Liuliang is more than the default disaster tolerance flow threshold;
The disaster tolerance processing unit includes:Disaster tolerance node computing unit, query word matching unit, disaster tolerance node selecting unit and appearance
Calamity result combining unit;
Disaster tolerance node computing unit, for when the real-time traffic is more than the default disaster tolerance flow threshold, according to the reality
The ratio of Shi Liuliang and normal discharge determines to perform the retrieval node number Q of this search mission;
Query word matching unit, for when the real-time traffic is more than the default disaster tolerance flow threshold, in the caching system
The default query word matched with target query word part is searched in system;
Disaster tolerance node selecting unit, the Q retrieval node in retrieval node array for selecting the web page search system, and
The Q retrieval node execution chosen is triggered to the real-time search mission of the target query word, real-time search result is obtained;
Disaster tolerance result combining unit, for by the real-time search result and the corresponding preset search result of the default query word
Merge, obtain the corresponding target search result of the target query word.
15. system according to claim 14, it is characterised in that also include:
Index construct unit, for the cryptographic Hash and its participle knot according to each default query word stored in the caching system
Really, the inverted index of each participle is built in the caching system, the corresponding cryptographic Hash of each participle row chain is obtained;
Accordingly, the query word matching unit includes:
Participle Weight Acquisition unit, for determining that each corresponding target participle of target query word and each target participle are corresponding
Weighted value;
Participle selecting unit, for the target participle as described in weighted value sequential selection from high to low, until the target point chosen
The weighted value sum of word is not less than default weight threshold;
Common factor judging unit, for judging whether at least one common factor cryptographic Hash while positioned at the multiple targets chosen
The corresponding cryptographic Hash of participle is fallen in row chain, if there is the common factor cryptographic Hash, then the common factor cryptographic Hash is corresponding default
Query word is labeled as the default query word matched with target query word part.
16. system according to claim 15, it is characterised in that the disaster tolerance processing unit also includes:
Disaster tolerance result cache unit, for the target query word and corresponding retrieval node number Q, common factor cryptographic Hash to be stored
In the caching system.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510945454.0A CN105447187B (en) | 2015-12-15 | 2015-12-15 | Web search method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510945454.0A CN105447187B (en) | 2015-12-15 | 2015-12-15 | Web search method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105447187A CN105447187A (en) | 2016-03-30 |
CN105447187B true CN105447187B (en) | 2017-09-22 |
Family
ID=55557363
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510945454.0A Active CN105447187B (en) | 2015-12-15 | 2015-12-15 | Web search method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105447187B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108470043A (en) * | 2018-02-27 | 2018-08-31 | 阿里巴巴集团控股有限公司 | A kind of acquisition methods and device of business result |
CN110309390B (en) * | 2018-03-15 | 2021-10-08 | 阿里巴巴(中国)有限公司 | Index reduction method and device suitable for search and server |
CN108846094A (en) * | 2018-06-15 | 2018-11-20 | 江苏中威科技软件系统有限公司 | A method of based on index in classification interaction |
CN113032436B (en) * | 2021-04-16 | 2022-05-31 | 苏州臻璇数据信息技术有限公司 | Searching method and device based on article content and title |
CN114218013A (en) * | 2021-12-13 | 2022-03-22 | 北京字节跳动网络技术有限公司 | Searching method, searching device and electronic equipment storage medium |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2633444A4 (en) * | 2010-10-30 | 2017-06-21 | International Business Machines Corporation | Transforming search engine queries |
KR20140001498A (en) * | 2012-06-27 | 2014-01-07 | 네이버 주식회사 | System, apparatus, method and computer readable recording medium for providing an information related to a music by recognition of the music outputted through the television |
US8713010B1 (en) * | 2013-02-19 | 2014-04-29 | Luxian Limited | Processor engine, integrated circuit and method therefor |
CN103812949B (en) * | 2014-03-06 | 2016-09-07 | 中国科学院信息工程研究所 | A kind of task scheduling towards real-time cloud platform and resource allocation methods and system |
-
2015
- 2015-12-15 CN CN201510945454.0A patent/CN105447187B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN105447187A (en) | 2016-03-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105447187B (en) | Web search method and system | |
CN103729438B (en) | Webpage preloads method and device | |
US20030161338A1 (en) | Network path selection based on bandwidth | |
CN106202112A (en) | CACHE DIRECTORY method for refreshing and device | |
US9733833B2 (en) | Selecting pages implementing leaf nodes and internal nodes of a data set index for reuse | |
JP2005353039A5 (en) | ||
US10437820B2 (en) | Asymmetric distributed cache with data chains | |
CN102761627A (en) | Cloud website recommending method and system based on terminal access statistics as well as related equipment | |
CN102955829B (en) | For the method being ranked up to resource items, device and equipment | |
CN102314336A (en) | Data processing method and system | |
CN104956340B (en) | Expansible Data duplication is deleted | |
CN115660380B (en) | Order processing method and device for picking goods to person | |
CN106886376B (en) | A kind of marine monitoring data copy management method optimized based on more attributes | |
CN116321303A (en) | Data caching method, device, equipment and readable storage medium | |
Zarezadeh et al. | Dynamic network reliability modeling under nonhomogeneous Poisson processes | |
CN109325266B (en) | Response time distribution prediction method for online cloud service | |
CN106453611A (en) | A method and apparatus for load balancing at a plurality of storage nodes | |
CN106020974A (en) | Memory caching method and system for NUMA (Non Uniform Memory Access Architecture) platform | |
CN111309769A (en) | Method, device and computer storage medium for processing target information based on multi-satellite search to perform imaging task planning | |
CN106708874A (en) | Method and device for adjusting arrangement of searching categories in searching page | |
CN108733763B (en) | Method and device for calculating key nodes based on microblog hot events | |
CN110084455B (en) | Data processing method, device and system | |
KR20100038800A (en) | Method for updating data stored in cache server, cache server and content delivery system thereof | |
KR101795482B1 (en) | Device and method for encrypted data retrival | |
CN106547906A (en) | Content of pages generation method, device and terminal |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right |
Effective date of registration: 20200527 Address after: 310051 room 508, floor 5, building 4, No. 699, Wangshang Road, Changhe street, Binjiang District, Hangzhou City, Zhejiang Province Patentee after: Alibaba (China) Co.,Ltd. Address before: 510627 Guangdong city of Guangzhou province Whampoa Tianhe District Road No. 163 Xiping Yun Lu Yun Ping square B radio tower 12 layer self unit 01 Patentee before: GUANGZHOU SHENMA MOBILE INFORMATION TECHNOLOGY Co.,Ltd. |
|
TR01 | Transfer of patent right |