CN108804234A

CN108804234A - Data-storage system and its operating method

Info

Publication number: CN108804234A
Application number: CN201710296541.7A
Authority: CN
Inventors: 宋增宽; 宋立鑫
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2017-04-28
Filing date: 2017-04-28
Publication date: 2018-11-13
Anticipated expiration: 2037-04-28
Also published as: CN108804234B

Abstract

The invention discloses a kind of data-storage system and its operating methods.Wherein data-storage system includes multiple processors, wherein each processor includes shared drive, and the data of the shared drive storage between multiple processors are mutually in step, and wherein method includes：Receive data operation request；When data operation request is read request, cryptographic Hash is generated according to read request, and searched in shared drive according to cryptographic Hash；When finding corresponding data in shared drive, the data found are returned；When not finding corresponding data in shared drive, corresponding data are searched from database and are returned.Thus, the embodiment of the present invention can ensure that the operating time of most read requests is rdma read time stage, the concurrency asked under low latency response demand is supported to greatest extent, and significantly reduces technical costs budget, avoids the consumption of unnecessary high-performance hardware machine.

Description

Data-storage system and its operating method

Technical field

The present invention relates to technical field of data storage, more particularly to a kind of operating method of data-storage system, data are deposited Storage system.

Background technology

In field of data storage, industrial quarters has the ripe solution party towards pervasive scene (scene generally adapted to) Case is broadly divided into relevant database (such as SQL (Structured Query Language, structured query language)) and non- Relevant database (such as NoSQL (Not Only SQL, the database of non-relational)) two classes, the former (is a pass with MySQL Be type data base management system) be representative, the latter with Redis (be one increase income using ANSI C (be by American National Standard The standard about C language that association (ANSI) and International Organization for standardization (ISO) release) language writes, supports network, can be based on Memory also can persistence log type, Key (key)-Value (value) database)/Memcache (is that a set of distributed high speed is slow Deposit system) it is representative.Relational database structure thought, it is easy to operate, there is abundant characteristic to support, but complex, Concurrent tenability is insufficient；Non-relational database more light weight, is stored in the form of key-value, more preferable to being accompanied with Support.Under distributed environment, industry carries out network encapsulation to database, so that business side is called.

In the related technology, common business storage scheme can be as follows under pervasive scene：The proceeding internal memory of processor is made For the caching of data, and using non-relational database as L2 cache, and relevant database lands to data.It can manage The thinking of solution, this storage scheme is exactly in distributed environment, and business side is by network request and by certain routing Storage side is accessed after rule, it can be seen that, this storage strategy can be suitable for most of business scenario.

However, above-mentioned this storage scheme is using the proceeding internal memory in processor as the caching of data, once including data Deposit middle hit, so that it may to greatly reduce the time of data access.However, the technical issues of this storage scheme, is distribution The stateless strategy of routing, can lead to the request of same data source, may be distributed on different processors, however different Processor in the data that are stored of proceeding internal memory be different, therefore this just seriously constrains the utilization rate of memory cache.It lifts For example, it is assumed that the first time from data source A asks to have found corresponding caching number in the proceeding internal memory in processor B According to, the time of data access is reduced, however due to the stateless strategy of distribution routing, second of request of data source A It may be assigned to processor C, therefore when asking for second, have corresponding data due to uncached in processor C, Therefore it or needs to read corresponding data from database, to seriously constrain the utilization rate of memory cache.

Invention content

The purpose of the present invention is intended to solve one of the technical problem in above-mentioned technology at least to a certain extent.

For this purpose, first purpose of the present invention is to propose a kind of operating method of data-storage system.This method can be with The utilization rate for greatly improving shared drive caching, ensure that the operating time of most read requests is the rdma read time Grade, supports the concurrency asked under low latency response demand, and significantly reduce technical costs budget, avoids to greatest extent The consumption of unnecessary high-performance hardware machine.

Second object of the present invention is to propose a kind of data-storage system.

Third object of the present invention is to propose a kind of processor.

Fourth object of the present invention is to propose a kind of storage medium.

In order to achieve the above objectives, first aspect present invention embodiment proposes a kind of operating method of data-storage system, The data-storage system includes multiple processors, wherein each processor includes shared drive, between the multiple processor Shared drive storage data be mutually in step, the method includes：

Receive data operation request；

When the data operation request is read request, cryptographic Hash is generated according to the read request, and according to the Hash Value is searched in the shared drive；

When finding corresponding data in the shared drive, the data found are returned；

When not finding corresponding data in the shared drive, corresponding data are searched from database and are returned.

The operating method of the data-storage system of the embodiment of the present invention, can be applied in data-storage system, which deposits Storage system includes multiple processors, wherein each processor includes shared drive, and is stored in shared drive in each processor Data be synchronous.In this way, when receiving data operation request, judge which kind of request class is the data operation request belong to Type generates cryptographic Hash and according to cryptographic Hash in shared drive when the data operation request is read request according to the read request It is searched, if finding, returns to the data found, corresponding data are searched in otherwise being stored from subordinate and returned.At this In the embodiment of invention, since the data stored in shared drive in each processor are synchronous, as long as same data The request in source can be searched in shared drive, even if then secondary request is routed to other processors, then Due to the synchronism of shared drive, can also be found in the shared drive of other processors, therefore greatly improve shared The utilization rate of memory cache.In addition, the embodiment of the present invention is indexed in shared drive according to the cryptographic Hash of data, using Kazakhstan Data are indexed in the time complexity of uncommon indexed mode O (1), to ensure that the operating time of most read requests is in reading Time stage is deposited, supports the concurrency asked under low latency response demand to greatest extent, and it is pre- significantly to reduce technical costs It calculates, avoids the consumption of unnecessary high-performance hardware machine.

According to one embodiment of present invention, described described after searching corresponding data and return in database Operating method further includes：

When currently more new strategy is to read more new strategy, generated according to the corresponding data searched from database corresponding Cryptographic Hash and the corresponding data of the lookup is inserted by the shared drive according to the cryptographic Hash.

According to one embodiment of present invention, the shared drive includes multi-stage data node, wherein described in the basis The corresponding data searched from database generates corresponding cryptographic Hash and according to the cryptographic Hash by the corresponding data of the lookup It is inserted into the shared drive, is specifically included：

S11, the corresponding data searched from database according to generate corresponding first cryptographic Hash, and work as where obtaining The corresponding current exponent number value of preceding rank back end；

S12, according to first cryptographic Hash and current rank numerical generation first index value；

S13, when the back end is empty, the corresponding data searched from database is inserted into the data section Point；

S14, when the back end is not empty, update the current exponent number value；

S15, S12-S15 is repeated according to the current exponent number value after the update.

According to one embodiment of present invention, the corresponding data searched from database described in the basis generates corresponding The corresponding data of the lookup is simultaneously inserted into the shared drive by cryptographic Hash according to the cryptographic Hash, is specifically included：

S21, the corresponding data searched from database according to generate corresponding second cryptographic Hash and third cryptographic Hash；

S22, second index value and third index value are generated according to second cryptographic Hash and the third cryptographic Hash；

It, will be described from number when S23, the back end corresponding to the second index value or the third index value are empty The empty back end is inserted as according to the corresponding data searched in library；

It, will when S24, the back end corresponding to the second index value and the third index value are not empty Data in the back end corresponding to the second index value or the third index value are taken out, and will be from described second The data taken out in the back end corresponding to index value or the third index value are inserted into its of the shared drive His back end, and the corresponding data searched from database is inserted into the second index value or the third rope Draw the corresponding back end of value.

According to one embodiment of present invention, described will be corresponding to the second index value or the third index value Other back end that the data taken out in the back end are inserted into the shared drive specifically include：

S251, basis are taken out from the back end corresponding to the second index value or the third index value Data generate the 4th cryptographic Hash, and generate the 4th index value according to the 4th cryptographic Hash；

S252, when the corresponding back end of the 4th index value is empty, will be from the second index value or described the The data taken out in the back end corresponding to three index values are inserted into the corresponding back end of the 4th index value；

S253, when the corresponding back end of the 4th index value is not empty, by the corresponding number of the 4th index value It takes out according to the data of node and is inserted into again according to the data taken out from the corresponding back end of the 4th index value, with And the data taken out from the back end corresponding to the second index value or the third index value are inserted into institute State the corresponding back end of the 4th index value.

According to one embodiment of present invention, the operating method further includes：

When the data operation request is write request, the database is written into the corresponding data of the write request；

When currently more new strategy is to write more new strategy, corresponding cryptographic Hash is generated according to the corresponding data of the write request And the corresponding data of the write request are inserted by the shared drive according to the cryptographic Hash.

In one embodiment of the invention, the shared drive includes multi-stage data node, wherein described in the basis Corresponding cryptographic Hash is generated from the corresponding data of the write request and according to the cryptographic Hash by the corresponding data of the write request It is inserted into the shared drive, is specifically included：

S31, corresponding 5th cryptographic Hash is generated according to the corresponding data of the write request, and obtains the current exponent number evidence in place The corresponding current exponent number value of node；

S32, according to the 5th cryptographic Hash and current the 5th index value of rank numerical generation；

S33, when the back end is empty, the corresponding data of the write request are inserted into the back end；

S34, when the back end is not empty, update the current exponent number value；

S35, step S32-S35 is repeated according to the current exponent number value after the update.

In another embodiment of the present invention, described to generate corresponding cryptographic Hash according to the corresponding data of the write request And the corresponding data of the write request are inserted by the shared drive according to the cryptographic Hash, it specifically includes：

S41, corresponding 6th cryptographic Hash and the 7th cryptographic Hash are generated from the corresponding data of write request according to described；

S42, the 6th index value and the 7th index value are generated according to the 6th cryptographic Hash and the 7th cryptographic Hash；

When S43, the back end corresponding to the 6th index value or the 7th index value are empty, described write is asked Corresponding data are asked to be inserted as the empty back end；

It, will when S44, the back end corresponding to the 6th index value and the 7th index value are not empty Data in the back end corresponding to 6th index value or the 7th index value are taken out, and will be from the described 6th The data taken out in the back end corresponding to index value or the 7th index value are inserted into its of the shared drive His back end, and the corresponding data of the write request are inserted into the 6th index value or the 7th index value institute is right The back end answered.

Obtain total amount of data value in the shared drive；

The clearance time interval of the shared drive is obtained, and demand data amount is generated according to the clearance time interval Value；

It is to write update by the current update strategy setting when the demand data magnitude is less than the total amount of data value Strategy；And

When the demand data magnitude is greater than or equal to the total amount of data value, it is by the current update strategy setting Read more new strategy.

According to one embodiment of present invention, the shared drive in the multiple processor passes through broadcast subscription message Mode synchronize.

In order to achieve the above objectives, second aspect of the present invention embodiment puts forward a kind of data-storage system, including multiple places Reason machine, wherein each the processor includes：

Shared drive；

Request receiving module, for receiving data operation request；

Operation module, for when the data operation request is read request, cryptographic Hash to be generated according to the read request, and It is searched in the shared drive according to the cryptographic Hash, is returned when finding corresponding data in the shared drive The data found, and when not finding corresponding data in the shared drive, corresponding data are searched from database And it returns；And

Synchronization module, for synchronizing the data in the shared drive with the shared drive in other processors.

Data-storage system according to the ... of the embodiment of the present invention, it may include multiple processors, wherein each processor includes altogether Memory is enjoyed, and the data stored in shared drive in each processor are synchronous.In this way, each processor is receiving data When operation requests, judge which kind of request type is the data operation request belong to, when the data operation request is read request, according to The read request generates cryptographic Hash and is searched in shared drive according to cryptographic Hash, if finding, returns to the data found, Otherwise corresponding data are searched in being stored from subordinate and are returned.In an embodiment of the present invention, due to being shared in each processor The data stored in memory are synchronous, as long as therefore the request of same data source can be searched in shared drive, that It, then, also can be in other processing due to the synchronism of shared drive even if secondary request is routed to other processors It is found in the shared drive of machine, therefore greatly improves the utilization rate of shared drive caching.In addition, the embodiment of the present invention exists It is indexed according to the cryptographic Hash of data in shared drive, is indexed in the time complexity using hash index mode O (1) Data, to ensure that the operating time of most read requests is rdma read time stage, supporting low latency response to greatest extent needs The concurrency of lower request is sought, and significantly reduces technical costs budget, avoids the consumption of unnecessary high-performance hardware machine.

According to one embodiment of present invention, the operation module is additionally operable to：Corresponding data are being searched from database And after returning, when currently more new strategy is to read more new strategy, generated according to the corresponding data searched from database Corresponding cryptographic Hash, and the corresponding data of the lookup is inserted by the shared drive according to the cryptographic Hash.

According to one embodiment of present invention, the shared drive includes multi-stage data node, wherein the operation module It is specifically used for：

According to one embodiment of present invention, the operation module is specifically used for：

Corresponding second cryptographic Hash and third cryptographic Hash are generated according to the corresponding data searched from database；

Second index value and third index value are generated according to second cryptographic Hash and the third cryptographic Hash；

It, will be described from database when back end corresponding to the second index value or the third index value is empty The corresponding data of middle lookup is inserted as the empty back end；

It, will be described when the back end corresponding to the second index value and the third index value is not empty Data in the back end corresponding to second index value or the third index value are taken out, and will be from second index The data taken out in the back end corresponding to value or the third index value are inserted into other numbers of the shared drive It is inserted into the second index value or the third index value according to node, and by the corresponding data searched from database The corresponding back end.

According to the data taken out from the back end corresponding to the second index value or the third index value The 4th cryptographic Hash is generated, and the 4th index value is generated according to the 4th cryptographic Hash；

When the corresponding back end of the 4th index value is empty, will be indexed from the second index value or the third The data taken out in the corresponding back end of value are inserted into the corresponding back end of the 4th index value；

When the corresponding back end of the 4th index value is not empty, by the corresponding back end of the 4th index value Data take out and be inserted into again according to the data taken out from the corresponding back end of the 4th index value, and will be from The data taken out in the back end corresponding to the second index value or the third index value are inserted into the described 4th The corresponding back end of index value.

According to one embodiment of present invention, the operation module is additionally operable to：It is write request in the data operation request When, the database is written into the corresponding data of the write request, and when judging that currently more new strategy is writes more new strategy, root Corresponding cryptographic Hash is generated according to the corresponding data of the write request and according to the cryptographic Hash by the corresponding data of the write request It is inserted into the shared drive.

According to described corresponding 6th cryptographic Hash and the 7th cryptographic Hash are generated from the corresponding data of write request；

The 6th index value and the 7th index value are generated according to the 6th cryptographic Hash and the 7th cryptographic Hash；

When back end corresponding to the 6th index value or the 7th index value is empty, by the write request pair The data answered are inserted as the empty back end；

It, will be described when the back end corresponding to the 6th index value and the 7th index value is not empty Data in the back end corresponding to 6th index value or the 7th index value are taken out, and will be from the 6th index The data taken out in the back end corresponding to value or the 7th index value are inserted into other numbers of the shared drive It is inserted into corresponding to the 6th index value or the 7th index value according to node, and by the corresponding data of the write request The back end.

According to one embodiment of present invention, the operation module is additionally operable to：Obtain total amount of data in the shared drive Value, and the clearance time interval of the shared drive is obtained, and demand data magnitude is generated according to the clearance time interval, and It is to write more new strategy by the current update strategy setting when the demand data magnitude is less than the total amount of data value, with It is to read to update by the current update strategy setting and when the demand data magnitude is greater than or equal to the total amount of data value Strategy.

In order to achieve the above objectives, third aspect present invention embodiment proposes a kind of processor, including：Shell, processor, Memory, shared drive, circuit board and power circuit, wherein the circuit board is placed in the space interior that the shell surrounds, The processor, the memory and the shared drive are arranged on the circuit board；The power circuit, for being described The each circuit or device of processor are powered；The memory is for storing executable program code；The shared drive is used for The data of memory storage are cached；The processor is by reading the executable program stored in the memory Code runs program corresponding with the executable program code, for executing following steps：

Receive data operation request；

Processor according to the ... of the embodiment of the present invention judges the data operation request category when receiving data operation request In which kind of request type, when the data operation request is read request, cryptographic Hash is generated according to the read request and according to cryptographic Hash It is searched in shared drive, if finding, returns to the data found, corresponding data are searched in otherwise being stored from subordinate And it returns.In an embodiment of the present invention, due to the data stored in shared drive in each processor be it is synchronous, only Want the request of same data source that can be searched in shared drive, even if then secondary request is routed to others Processor can also be found, therefore greatly carry then due to the synchronism of shared drive in the shared drive of other processors The high utilization rate of shared drive caching.In addition, the embodiment of the present invention carries out rope in shared drive according to the cryptographic Hash of data Draw, data are indexed in the time complexity using hash index mode O (1), to ensure the operation of most read requests Time is rdma read time stage, supports the concurrency asked under low latency response demand to greatest extent, and significantly reduce Technical costs budget avoids the consumption of unnecessary high-performance hardware machine.

The additional aspect of the present invention and advantage will be set forth in part in the description, and will partly become from the following description Obviously, or practice through the invention is recognized.

Description of the drawings

Above-mentioned and/or additional aspect and advantage of the invention will become from the following description of the accompanying drawings of embodiments Obviously and it is readily appreciated that, wherein

Fig. 1 is the flow chart of the operating method of data-storage system according to an embodiment of the invention；

Fig. 2 is the flow chart of the operating method of data-storage system in accordance with another embodiment of the present invention；

Fig. 3 (a) is the flow chart that data according to an embodiment of the invention are inserted into shared drive；

Fig. 3 (b) is under multistage Hash management structure according to an embodiment of the invention, reading and writing data shared drive is gone through Journey exemplary plot；

Fig. 4 be it is in accordance with another embodiment of the present invention by from database lookup to data be inserted into shared drive in stream Cheng Tu；

Fig. 5 be it is according to an embodiment of the invention by from database lookup to data be inserted into shared drive other The flow chart of back end；

Fig. 6 is the flow chart according to the operating method of the data-storage system of another embodiment of the invention；

Fig. 7 is the flow chart according to an embodiment of the invention that the corresponding data of write request are inserted into shared drive；

Fig. 8 is the flow in accordance with another embodiment of the present invention that the corresponding data of write request are inserted into shared drive Figure；

Fig. 9 is the flow chart of the operating method of data-storage system in accordance with another embodiment of the present invention；

Figure 10 (a) is that the operating method of data-storage system according to an embodiment of the invention is applied to business scenario Exemplary plot；

Figure 10 (b) is that the operating method of data-storage system in accordance with another embodiment of the present invention is applied to business scenario Exemplary plot；

Figure 11 is the structural schematic diagram of data-storage system according to an embodiment of the invention；

Figure 12 is the major technique Organization Chart of the present invention；

Figure 13 is the structural schematic diagram of processor according to an embodiment of the invention.

Specific implementation mode

The embodiment of the present invention is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached The embodiment of figure description is exemplary, it is intended to for explaining the present invention, and is not considered as limiting the invention.

In order to solve the above-mentioned technical problem, the present invention proposes a kind of operating method of data-storage system, data storage System and processor.In an embodiment of the present invention, since the data stored in shared drive in each processor are synchronous , as long as therefore the request of same data source can be searched in shared drive, even if then secondary request is by road By can also be found in the shared drive of other processors then due to the synchronism of shared drive to other processors, because This greatly improves the utilization rate of shared drive caching.In addition, the embodiment of the present invention in shared drive according to the Kazakhstan of data Uncommon value is indexed, and is directly read internal storage data in the time complexity of (1) O by hash index, is greatly reduced business Taking for logic, can show preferable readwrite performance under limited machine, can be needed in technical costs and business It asks and tradeoff effect more more optimal than the prior art is provided.Specifically, the embodiment of the present invention is described below in conjunction with the accompanying drawings Operating method, data-storage system and the processor of data-storage system.

Fig. 1 is the flow chart of the operating method of data-storage system according to an embodiment of the invention.It needs to illustrate It is that the operating method of the data-storage system of the embodiment of the present invention can be applied in data-storage system.Wherein, the present invention's In embodiment, which may include multiple processors, wherein each processor may include shared drive, multiple places The data of shared drive storage between reason machine are mutually in step, i.e., the data of shared drive are identical in multiple processors.

As shown in Figure 1, the operating method of the data-storage system of the embodiment of the present invention may comprise steps of：

In step S101, data operation request is received.

For example, the operating method of the data-storage system of the embodiment of the present invention is applicable to distributed business system In, wherein the premise of distributed storage needs to carry out network request when being Operational Visit data, which can be business The request that policy operates business datum.It is appreciated that business side can generally carry out business datum read operation or write behaviour Make, therefore, which can be read request either write request.It should be noted that the embodiment of the present invention can needle To the difference of data operation request, read operation or the write operation of data are carried out using Different Strategies.

In step s 102, when data operation request is read request, cryptographic Hash is generated according to read request, and according to Hash Value is searched in shared drive.

When receiving data operation request, it can determine whether which kind of request is the data operation request belong to, when judging the data When operation requests are read request, corresponding cryptographic Hash can be generated according to the read request, and according to the cryptographic Hash in processor It is searched in shared drive.

It should be noted that the embodiment of the present invention is indexed data to realize by Hash strategy to shared drive, this Sample can make business directly read internal storage data in the time complexity of (1) O by hash index mode, can subtract significantly Few service logic takes.The mode that a variety of Hash can be used in the present invention is indexed, and will be carried out in subsequent embodiment detailed Thin explanation.

In step s 103, when finding corresponding data in shared drive, the data found are returned.

It should be noted that data basis storage organization strategy used by the shared drive of the embodiment of the present invention can have Many kinds, and the difference of basic storage organization strategy, can make the reality for reading corresponding data from shared drive according to read request Existing mode is also different：

As an example, when shared drive manages structure strategy using multistage Hash, structure is managed in multistage Hash Under strategy, the realization method for reading corresponding data from shared drive according to read request can be as follows：

1) corresponding cryptographic Hash is calculated by hash units according to read request；

2) it is searched stepwise in shared drive according to the cryptographic Hash, the data found is returned if finding；Otherwise it holds Row 3)；

3) failure is searched in shared drive, you can think that the shared drive is to store the data to be read, it at this time can be from It is read out in database (i.e. subordinate stores).

(it is the one of Carnegie Mellon University rejecting when shared drive is used based on CUCKOO strategies as another example The solution of kind of hash-collision) management structure when, under the management structure based on CUCKOO strategies, according to read request from being total to The realization method for enjoying reading corresponding data in memory can be as follows：

1) corresponding cryptographic Hash h1 and h2 is calculated by hash units according to read request；

2) data for searching the positions hash (Hash) value h1 and h2 in shared drive respectively, return to the lookup if finding Otherwise 3) data arrived execute；

3) failure is searched in shared drive, then can be read out from database (i.e. subordinate stores).

In step S104, when not finding corresponding data in shared drive, corresponding number is searched from database According to and return.

That is, judge data operation request for read request when, the reading can have first been searched whether from shared drive Corresponding data to be read in request, if finding the corresponding data (number i.e. to be read in shared drive According to), then return to the data found；If not finding corresponding data (data i.e. to be read) in shared drive, The corresponding data are then further searched from database and are returned.

Wherein, in an embodiment of the present invention, above-mentioned database can be regarded as the caching less than shared drive grade or deposit Storage, i.e. subordinate's storage of shared drive, which can be relational data, can also be non-relational data, as long as energy Enough realize that the purpose of subordinate's storage, the embodiment of the present invention are not specifically limited this.

To sum up, the embodiment of the present invention is by regarding the shared drive of processor as storage organization, and makes multiple processors Between shared drive storage data be mutually in step, to keep multiple processor datas consistent, farthest avoid net Network is asked, and realizes the reading speed of memory grade, enables the service request in face of high concurrent and low latency.

Fig. 2 is the flow chart of the operating method of data-storage system in accordance with another embodiment of the present invention.

In order to ensure hot spot data storage among shared drive, in an embodiment of the present invention, from data After searching corresponding data in library and returning, it can determine whether current more new strategy is reading more new strategy, if so, it is believed that It is hot spot data that corresponding data should be searched from database, this need to be searched from database corresponding data be inserted into it is shared In memory.Specifically, as shown in Fig. 2, the operating method of the data-storage system may comprise steps of：

In step s 201, data operation request is received.

In step S202, when data operation request is read request, cryptographic Hash is generated according to read request, and according to Hash Value is searched in shared drive.

In step S203, when finding corresponding data in shared drive, the data found are returned to.

In step S204, when not finding corresponding data in shared drive, corresponding number is searched from database According to and return.

In step S205, judge whether current more new strategy is to read more new strategy.

Specifically, it after finding data corresponding with read request and return in database, can determine whether currently to update Whether strategy is reading more new strategy, if so, need this being inserted into shared drive from corresponding data are searched in database, Using as hot spot data.

It should be noted that being directed to demand of the different business scene to data, the embodiment of the present invention is supported to read update and be write Update two kinds of more new strategies, and be currently specifically chosen which kind of strategy, need storing data by combination cell size, processor memory Decision after space and storage number are assessed.Wherein, the realization process of current more new strategy how to be selected to can be found in follow-up reality Apply the description of example.

In step S206, when currently more new strategy is to read more new strategy, according to the correspondence number searched from database According to the corresponding cryptographic Hash of generation and the corresponding data of lookup is inserted by shared drive according to cryptographic Hash.

It should be noted that the basic storage organization strategy of data can there are many kinds of, in this way, due to basic storage organization The difference of strategy, can make the realization method that the corresponding data that will be searched from database is inserted into shared drive also different.Under Face will provide two kinds of different examples to facilitate the understanding of those skilled in the art, but cannot be using both examples as of the invention real Apply the specific restriction of example：

Example one

When shared drive includes multi-stage data node, as shown in Fig. 3 (a), pair that above-mentioned basis is searched from database It answers data to generate corresponding cryptographic Hash and the corresponding data of lookup is inserted into the specific implementation of shared drive according to cryptographic Hash Journey may include following steps：

In step S301, corresponding first cryptographic Hash is generated according to the corresponding data searched from database, and obtain The corresponding current exponent number value of the current rank back end in place.

It is appreciated that the basic component units of shared drive are data blocks, the corresponding key-value numbers of storage in data block According to, wherein key be by business side index is provided after by hash function calculate from, wherein the type selecting of the hash function can have It is a variety of, it can be decided in its sole discretion by business side and which kind of Hash strategy selected.As an example, which can be used A kind of Murmurhash2 (hash algorithm) algorithm, in this way, can be by Murmurhash2 algorithms, pair that will be searched from database It answers data to generate corresponding first cryptographic Hash, and obtains the corresponding current exponent number value of the current rank back end in place.

In step s 302, according to the first cryptographic Hash and current rank numerical generation first index value.

Specifically, after getting the corresponding current exponent number value of the current rank back end in place, it can first determine that this is current Data number is stored corresponding to exponent number value, it later, can be by the first cryptographic Hash and storing corresponding to the current exponent number value Data number carries out modulus, and obtained remainder is as the first index value, i.e., the first index value can be calculated by following formula： Index (i)=key%s (i), wherein index (i) is the index value corresponding to current rank back end, i.e. first index value, Key is that the corresponding data searched from database generates corresponding first cryptographic Hash, and s (i), which is current rank back end, to be stored Data number.

In step S303, judge whether the back end corresponding to first index value is empty.

In step s 304, when back end is empty, the corresponding data searched from database is inserted into back end.

That is, if it is determined that the back end corresponding to first index value is sky, i.e., the back end does not have also currently There are storage data, then the corresponding data that can be searched this from database is stored in the back end corresponding to the first index value On, execution is inserted into successfully；It is no to then follow the steps S305.

In step S305, when back end is not empty, current exponent number value is updated.

In step S306, step S302-S306 is repeated according to the current exponent number value after update.

When judging the back end corresponding to first index value not for sky, i.e., data are currently housed on the back end, The current exponent number value then may be updated, for example, the current exponent number value can be added to one, and repeated according to the current exponent number value after update Step S302-S306 is executed, to judge the corresponding data searched from database whether can be stored on next rank back end.

For example, it is assumed that the corresponding data searched from database (is inserted by shared drive by hash function In data) to generate corresponding first cryptographic Hash be key, shared drive includes 25 rank back end, and i indicates the i-th rank (wherein 0< i<25), s (i) indicates that the i-th rank back end can store data number, initializes i=0, in this way, when shared drive is using multistage When Hash manages structure strategy, under the multistage Hash management structure strategy, the realization method that data are inserted into can be as follows：

A) index (i)=key%s (i) is calculated, wherein index (i) is the index value corresponding to the i-th rank back end；

If b) back end corresponding to index (i) is sky, which is stored Onto corresponding back end, execution is inserted into successfully；Otherwise it executes c)；

If c) i<24, i+1 layer is transferred to, and continue to execute and a) operate；Otherwise it executes d)；

D) there are data in the corresponding position of the 25 ranks storage in shared drive, is inserted into failure.

Wherein, it as shown in Fig. 3 (b), describes multistage Hash and manages structure for the directviewing description of shared drive, black line portion Point describe the course of reading and writing data shared drive, each level corresponding to the corresponding data searched from database the One index value finds corresponding back end, and judges whether the back end is empty, if not empty, then transfers to lower single order Back end, to the last until rank back end.As a result, under the management structure strategy based on multistage Hash, reading and writing data is most Multioperation 25 times is completed in the time complexity of O (1).

Example two

When shared drive is using management structure based on CUCKOO strategies, as shown in figure 4, above-mentioned basis is from data The corresponding data searched in library generates corresponding cryptographic Hash and the corresponding data of lookup is inserted into shared drive according to cryptographic Hash Specific implementation process may include following steps：

In step S401, corresponding second cryptographic Hash is generated according to the corresponding data searched from database and third is breathed out Uncommon value.

Specifically, cryptographic Hash meter can be carried out to the corresponding data searched from database by Cuckoo hash algorithms It calculates, to obtain corresponding two cryptographic Hash, i.e. the second cryptographic Hash and third cryptographic Hash.

In step S402, second index value and third index value are generated according to the second cryptographic Hash and third cryptographic Hash.

Specifically, the second cryptographic Hash is mapped in shared memory space to obtain corresponding second index value, and by Three cryptographic Hash are mapped in shared memory space, to obtain corresponding third index value.

In step S403, judge whether the back end corresponding to second index value and third index value is empty.

It in step s 404, will be from data when the back end corresponding to second index value or third index value is empty The corresponding data searched in library is inserted as empty back end.

That is, if back end corresponding to back end and third index value corresponding to second index value, At least one in the two is sky, then can directly be inserted as the corresponding data searched in the database on empty back end, Expression is inserted into successfully；It is no to then follow the steps S405.

In step S405, when back end corresponding to second index value and third index value is not empty, by the Data in back end corresponding to two index values or third index value are taken out, and will be from second index value or third index value The data taken out in corresponding back end are inserted into other back end of shared drive, and will be searched from database Corresponding data be inserted into the back end corresponding to second index value or third index value.

It specifically, can be by the second rope when the back end corresponding to second index value and third index value is not empty The data drawn in the corresponding back end of value are taken out, or the data in the back end corresponding to third index value are taken Go out, later, pair that the data of the taking-up are inserted on other back end of shared drive, and this is searched from database Data are answered to be inserted on the back end for being removed data.

Specifically, in one embodiment of the invention, as shown in figure 5, above-mentioned will be from second index value or third rope The specific implementation process for drawing other back end that the data taken out in the corresponding back end of value are inserted into shared drive can Including：

In step S4051, according to the number taken out from the back end corresponding to second index value or third index value The 4th index value is generated according to the 4th cryptographic Hash of generation, and according to the 4th cryptographic Hash.

Specifically, cryptographic Hash calculating can be carried out to the data that this is removed, to obtain corresponding 4th cryptographic Hash, and should 4th cryptographic Hash is mapped in shared memory space, to obtain corresponding 4th index value.It is appreciated that the above-mentioned number being removed According to can be taken out from the back end corresponding to second index value or third index value.

In step S4052, judge whether the corresponding back end of the 4th index value is empty.

It, will be from second index value or third rope when the corresponding back end of the 4th index value is empty in step S4053 Draw the data taken out in the corresponding back end of value and is inserted into the corresponding back end of the 4th index value.

That is, when judging the corresponding back end of the 4th index value for sky, the data that this is removed can be inserted into Onto the corresponding back end of the 4th index value.

In step S4054, when the corresponding back end of the 4th index value is not empty, by the corresponding number of the 4th index value It takes out according to the data of node and is inserted into again according to the data taken out from the corresponding back end of the 4th index value, and will The data taken out from the back end corresponding to second index value or third index value are inserted into the corresponding number of the 4th index value According to node.

That is, when judging the corresponding back end of the 4th index value not for sky, the 4th index value can be corresponded to The data of back end take out and be inserted into again according to the data taken out from the corresponding back end of the 4th index value, i.e., The data being removed are subjected to cryptographic Hash calculating, and generating corresponding index value according to the cryptographic Hash being calculated later can Judge whether the corresponding back end of the index value is empty, if not empty, is then handed on successively, until finding empty data section Point, and the data being finally removed are inserted on the back end of the sky.

Management structure as a result, based on CUCKOO strategies at most only needs to read 2 times, is writing few scene in face of the extreme reading more When, CUCKOO strategies more customize, and are suitable for the demand read.

It is appreciated that in view of the inherent characteristics and organizational form of shared drive, the embodiment of the present invention is supporting more new strategy While, go back support data replacement policy.Optionally, data replacement policy can be used timed task and be eliminated, business side Timestamp can be customized according to business demand, determine how long eliminate a data.As a result, timing is combined by reading more new strategy Replacement policy, can ensure will be among hot spot data storage to shared drive.

In order to ensure that the data that shared drive stores in multiple processors are mutually in step, further, the one of the present invention In a embodiment, the corresponding data of read request are being searched from database, and current more new strategy is when reading more new strategy, it can will The corresponding data of read request should be searched from database, and the shared of all processors is broadcast to by way of broadcast subscription message Memory, to realize purpose that data are mutually in step.It is subscribed to that is, notice can be broadcasted by way of triggering asynchronous message All processors of the message, update should search the corresponding data of read request from database.It is broadcasted as a result, by message queue Mode so that each processor possesses equal amount of data copy under distributed environment.

The operating method of data-storage system according to the ... of the embodiment of the present invention is searching corresponding data simultaneously from database After return, it can determine whether current more new strategy is to read more new strategy, if so, being corresponded to it is believed that should be searched from database Data be hot spot data, this need to be inserted into from corresponding data are searched in database in shared drive, can be ensured in this way Among hot spot data storage to shared drive, memory grade read or write speed is farthest realized.

Fig. 6 is the flow chart according to the operating method of the data-storage system of another embodiment of the invention.

In order to improve the availability and feasibility of the present invention, in an embodiment of the present invention, when judging that data manipulation asks Ask for write request when, data can be written into according to the write request and be written in database, and currently more new strategy be write more When new strategy, the data corresponding to write request are inserted into shared drive, the data of shared drive are updated with realizing. Specifically, as shown in fig. 6, the operating method of the data-storage system may include：

In step s 601, data operation request is received.

In step S602, when data operation request is read request, cryptographic Hash is generated according to read request, and according to Hash Value is searched in shared drive.

In step S603, when finding corresponding data in shared drive, the data found are returned to.

In step s 604, when not finding corresponding data in shared drive, corresponding number is searched from database According to and return.

In step s 605, judge whether current more new strategy is to read more new strategy.

In step S606, when currently more new strategy is to read more new strategy, according to the correspondence number searched from database According to the corresponding cryptographic Hash of generation and the corresponding data of lookup is inserted by shared drive according to cryptographic Hash.

In step S607, when data operation request is write request, database is written into the corresponding data of write request.

In step S608, judge whether current more new strategy is to write more new strategy.

In step S609, when currently more new strategy is to write more new strategy, phase is generated according to the corresponding data of write request The corresponding data of write request are simultaneously inserted into shared drive by the cryptographic Hash answered according to cryptographic Hash.

For example, if it is determined that current more new strategy is then write request receiving data operation request to write more new strategy When, while can database being written in the corresponding data of write request, the corresponding data of the write request are inserted into shared drive.

It should be noted that the basic storage organization strategy of data can there are many kinds of, in this way, due to basic storage organization The difference of strategy, can make the realization method that the corresponding data of the write request are inserted into shared drive also different.Below will Two kinds of different examples are provided to facilitate the understanding of those skilled in the art, but cannot be using both examples as the specific of the present invention It limits：

As an example, when shared drive includes multi-stage data node, as shown in fig. 7, above-mentioned basis is from write request Corresponding data generate corresponding cryptographic Hash and the corresponding data of write request are inserted into the specific of shared drive according to cryptographic Hash Realization process may include following steps：

In step s 701, corresponding 5th cryptographic Hash is generated according to the corresponding data of write request, and it is current to obtain place The corresponding current exponent number value of rank back end.

In step S702, according to the 5th cryptographic Hash and current the 5th index value of rank numerical generation.

In step S703, judge whether the back end corresponding to the 5th index value is empty.

In step S704, when back end is empty, by the corresponding insertion of data into data node of write request.

In step S705, when back end is not empty, current exponent number value is updated.

In step S706, step S702-S706 is repeated according to the current exponent number value after update.

As another example, when shared drive is using management structure based on CUCKOO strategies, such as Fig. 8 institutes Show, it is above-mentioned to generate corresponding cryptographic Hash and according to the cryptographic Hash by the write request pair according to the corresponding data of the write request The specific implementation process that the data answered are inserted into the shared drive may include following steps：

In step S801, corresponding 6th cryptographic Hash and the 7th cryptographic Hash are generated according to from the corresponding data of write request.

In step S802, the 6th index value and the 7th index value are generated according to the 6th cryptographic Hash and the 7th cryptographic Hash.

In step S803, judge whether the back end corresponding to the 6th index value and the 7th index value is empty.

In step S804, when the back end corresponding to the 6th index value or the 7th index value is empty, by write request Corresponding data are inserted as empty back end.

That is, if back end corresponding to back end and the 7th index value corresponding to the 6th index value, At least one in the two is sky, then can directly be inserted as the corresponding data of the write request on empty back end, indicates to insert Enter success；It is no to then follow the steps S405.

In step S805, when back end corresponding to the 6th index value and the 7th index value is not empty, by the Data in back end corresponding to six index values or the 7th index value are taken out, and will be from the 6th index value or the 7th index value The data taken out in corresponding back end are inserted into other back end of shared drive, and by the corresponding number of write request According to the back end being inserted into corresponding to the 6th index value or the 7th index value.

It specifically, can be by the 6th if the back end corresponding to the 6th index value and the 7th index value is not sky Data in back end corresponding to index value are taken out, or the data in the back end corresponding to the 7th index value are taken Go out, later, the data of the taking-up are inserted on other back end of shared drive, and the corresponding data of the write request are inserted Enter to being removed on the back end of data.

In order to improve the present invention availability and feasibility, need storing data by combination cell size, processor it is interior Deposit space and storage number assessed after, come decision currently update policy selection which kind of strategy.Further, the present invention's In one embodiment, as shown in figure 9, the operating method of the data-storage system may also include：

In step S901, total amount of data value in shared drive is obtained.

During being written and read operation in shared drive, total amount of data value in the shared drive can be also obtained.For example, It is data-storable the shared drive can be estimated according to data unit size and the assignable shared memory space of processor Total quantity n_max.

In step S902, the clearance time interval of shared drive is obtained, and demand number is generated according to clearance time interval According to magnitude.

For example, can business scenario demand, assess the clearance time interval t of shared drive, and estimate between clearance time Every the space size needed for full dose data in t, i.e., above-mentioned demand data magnitude n_data.It later, can will be in the shared drive Total amount of data value n_max and demand data magnitude n_data carries out size and compares.Wherein, which can be regarded as business field All business datums under scape.

It is to write more by current update strategy setting when demand data magnitude is less than total amount of data value in step S903 New strategy.

If demand data magnitude n_data is less than total amount of data value n_max, show that shared drive is enough to support that this is complete Data storage is measured, then optional more new strategy of writing is proper at this time, in this way, (i.e. data can be stored write grade every time Library) when be synchronized in the shared drive of processing.

In step S904, when demand data magnitude is greater than or equal to total amount of data value, strategy setting will be currently updated To read more new strategy.

If demand data magnitude n_data is greater than or equal to total amount of data value n_max, show shared drive as portion The caching of divided data then selecting reading more new strategy proper at this time, and combines the time interval t of cleaning, according to business field Cleaning condition is arranged in scape, after multiple rounds, it is ensured that the hot spot data of needs is cached in shared drive, by cold data Gradually cleared up.

In order to ensure that the data that shared drive stores in multiple processors are mutually in step, further, the one of the present invention In a embodiment, it is write request and when current more new strategy is to write more new strategy receiving data operation request, this can be write Corresponding data are asked to be broadcast to the shared drive of all processors by way of broadcast subscription message, to realize that data are mutual Synchronous purpose.

The operating method of data-storage system according to the ... of the embodiment of the present invention is judging data operation request for write request When, data can be written into according to the write request and be written in database, and when currently more new strategy is to write more new strategy, it will Data corresponding to write request are inserted into shared drive, are updated to the data of shared drive with realizing, are improved this hair Bright availability and feasibility, and by reading to update and write more new strategy, constantly update the data in shared drive so that heat Point data is cached in shared drive, is ensured that the operating time of most read requests is rdma read time stage, is propped up to greatest extent The concurrency asked under low latency response demand is held.

It, below can will be of the invention in order to enable those skilled in the art can clearly understand the features of the present invention The operating method of the data-storage system of embodiment is applied in two kinds of different business scenes and is described.

For example, the operating method of the data-storage system of the embodiment of the present invention is applied in instant communication client In the business scenario of message sequence number read-write service, wherein can be in the application scenarios of message sequence number read-write service：User opens When telecommunication customer end arbitrary chat group chat window, can all ask one nearest chat messages of chat group of this service acquisition Serial number, and the chat record after the serial number is handed down in the instant communication client of user.Based on the application scenarios, i.e., When telecommunication customer end corresponding to server can receive the instant communication client transmission data operation request, the data manipulation The serial number that nearest chat messages are contained in request reads the chat record after the serial number with request, it can be seen that the data Operation requests are read request, it will be appreciated that, which is equal to the group chat skylight opening that user opens instant communication client Number it is consistent.As shown in Figure 10 (a), judge the operation requests for read request when, can according to the read request generate cryptographic Hash, And all message in the chat record after the serial number are searched in shared drive according to the cryptographic Hash, if being looked in shared drive To these data, then these data are returned into instant communication client, in this way, instant communication client can will be after the serial number Chat record in be presented to user.If not finding these data in shared drive, these can be searched from database Data, and when currently more new strategy is to read more new strategy, will be inserted into shared drive from the data found in database.

For another example, by the operating method of the data-storage system of the embodiment of the present invention be applied to instant communication client in group at In the business scenario of member's title, wherein can be in the application scenarios of group members title business：Instant messaging client is opened in user In end when the window of group chat, it can be seen that the group members of history speech and user make a speech every time in group chat skylight opening, this is It, all can be from obtaining group title of the primary speech member in this chat group because when being made a speech user or group members, it can be with Solution, the request magnitude are consistent with group's message magnitude.As shown in Figure 10 (b), group's title for speech user can receive Obtain request, it is possible to determine that the request is read request, can generate cryptographic Hash according to the read request at this time, and exist according to the cryptographic Hash The data (such as cotton clothes Jie) corresponding to group's title of the user are searched in shared drive, if finding Jie in shared drive The data can then be returned to instant communication client, in this way, instant communication client can be in group chat by group's title data of cotton clothes At position corresponding to its user, group's title of cotton clothes Jie is shown.

To sum up, the business of the platform in the majority in above-mentioned this magnanimity business, many scenes has high concurrent request scene More harsh index request, the operating method of data-storage system provided in an embodiment of the present invention can solve many similar The problem of scene, it is that can significantly decrease technical costs to bring most intuitive income effect, is completed with less machine Higher-quality service.

A kind of embodiment corresponding, of the invention with the operating method of data-storage system of above-mentioned several embodiments offers A kind of data-storage system is also provided, since data-storage system provided in an embodiment of the present invention and above-mentioned several embodiments provide Data-storage system operating method it is corresponding, therefore also fitted in the embodiment of the operating method of aforementioned data storage system For data-storage system provided in this embodiment, it is not described in detail in the present embodiment.Figure 11 is according to one of the invention The structural schematic diagram of the data-storage system of embodiment.As shown in figure 11, which may include：Multiple places Reason machine 100.Wherein, as shown in figure 11, each processor 100 may include shared drive 110, request receiving module 120, operation mould Block 130 and synchronization module 140.

Specifically, request receiving module 120 is for receiving data operation request.

Operation module 130 is used to, when data operation request is read request, generate cryptographic Hash according to read request, and according to Kazakhstan Uncommon value is searched in shared drive 110, and the number found is returned when finding corresponding data in shared drive 110 According to, and when not finding corresponding data in shared drive 110, corresponding data are searched from database and are returned.

Synchronization module 140 is used to the shared drive 110 in other processors carry out the data in shared drive 110 together Step.

In order to ensure that the data that shared drive stores in multiple processors are mutually in step, further, the one of the present invention In a embodiment, operation module 130 is additionally operable to：After searching corresponding data and return in database, judge currently more Whether new strategy is reading more new strategy, and when currently more new strategy is to read more new strategy, according to pair searched from database It answers data to generate corresponding cryptographic Hash, and the corresponding data of lookup is inserted by shared drive according to cryptographic Hash.

As an example, when shared drive includes multi-stage data node, operation module 130 is looked into according to from database The corresponding data looked for generates corresponding cryptographic Hash and the corresponding data of lookup is inserted into the specific of shared drive according to cryptographic Hash Realization process can be as follows：

S11 corresponding first cryptographic Hash) is generated according to the corresponding data searched from database, and obtains the current rank in place The corresponding current exponent number value of back end；

S12) according to the first cryptographic Hash and current rank numerical generation first index value；

S13) judge whether the back end corresponding to first index value is empty；

S14) when back end is empty, the corresponding data searched from database is inserted into back end；

S15) when back end is not empty, current exponent number value is updated；

S16 S12-S16) is repeated according to the current exponent number value after update.

As another example, when shared drive is using management structure based on CUCKOO strategies, operation module 130 generate corresponding cryptographic Hash and are inserted the corresponding data of lookup according to cryptographic Hash according to the corresponding data searched from database The specific implementation process entered to shared drive can be as follows：

S21 corresponding second cryptographic Hash and third cryptographic Hash) are generated according to the corresponding data searched from database；

S22 second index value and third index value) are generated according to the second cryptographic Hash and third cryptographic Hash；

S23) judge whether the back end corresponding to second index value and third index value is empty；

S24 when) back end corresponding to second index value or third index value is empty, by what is searched from database Corresponding data is inserted as empty back end；

S25 when) back end corresponding to second index value and third index value is not empty, by second index value or Data in back end corresponding to third index value are taken out, and by the number corresponding to the second index value or third index value Other back end of shared drive are inserted into according to the data taken out in node, and the corresponding data that will be searched from database It is inserted into the back end corresponding to second index value or third index value.

In an embodiment of the present invention, operation module 130 is by the data corresponding to the second index value or third index value The specific implementation process that the data taken out in node are inserted into other back end of shared drive can be as follows：

S251) the 4th is generated according to the data taken out from the back end corresponding to second index value or third index value Cryptographic Hash, and the 4th index value is generated according to the 4th cryptographic Hash；

S252) judge whether the corresponding back end of the 4th index value is empty；

It S253, will be right from second index value or third index value institute) when the corresponding back end of the 4th index value is empty The data taken out in the back end answered are inserted into the corresponding back end of the 4th index value；

S254) when the corresponding back end of the 4th index value is not empty, by the corresponding back end of the 4th index value Data are taken out and are inserted into again according to the data taken out from the corresponding back end of the 4th index value, and will be from the second rope Draw the data taken out in the back end corresponding to value or third index value and is inserted into the corresponding back end of the 4th index value.

In order to improve the availability and feasibility of the present invention, further, in one embodiment of the invention, operation Module 130 is additionally operable to：When data operation request is write request, database is written into the corresponding data of write request, and judging Current more new strategy is to generate corresponding cryptographic Hash when writing more new strategy according to the corresponding data of write request and incited somebody to action according to cryptographic Hash The corresponding data of write request are inserted into shared drive 110.

As an example, when shared drive includes multi-stage data node, operation module 130 is corresponding according to write request Data generate corresponding cryptographic Hash and the corresponding data of write request are inserted into the specific implementation of shared drive 110 according to cryptographic Hash Process can be as follows：

S31 corresponding 5th cryptographic Hash) is generated according to the corresponding data of write request, and obtains the current rank back end in place Corresponding current exponent number value；

S32) according to the 5th cryptographic Hash and current the 5th index value of rank numerical generation；

S33) judge whether the back end corresponding to the 5th index value is empty；

S34) when back end is empty, by the corresponding insertion of data into data node of write request；

S35) when back end is not empty, current exponent number value is updated；

S36 step S32-S36) is repeated according to the current exponent number value after update.

As another example, when shared drive is using management structure based on CUCKOO strategies, operation module 130 generate corresponding cryptographic Hash according to the corresponding data of write request and are inserted into the corresponding data of write request altogether according to cryptographic Hash The specific implementation process for enjoying memory 110 can be as follows：

S41 corresponding 6th cryptographic Hash and the 7th cryptographic Hash) are generated according to from the corresponding data of write request；

S42 the 6th index value and the 7th index value) are generated according to the 6th cryptographic Hash and the 7th cryptographic Hash；

S43) judge whether the back end corresponding to the 6th index value and the 7th index value is empty；

S44 when) back end corresponding to the 6th index value or the 7th index value is empty, by the corresponding data of write request It is inserted as empty back end；

S45 when) back end corresponding to the 6th index value and the 7th index value is not empty, by the 6th index value or Data in back end corresponding to 7th index value are taken out, and by the number corresponding to the 6th index value or the 7th index value It is inserted into other back end of shared drive according to the data taken out in node, and the corresponding data of write request are inserted into Back end corresponding to six index values or the 7th index value.

In order to improve the present invention availability and feasibility, need storing data by combination cell size, processor it is interior Deposit space and storage number assessed after, come decision currently update policy selection which kind of strategy.Further, the present invention's In one embodiment, operation module 130 is additionally operable to：Total amount of data value in shared drive 110 is obtained, and obtains shared drive 110 Clearance time interval, and according to clearance time interval generate demand data magnitude, and demand data magnitude be less than data it is total It is to write more new strategy, and be greater than or equal to total amount of data value in demand data magnitude by current update strategy setting when magnitude When, it is to read more new strategy that will currently update strategy setting.

In order to ensure that the data that shared drive stores in multiple processors are mutually in step, further, the one of the present invention In a embodiment, the shared drive 100 in multiple processors 100 can be synchronized by way of broadcast subscription message.

Figure 12 is the major technique Organization Chart of the present invention.As shown in figure 12, flow (1) (2) (3) describes the mistake of read request Journey, i.e.,：When receiving the read request that business platform is sent by shared drive api interface, can be read into shared drive should Corresponding data in read request, if there are the data in the shared drive, having executed (1) (2) (3), flow terminates afterwards； Otherwise, if current more new strategy executes flow (4) (5) (6) using more new strategy is read, landing storage assembly is read, It i.e. after reading the data in database, is broadcasted by message queue, the number that will should be read from database According to being synchronized among the shared drive of all processors.As shown in figure 12, flow (1) (4) describes the process of write request, i.e.,： It, directly will be in the data write-in landing storage assembly (i.e. database) in the write request when receiving write request；If current More new strategy then executing the process of flow (5) (6) (3), i.e., is broadcasted each using more new strategy is write after notification message queue A processor and the data for updating shared drive, i.e., will be in the data update to shared drive in write request.

In order to realize above-described embodiment, the invention also provides a kind of processors.

Figure 13 is the structural schematic diagram of processor according to an embodiment of the invention.As shown in figure 13, the processor 100 It may include：Shell 150, processor 160, memory 170, shared drive 110, circuit board 180 and power circuit 190, wherein electricity Road plate 160 is placed in the space interior that shell 150 surrounds, and processor 160, memory 170 and shared drive 110 are arranged in circuit On plate 160；Power circuit 190, for being each circuit or the device power supply of processor；Memory 170 is executable for storing Program code；The data that shared drive 110 is used to store memory 170 cache；Processor 160 is by reading memory The executable program code stored in 170 runs program corresponding with executable program code, for executing following steps：

Receive data operation request；

When data operation request is read request, cryptographic Hash is generated according to read request, and according to cryptographic Hash in shared drive In searched；

When finding corresponding data in shared drive, the data found are returned；

When not finding corresponding data in shared drive, corresponding data are searched from database and are returned.

The processor of the embodiment of the present invention judges which the data operation request belongs to when receiving data operation request Kind request type generates cryptographic Hash and according to cryptographic Hash altogether when the data operation request is read request according to the read request It enjoys in memory and being searched, if finding, return to the data found, corresponding data are searched in otherwise being stored from subordinate and return It returns.In an embodiment of the present invention, since the data stored in shared drive in each processor are synchronous, as long as same The request of one data source can be searched in shared drive, even if then secondary request is routed to other processing Machine can also be found, therefore greatly improve then due to the synchronism of shared drive in the shared drive of other processors The utilization rate of shared drive caching.In addition, the embodiment of the present invention is indexed in shared drive according to the cryptographic Hash of data, adopt With data are indexed in the time complexity of hash index mode O (1), to ensure that the operating time of most read requests is Rdma read time stage supports the concurrency asked under low latency response demand to greatest extent, and significantly reduce technology at This budget avoids the consumption of unnecessary high-performance hardware machine.

In order to realize above-described embodiment, the invention also provides a kind of storage mediums, for storing application program, using journey Sequence is used to execute the operating method of the data-storage system described in embodiment any one of of the invention.

In the description of the present invention, it is to be understood that, term " first ", " second " are used for description purposes only, and cannot It is interpreted as indicating or implies relative importance or implicitly indicate the quantity of indicated technical characteristic.Define as a result, " the One ", the feature of " second " can explicitly or implicitly include one or more this feature.In the description of the present invention, The meaning of " plurality " is two or more, unless otherwise specifically defined.

In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show The description of example " or " some examples " etc. means that specific features described in conjunction with this embodiment or example or feature are contained in this In at least one embodiment or example of invention.In the present specification, schematic expression of the above terms are necessarily directed to It is identical embodiment or example.Moreover, the specific features or feature of description in any one or more embodiments or can show It can be combined in any suitable manner in example.In addition, without conflicting with each other, those skilled in the art can be by this specification Described in different embodiments or examples and the features of different embodiments or examples be combined.

Any process described otherwise above or method description are construed as in flow chart or herein, and expression includes It is one or more for realizing specific logical function or process the step of executable instruction code module, segment or portion Point, and the range of the preferred embodiment of the present invention includes other realization, wherein can not press shown or discuss suitable Sequence, include according to involved function by it is basic simultaneously in the way of or in the opposite order, to execute function, this should be of the invention Embodiment person of ordinary skill in the field understood.

Expression or logic and/or step described otherwise above herein in flow charts, for example, being considered use In the order list for the executable instruction for realizing logic function, may be embodied in any computer-readable medium, for Instruction execution system, device or equipment (system of such as computer based system including processor or other can be held from instruction The instruction fetch of row system, device or equipment and the system executed instruction) it uses, or combine these instruction execution systems, device or set It is standby and use.For the purpose of this specification, " computer-readable medium " can any can be included, store, communicating, propagating or passing Defeated program is for instruction execution system, device or equipment or the dress used in conjunction with these instruction execution systems, device or equipment It sets.The more specific example (non-exhaustive list) of computer-readable medium includes following：Electricity with one or more wiring Interconnecting piece (electronic device), portable computer diskette box (magnetic device), random access memory (RAM), read-only memory (ROM), erasable edit read-only storage (EPROM or flash memory), fiber device and portable optic disk is read-only deposits Reservoir (CDROM).In addition, computer-readable medium can even is that the paper that can print described program on it or other are suitable Medium, because can be for example by carrying out optical scanner to paper or other media, then into edlin, interpretation or when necessary with it His suitable method is handled electronically to obtain described program, is then stored in computer storage.

It should be appreciated that each section of the present invention can be realized with hardware, software, firmware or combination thereof.Above-mentioned In embodiment, software that multiple steps or method can in memory and by suitable instruction execution system be executed with storage Or firmware is realized.It, and in another embodiment, can be under well known in the art for example, if realized with hardware Any one of row technology or their combination are realized：With the logic gates for realizing logic function to data-signal Discrete logic, with suitable combinational logic gate circuit application-specific integrated circuit, programmable gate array (PGA), scene Programmable gate array (FPGA) etc..

Those skilled in the art are appreciated that realize all or part of step that above-described embodiment method carries Suddenly it is that relevant hardware can be instructed to complete by program, the program can be stored in a kind of computer-readable storage medium In matter, which includes the steps that one or a combination set of embodiment of the method when being executed.

In addition, each functional unit in each embodiment of the present invention can be integrated in a processing module, it can also That each unit physically exists alone, can also two or more units be integrated in a module.Above-mentioned integrated mould The form that hardware had both may be used in block is realized, can also be realized in the form of software function module.The integrated module is such as Fruit is realized in the form of software function module and when sold or used as an independent product, can also be stored in a computer In read/write memory medium.

Storage medium mentioned above can be read-only memory, disk or CD etc..Although having been shown and retouching above The embodiment of the present invention is stated, it is to be understood that above-described embodiment is exemplary, and should not be understood as the limit to the present invention System, those skilled in the art can be changed above-described embodiment, change, replace and become within the scope of the invention Type.

Claims

1. a kind of operating method of data-storage system, which is characterized in that the data-storage system includes multiple processors, In, each processor includes shared drive, and the data of the shared drive storage between the multiple processor are mutually in step, described Method includes the following steps：

Receive data operation request；

When the data operation request is read request, cryptographic Hash is generated according to the read request, and exist according to the cryptographic Hash It is searched in the shared drive；

2. the operating method of data-storage system as described in claim 1, which is characterized in that searched from database described After corresponding data and return, further include：

When currently more new strategy is to read more new strategy, corresponding breathe out is generated according to the corresponding data searched from database The corresponding data of the lookup is simultaneously inserted into the shared drive by uncommon value according to the cryptographic Hash.

3. the operating method of data-storage system as claimed in claim 2, which is characterized in that the shared drive includes multistage Back end, wherein the corresponding data searched from database described in the basis generates corresponding cryptographic Hash and according to described The corresponding data of the lookup is inserted into the shared drive by cryptographic Hash, is specifically included：

S11, the corresponding data searched from database according to generate corresponding first cryptographic Hash, and obtain the current rank in place The corresponding current exponent number value of back end；

When S13, the back end corresponding to the first index value are empty, by the corresponding data searched from database It is inserted into the back end；

When S14, the back end corresponding to the first index value are not empty, the current exponent number value is updated；

4. the operating method of data-storage system as claimed in claim 2, which is characterized in that from database described in the basis The corresponding data of middle lookup generates corresponding cryptographic Hash and the corresponding data of the lookup is inserted into institute according to the cryptographic Hash Shared drive is stated, is specifically included：

It, will be described from database when S23, the back end corresponding to the second index value or the third index value are empty The corresponding data of middle lookup is inserted as the empty back end；

It, will be described when S24, the back end corresponding to the second index value and the third index value are not empty Data in the back end corresponding to second index value or the third index value are taken out, and will be from second index The data taken out in the back end corresponding to value or the third index value are inserted into other numbers of the shared drive It is inserted into the second index value or the third index value according to node, and by the corresponding data searched from database The corresponding back end.

5. the operating method of data-storage system as claimed in claim 4, which is characterized in that described to be indexed from described second The data taken out in the back end corresponding to value or the third index value are inserted into other numbers of the shared drive It is specifically included according to node：

The data that S251, basis are taken out from the back end corresponding to the second index value or the third index value The 4th cryptographic Hash is generated, and the 4th index value is generated according to the 4th cryptographic Hash；

S252, when the corresponding back end of the 4th index value is empty, will be from the second index value or the third rope Draw the data taken out in the corresponding back end of value and is inserted into the corresponding back end of the 4th index value；

S253, when the corresponding back end of the 4th index value is not empty, by the corresponding data section of the 4th index value The data of point are taken out and are inserted into again according to the data taken out from the corresponding back end of the 4th index value, and will The data taken out from the back end corresponding to the second index value or the third index value are inserted into described The corresponding back end of four index values.

6. the operating method of data-storage system as described in any one in claim 1-5, which is characterized in that further include：

When currently more new strategy is to write more new strategy, corresponding cryptographic Hash and root are generated according to the corresponding data of the write request The corresponding data of the write request are inserted into the shared drive according to the cryptographic Hash.

7. the operating method of data-storage system as claimed in claim 6, which is characterized in that the shared drive includes multistage Back end, wherein from the corresponding cryptographic Hash of the corresponding data generation of the write request and according to the Kazakhstan described in the basis The corresponding data of the write request are inserted into the shared drive by uncommon value, are specifically included：

S31, corresponding 5th cryptographic Hash is generated according to the corresponding data of the write request, and obtains the current rank back end in place Corresponding current exponent number value；

When S33, the back end corresponding to the 5th index value are empty, described in the corresponding data insertion of the write request Back end；

8. the operating method of data-storage system as claimed in claim 6, which is characterized in that described according to the write request pair The data answered generate corresponding cryptographic Hash and the corresponding data of the write request are inserted into described share according to the cryptographic Hash Memory specifically includes：

When S43, the back end corresponding to the 6th index value or the 7th index value are empty, by the write request pair The data answered are inserted as the empty back end；

It, will be described when S44, the back end corresponding to the 6th index value and the 7th index value are not empty Data in the back end corresponding to 6th index value or the 7th index value are taken out, and will be from the 6th index The data taken out in the back end corresponding to value or the 7th index value are inserted into other numbers of the shared drive It is inserted into corresponding to the 6th index value or the 7th index value according to node, and by the corresponding data of the write request The back end.

9. the operating method of data-storage system as claimed in claim 6, which is characterized in that further include：

Obtain total amount of data value in the shared drive；

The clearance time interval of the shared drive is obtained, and demand data magnitude is generated according to the clearance time interval；

It is to write update plan by the current update strategy setting when the demand data magnitude is less than the total amount of data value Slightly；And

It is to read more by the current update strategy setting when the demand data magnitude is greater than or equal to the total amount of data value New strategy.

10. the operating method of data-storage system as described in claim 1, which is characterized in that in the multiple processor The shared drive is synchronized by way of broadcast subscription message.

11. a kind of data-storage system, which is characterized in that including multiple processors, wherein each the processor includes：

Shared drive；

Request receiving module, for receiving data operation request；

Operation module is used for when the data operation request is read request, according to read request generation cryptographic Hash, and according to The cryptographic Hash is searched in the shared drive, is returned and is searched when finding corresponding data in the shared drive The data arrived, and when not finding corresponding data in the shared drive, corresponding data are searched from database and are returned It returns；And

12. data-storage system as claimed in claim 11, which is characterized in that the operation module is additionally operable to：From data After searching corresponding data in library and returning, when currently more new strategy is to read more new strategy, according to described from database The corresponding data of lookup generates corresponding cryptographic Hash, and according to the cryptographic Hash is inserted into the corresponding data of the lookup described Shared drive.

13. data-storage system as claimed in claim 12, which is characterized in that the shared drive includes multi-stage data section Point, wherein the operation module is specifically used for：

S13, when the back end is empty, the corresponding data searched from database is inserted into the back end；

14. data-storage system as claimed in claim 12, which is characterized in that the operation module is specifically used for：

When back end corresponding to the second index value or the third index value is empty, looked into described from database The corresponding data looked for is inserted as the empty back end；

When the back end corresponding to the second index value and the third index value is not empty, by described second Data in the back end corresponding to index value or the third index value are taken out, and will from the second index value or The data taken out in the back end corresponding to the third index value are inserted into other data sections of the shared drive Point, and by the corresponding data searched from database be inserted into the second index value or the third index value institute it is right The back end answered.

15. such as claim 11-14 any one of them data-storage systems, which is characterized in that the operation module is additionally operable to： When the data operation request is write request, the corresponding data of the write request are written into the database, and currently more New strategy is to generate corresponding cryptographic Hash according to the corresponding data of the write request and according to the cryptographic Hash when writing more new strategy The corresponding data of the write request are inserted into the shared drive.