CN106354722A - Message processing method and device for stream-oriented computation system - Google Patents

Message processing method and device for stream-oriented computation system Download PDF

Info

Publication number
CN106354722A
CN106354722A CN201510413095.4A CN201510413095A CN106354722A CN 106354722 A CN106354722 A CN 106354722A CN 201510413095 A CN201510413095 A CN 201510413095A CN 106354722 A CN106354722 A CN 106354722A
Authority
CN
China
Prior art keywords
message
write
memory element
unit
service processing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201510413095.4A
Other languages
Chinese (zh)
Other versions
CN106354722B (en
Inventor
陈昱
刘键
封仲淹
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Alibaba Group Holding Ltd
Original Assignee
Alibaba Group Holding Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Alibaba Group Holding Ltd filed Critical Alibaba Group Holding Ltd
Priority to CN201510413095.4A priority Critical patent/CN106354722B/en
Publication of CN106354722A publication Critical patent/CN106354722A/en
Application granted granted Critical
Publication of CN106354722B publication Critical patent/CN106354722B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00Error detection; Error correction; Monitoring
    • G06F11/07Responding to the occurrence of a fault, e.g. fault tolerance
    • G06F11/14Error detection or correction of the data by redundancy in operation
    • G06F11/1402Saving, restoring, recovering or retrying
    • G06F11/1446Point-in-time backing up or restoration of persistent data
    • G06F11/1448Management of the data involved in backup or backup restore
    • G06F11/1453Management of the data involved in backup or backup restore using de-duplication of the data
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/10File systems; File servers
    • G06F16/17Details of further file system functions
    • G06F16/174Redundancy elimination performed by the file system
    • G06F16/1748De-duplication implemented within the file system, e.g. based on file segments
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00Traffic control in data switching networks
    • H04L47/50Queue scheduling
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/50Network services
    • H04L67/56Provisioning of proxy services
    • H04L67/568Storing data temporarily at an intermediate stage, e.g. caching

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Quality & Reliability (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Computer And Data Communications (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention provides a message processing method and device for a stream-oriented computation system. The method comprises the steps that 1, a message is received; 2, whether the message is stored in a first storage unit or not is judged, if yes, the message is not processed, and if not, the step 3 is executed; 3, whether a business processing result corresponding to the message exists in a second storage unit or not is judged, if yes, the step 6 is executed, and if not, the step 4 is executed; 4, business processing is conducted on the message; 5, the message and the business processing result thereof are written in the second storage unit; 6, the message is written in the first storage unit. Accordingly, it can be guaranteed that business processing is only conducted on the message in the stream-oriented computation system once, and the performance and the efficiency of the system are improved.

Description

A kind of message treatment method of streaming computing system and device
[technical field]
The present invention relates to Computer Applied Technology field, at a kind of message of streaming computing system Reason method and apparatus.
[background technology]
Jstorm is the real-time streaming Computational frame with reference to storm, in network io, threading model, resource Do sustained improvement in scheduling, availability and stability, used by more and more enterprises.Streaming calculates The rank of the current Message Processing in field is generally and processes at least one times with process at most once, needs a kind of letter badly Single efficient mode is ensureing that message is only processed once exactly.
Jstrom has been provided for acker mechanism to realize at least processing Message Processing circle once in itself Not it is therefore desirable to the key problem solving is, how repeated data is carried out with duplicate removal, that is, for processing Message no longer processed.
[content of the invention]
In view of this, the invention provides a kind of message treatment method of streaming computing system and device, with Be easy to reduce to message repeat process.
Concrete technical scheme is as follows:
The invention provides a kind of message treatment method of Stream Processing system, the method includes:
S1, reception message;
Whether s2, judge to be stored with the first memory element described message, if it is, not to this message Processed;Otherwise, execute s3;
S3, judge in the second memory element, whether there is the corresponding service processing result of described message, such as Fruit is to execute s6, otherwise executes s4;
S4, Business Processing is carried out to described message;
S5, described message and its service processing result are write the second memory element;
S6, by described message write the first memory element.
According to the present invention one preferred implementation, the content in described first memory element is mapped to the grand filtration of cloth Device bloomfilter;
Also included before described s2: judge whether described message hits bloomfilter, if not, execution Described s3;If it is, executing described s2.
According to the present invention one preferred implementation, when executing described s4, if in described second memory element There is the messaging list with described message with identical calculations mark, then using with described message, there is phase With the corresponding service processing result of messaging list calculating mark, Business Processing is carried out to described message;
In described s5, described message is added described messaging list, and the service processing result using s4 Update the corresponding service processing result of messaging list described in described second memory element;
When executing described s6, the described messaging list comprising described message is write the first memory element.
According to the present invention one preferred implementation, also include between described s4 and described s5: disappear described Breath and its corresponding service processing result put into write queue;
Using the content execution described s5 and described s6 in said write queue.
According to the present invention one preferred implementation, execute described s5 and institute using the content in said write queue State s6 to include:
S51, from said write queue take out a cache node, comprise message in one cache node And its corresponding service processing result;
S52, by described cache node content write described second memory element;
S53, the message in described cache node is put into duplicate removal storage write queue;
S54, store write queue from duplicate removal and take out message write the first memory element.
According to the present invention one preferred implementation, also include after described s2 and between described s3:
Judge whether there is described message, such as in said write queue and the storage write queue of described duplicate removal Fruit is then not carry out Business Processing to described message, otherwise continues executing with described s3.
According to the present invention one preferred implementation, also included before described s51 and described s52:
Judge whether there is the comprised message tool with described cache node in described duplicate removal storage write queue There is the message of identical calculations mark, if it is, described cache node is put back to said write queue, wait Re-execute described s51;Otherwise, continue executing with described s52.
Present invention also offers a kind of message processing apparatus of streaming computing system, this device includes:
Receiving unit, for receiving message;
First judging unit, for the described message that judges whether to be stored with the first memory element, if Judged result is yes, no longer described message is processed;
Second judging unit, for when the judged result of described first judging unit is no, judging described the Whether there is the corresponding service processing result of described message in two memory element;
Processing unit, for when the judged result of described second judging unit is for being, not entering to described message Row Business Processing;When the judged result of described second judging unit is no, described message is carried out at business Reason;
First writing unit, for writing the second memory element by the result of described message;
Second writing unit, for, after described first writing unit completes write, or sentencing described second When the judged result of disconnected unit is no, described message is write the first memory element.
According to the present invention one preferred implementation, the content in described first memory element is mapped to bloomfilter;
This device also includes: the 3rd judging unit, for after described receiving unit receives described message, Judge whether described message hits bloomfilter, if not, triggering described second judging unit to execute judgement Operation;If it is, the operation that the described first judging unit execution of triggering judges.
According to the present invention one preferred implementation, described processing unit, if single specifically for the described second storage There is the messaging list with described message with identical calculations mark in unit, then had using with described message There is the corresponding service processing result of messaging list of identical calculations mark, Business Processing is carried out to described message;
Described first writing unit, specifically for adding described messaging list by described message, and using described This service processing result of processing unit updates the corresponding industry of messaging list described in described second memory element Business result;
Described second writing unit, specifically for depositing the described messaging list write first comprising described message Storage unit.
According to the present invention one preferred implementation, described processing unit, it is additionally operable to described message and its correspondence Service processing result put into write queue;
Described first writing unit and described second writing unit are write using the content execution in said write queue The operation entering.
According to the present invention one preferred implementation, described first writing unit, specifically for from said write queue One cache node of middle taking-up, comprises message and its corresponding service processing result in one cache node, Content in described cache node is write described second memory element, the message in described cache node is put Enter duplicate removal storage write queue;
Described second writing unit, deposits specifically for storing to take out message in write queue and write first from duplicate removal Storage unit.
According to the present invention one preferred implementation, this device also includes:
4th judging unit, for when the judged result of described first judging unit is no, judging said write Queue and described go to re-write in queue whether there is described message, if it is, not to described message Carry out Business Processing;Otherwise, the operation that the described second judging unit execution of triggering judges.
According to the present invention one preferred implementation, described first writing unit is by the content in described cache node Before writing described second memory element, it is additionally operable to judge whether deposited in described duplicate removal storage write queue There is the message of identical calculations mark in message comprised with described cache node, if it is, will be described slow Deposit node and put back to said write queue, wait retrieves;Otherwise, continue executing with described cache node Content write described second memory element operation.
As can be seen from the above technical solutions, message is single together with result corresponding write the second storage Unit, even if message write the first memory element failure, when receiving identical message next time, due to second There is the corresponding service processing result of this message it is not necessary to again industry be entered to this message in memory element Business process, only by this message write the first memory element can it is ensured that message only by Business Processing once, Improve systematic function and efficiency.
[brief description]
Fig. 1 is a process chart provided in an embodiment of the present invention;
Fig. 2 is the flow chart of message treatment method provided in an embodiment of the present invention;
Fig. 3 is the write flow process of calculating storaging unit provided in an embodiment of the present invention and duplicate removal memory element Figure;
Fig. 4 is the structure chart of message processing apparatus provided in an embodiment of the present invention.
[specific embodiment]
In order that the object, technical solutions and advantages of the present invention are clearer, below in conjunction with the accompanying drawings and specifically Embodiment describes the present invention.
In the present invention using two memory element: calculating storaging unit (the second memory element) and duplicate removal Memory element (the first memory element).Wherein, calculating storaging unit is used for storing to the business of message Reason result, duplicate removal memory element storage carried out the message of Business Processing, and was used for doing duplicate removal to avoid message Repetition Business Processing.
Upon receiving a message, handling process can as shown in fig. 1, including below scheme:
Whether in 101, judge to be stored with duplicate removal memory element the message that this receives, if It is no longer this message to be processed;Otherwise, 102 are executed.
In 102, Business Processing is carried out to this message.
The Business Processing that message is carried out being related in embodiments of the present invention can be needed by specific business Ask to determine, history result may be utilized, if necessary to using history result, then in industry When business is processed, history result can be obtained first from calculating storaging unit, then using acquisition History result is processed to this message.Under some business demands, at business that message is carried out Reason can be the calculating being carried out using message, for example, when message 2 is that blanket order amount to a class user is entered Row statistics, if there is the blanket order amount that message 1 is also to a class user before this counted, that The calculating knot of message 1 is obtained for message 2 when being calculated it is possible to from calculating storaging unit Really, obtain message after this result of calculation being added with the order volume of a class user comprising in message 2 2 result of calculation.
In 103, service processing result is write calculating storaging unit.
In 104, this message is write duplicate removal memory element.
When there is a new message, again go to 102 to start to execute.
Said method can reduce to a certain extent to message repeat process, but if execution 103 Afterwards, system faults such as machine of delaying are occurred to lead to 104 to execute failure when 104, this will result in industry Business result write calculating storaging unit success, but message write duplicate removal memory element failure.This feelings Under condition, can still result in message and repeated Business Processing, thus affecting systematic function and efficiency.For example, when When receiving this message again, judge not existing this message in duplicate removal memory element, then again to this message Carry out Business Processing, and service processing result is write calculating storaging unit, this message write duplicate removal is deposited Storage unit.
For above-mentioned problem, the embodiment of the present invention further provides a solution, and it executes stream Journey can as shown in Figure 2, upon receiving a message, execution following steps:
Whether in 201, judge to be stored with duplicate removal memory element the message that this receives, if It is no longer this message to be processed;Otherwise, 202 are executed.
In 202, judge in calculating storaging unit, whether there is this message corresponding Business Processing knot Really, if it is, no longer this message is carried out with Business Processing, direct execution step 205;Otherwise, execute 203.
When whether there is the corresponding service processing result of this message in judging calculating storaging unit, can Whether there is service processing result in calculating storaging unit to determine first, then pass through to judge to calculate storage Whether there is the message receiving in unit to realize.If existed at business in calculating storaging unit There is, in reason result and calculating storaging unit, the message that this receives it is determined that calculating storaging unit In there is the corresponding service processing result of this message, again business need not be carried out to the message receiving Process.
In 203, Business Processing is carried out to this message.
In 204, by corresponding with the service processing result of this message for this message write calculating storaging unit.
In 205, this message is write duplicate removal memory element.
It is exemplified below with several scenes:
Assume for the first time to receive message a, judge in duplicate removal memory element that this message and calculating is not deposited There is not the corresponding service processing result of this message in storage unit, then Business Processing carried out to this message yet, By the service processing result of message a and message a corresponding write calculating storaging unit, and message a is write Enter duplicate removal memory element.
Assume the write duplicate removal memory element failure of message a.
Receive message a again, judge there is not this message in duplicate removal memory element, but calculate storage There is the corresponding service processing result of this message a in unit, then again industry need not have been entered to this message Business is processed, and directly message a is write duplicate removal memory element.Obviously, this mode avoids to message a Carry out Business Processing and write calculating storaging unit again, improve systematic function and efficiency.
In addition, in order to improve efficiency, can be by the content map in duplicate removal memory element to bloomfilter (Bloom filter), when receiving message, first determines whether whether this message hits bloomfilter, If not, illustrating that this message does not exist in duplicate removal memory element, then in the relaying of execution flow process shown in Fig. 2 Continuous execution step 202;If hit bloomfilter, illustrate may be stored with duplicate removal memory element This message, then continue execution step 201 in execution flow process shown in Fig. 2.
In flow process shown in above-mentioned Fig. 2, it is understood that there may be such a scene, the Business Processing of message a according to Rely the service processing result in message b, in this case, message a and message b just have identical Calculate key (calculate mark), such as certain message is that the order volume to user u adds up, need by The order volume of user u comprising in this message with using the user u's of the service processing result of message before Order volume is added up, and these message all have identical and calculate key, and this calculating key can be using use Mark of family u etc..For this scene, in order to more accurately ensure duplicate removal, to message in 203 When carrying out Business Processing, if existed in calculating storaging unit with the message that this receives have identical Calculate the messaging list of key, then with the message receiving, there are identical calculations using in calculating storaging unit The corresponding service processing result of messaging list of key, carries out Business Processing to the message receiving.204 In, this message receiving is added above-mentioned messaging list, and is updated using this service processing result The corresponding service processing result of this messaging list in calculating storaging unit.To comprise to receive in 205 The above-mentioned messaging list write duplicate removal memory element of message.
Give one example:
Assume to receive message a, judge not comprising message a in duplicate removal memory element, in calculating storaging unit Do not comprise message a, then Business Processing is carried out to this message a, obtains service processing result yet, should Message a and its corresponding service processing result a write calculating storaging unit, now in calculating storaging unit Message a is only comprised in the corresponding messaging list of this service processing result a;By the write duplicate removal storage of message a Unit.
Receive message b on this basis again, judge not comprising message b in duplicate removal memory element, calculate Also do not comprise message b in memory element, but with message b, there is phase due to containing in calculating storaging unit With the messaging list calculating key, therefore, offseted using the service processing result a in calculating storaging unit Breath b carries out Business Processing, obtains service processing result b.Message b is added the upper of calculating storaging unit State messaging list, now comprise message a and message b in messaging list, using service processing result b more New business result a, now comprises the messaging list corresponding Business Processing knot of message a and message b Fruit is service processing result b.Message a and message b are write duplicate removal memory element.
If receive message a, the write duplicate removal memory element failure of message a then can cause to calculate Comprise message a in memory element, and in duplicate removal memory element, do not comprise message a.So on this basis such as Fruit receives message b, is capable of the Business Processing of message b by above-mentioned flow process, and by message a and Message b writes duplicate removal memory element, that is, when receiving message b, message a benefit is written with duplicate removal and deposits Storage unit.Even if subsequently receiving message a again it is also possible to ensure this message a will not be repeated Business Processing.
In fig. 2 after shown step 203, the service processing result of this message and message can be put Enter to write queue.Further, since there may be the mode of multiple threads in flow-type business processing system, Therefore, before execution step 202, can first determine whether write queue and duplicate removal storage write queue (after Will be related in continuous description) in whether this message included, if it is, not executing 202, no longer The message that this receives is processed;Otherwise, then execution step 202.
Specific write flow process, that is, the step 204 shown in Fig. 2 and 205 can as shown in Figure 3, Comprise the following steps:
In 301, one cachenode (cache node) of taking-up from write queue, one Message and the corresponding service processing result of this message is comprised, wherein message can be to disappear in cachenode Presented in breath list.
In 302, judge whether existed and this cachenode bag in duplicate removal storage write queue There is the message of identical calculations key containing message, if it is, this cachenode is put back to write queue, Wait re-executes 301;Otherwise, 303 are executed.
In order to prevent message to be not yet stored into duplicate removal memory element, just be there is the message of identical calculations key to cover Lid falls, and in this step, first judges whether existed and this cachenode in duplicate removal storage write queue Comprised message has the message of identical calculations key, if it has, explanation has the message of identical calculations key It is being written into duplicate removal memory element it should write new message after waiting to be written completing, therefore first This cachenode is put back to write queue, wait retries.
In 303, by the content write calculating storaging unit in cachenode.
In 304, the message in cachenode is put into duplicate removal storage write queue.
In 305, store write queue from duplicate removal and take out message, message is write duplicate removal memory element.
It should be noted that above-mentioned write queue is actually a total queue, it is mainly used in calculating The write of memory element, and duplicate removal storage write queue is only used for the write of duplicate removal memory element.Can see Go out, the write of the write of calculating storaging unit and duplicate removal memory element is by the way of asynchronous, thus improving Handling capacity.
Fig. 4 is structure drawing of device provided in an embodiment of the present invention, and this device can be arranged at flow-type business Need in reason system in the equipment that message is processed.As shown in Figure 4, this device may include that Receiving unit 01, the first judging unit 02, the second judging unit 03, processing unit 04, first write Unit 05 and the second writing unit 06, can further include the 3rd judging unit 07 and the 4th judgement Unit 08.Wherein, the major function of each component units is as follows:
Receiving unit 01 is responsible for receiving message.
Responsible this message that judges whether to be stored with duplicate removal memory element of first judging unit 02, if sentenced Disconnected result is yes, no longer this message is processed.When the judged result of the first judging unit 02 is no, Second judging unit 03 judges whether there is the corresponding service processing result of message in calculating storaging unit. When the judged result of the second judging unit 03 is to be, processing unit 04 is not processed to message;? When the judged result of two judging units 03 is no, processing unit 04 carries out Business Processing to message.
First writing unit 05 is responsible for for the service processing result of message writing calculating storaging unit.Second write Unit 06 is after the first writing unit 05 completes write, or in the judged result of the second judging unit 03 is When no, message is write duplicate removal memory element.
Preferably, in order to improve efficiency, can be by the content map in duplicate removal memory element to bloomfilter. After receiving unit 01 receives message, first can judge whether this message hits by the 3rd judging unit 07 Bloomfilter, if not, illustrating that this message does not necessarily exist in duplicate removal memory element, triggering second judgement The operation that unit 03 execution judges;If it is, this message that may be stored with explanation duplicate removal memory element, then Trigger the operation that the first judging unit 02 execution judges.
For needing to carry out the scene of Business Processing using history service result, in order to more accurately ensure Duplicate removal, it is preferred to use in the following manner: if existed in calculating storaging unit and the message tool receiving There is the messaging list of identical calculations key, then processing unit 04 has identical calculations key using with this message The corresponding service processing result of messaging list, carries out Business Processing to this message.
Now, the first writing unit 05 by this message add messaging list, and using processing unit 04 this Service processing result update calculating storaging unit in the corresponding service processing result of messaging list.Second write Unit 06 will comprise the messaging list write duplicate removal memory element of the message that this receives.
In addition, processing unit 04 can also be by this message (comprising the message queue of this message) and its corresponding Service processing result puts into write queue.First writing unit 05 and the second writing unit 06 are using write team The operation of row execution write.
Specifically, the first writing unit 05 can take out a cachenode from writing queue, one Message (messaging list) and its corresponding service processing result is comprised, by cachenode in cachenode In content write calculating storaging unit, by the message in cachenode put into duplicate removal storage write queue. Second writing unit 06 stores from duplicate removal and takes out message write queue and write duplicate removal memory element.Permissible Find out, the write of the write of calculating storaging unit and duplicate removal memory element is by the way of asynchronous, thus carrying High-throughput.
Preferably, when the judged result of above-mentioned first judging unit 02 is no, the 4th judging unit 08 judges Write queue and go to re-write in queue whether there is message, if it is, not processing to message; Otherwise, the operation that the execution of triggering the second judging unit 03 judges.
Further, the content write in cachenode is being calculated storage by the first writing unit 05 Before unit, can first judge whether existed and cachenode bag in duplicate removal storage write queue There is the message of identical calculations key containing message, if it is, cachenode is put back to write queue, Wait retrieves;Otherwise, continue executing with and the content in cachenode is write calculating storaging unit Operation.
It should be noted that method and apparatus provided in an embodiment of the present invention be not limited to jstorm and Storm, goes for the Message Processing of any flow-type business processing system using similar mechanism.
It should be understood that disclosed apparatus and method in several embodiments provided by the present invention, Can realize by another way.For example, device embodiment described above is only schematically, For example, the division of described unit, only a kind of division of logic function, in addition actual can have when realizing Dividing mode.
The described unit illustrating as separating component can be or may not be physically separate, make For the part that unit shows can be or may not be physical location, you can with positioned at a place, Or can also be distributed on multiple NEs.Can select according to the actual needs part therein or The whole unit of person is realizing the purpose of this embodiment scheme.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit In or unit is individually physically present it is also possible to two or more units are integrated in one In individual unit.Above-mentioned integrated unit both can be to be realized in the form of hardware, it would however also be possible to employ hardware adds The form of SFU software functional unit is realized.
The above-mentioned integrated unit realized in the form of SFU software functional unit, can be stored in a computer In read/write memory medium.Above-mentioned SFU software functional unit is stored in a storage medium, including some fingers Order is with so that a computer equipment (can be personal computer, server, or network equipment etc.) Or processor (processor) executes the part steps of each embodiment methods described of the present invention.And it is aforementioned Storage medium include: u disk, portable hard drive, read only memory (read-only memory, rom), Random access memory (random access memory, ram), magnetic disc or CD etc. are various can Medium with store program codes.
The foregoing is only presently preferred embodiments of the present invention, not in order to limit the present invention, all this Within the spirit of invention and principle, any modification, equivalent substitution and improvement done etc., should be included in Within the scope of protection of the invention.

Claims (14)

1. a kind of message treatment method of streaming computing system is it is characterised in that the method includes:
S1, reception message;
Whether s2, judge to be stored with the first memory element described message, if it is, not to this message Processed;Otherwise, execute s3;
S3, judge in the second memory element, whether there is the corresponding service processing result of described message, such as Fruit is to execute s6, otherwise executes s4;
S4, Business Processing is carried out to described message;
S5, described message and its service processing result are write the second memory element;
S6, by described message write the first memory element.
2. method according to claim 1 is it is characterised in that content in described first memory element It is mapped to Bloom filter bloomfilter;
Also included before described s2: judge whether described message hits bloomfilter, if not, execution Described s3;If it is, executing described s2.
3. method according to claim 1 it is characterised in that execute described s4 when, if institute State and in the second memory element, there is the messaging list with described message with identical calculations mark, then utilize There is with described message the corresponding service processing result of messaging list of identical calculations mark, described message is entered Row Business Processing;
In described s5, described message is added described messaging list, and the service processing result using s4 Update the corresponding service processing result of messaging list described in described second memory element;
When executing described s6, the described messaging list comprising described message is write the first memory element.
4. the method according to claim 1,2 or 3 is it is characterised in that in described s4 and described s5 Between also include: described message and its corresponding service processing result are put into write queue;
To the content execution described s5 and described s6 in said write queue.
5. method according to claim 4 is it is characterised in that hold to the content in said write queue Row described s5 and described s6 includes:
S51, from said write queue take out a cache node, comprise message in one cache node And its corresponding service processing result;
S52, by described cache node content write described second memory element;
S53, the message in described cache node is put into duplicate removal storage write queue;
S54, store write queue from duplicate removal and take out message write the first memory element.
6. method according to claim 5 it is characterised in that after described s2 and described s3 it Front also include:
Judge whether there is described message, such as in said write queue and the storage write queue of described duplicate removal Fruit is then not carry out Business Processing to described message, otherwise continues executing with described s3.
7. method according to claim 5 is it is characterised in that between described s51 and described s52 Also include:
Judge whether there is the comprised message tool with described cache node in described duplicate removal storage write queue There is the message of identical calculations mark, if it is, described cache node is put back to said write queue, wait Re-execute described s51;Otherwise, continue executing with described s52.
8. a kind of message processing apparatus of streaming computing system are it is characterised in that this device includes:
Receiving unit, for receiving message;
First judging unit, for the described message that judges whether to be stored with the first memory element, if Judged result is yes, no longer described message is processed;
Second judging unit, for when the judged result of described first judging unit is no, judging described the Whether there is the corresponding service processing result of described message in two memory element;
Processing unit, for when the judged result of described second judging unit is for being, not entering to described message Row Business Processing;When the judged result of described second judging unit is no, described message is carried out at business Reason;
First writing unit, for writing the second memory element by the service processing result of described message;
Second writing unit, for, after described first writing unit completes write, or sentencing described second When the judged result of disconnected unit is no, described message is write the first memory element.
9. device according to claim 8 is it is characterised in that content in described first memory element It is mapped to bloomfilter;
This device also includes: the 3rd judging unit, for after described receiving unit receives described message, Judge whether described message hits bloomfilter, if not, triggering described second judging unit to execute judgement Operation;If it is, the operation that the described first judging unit execution of triggering judges.
10. device according to claim 8 is it is characterised in that described processing unit, specifically for such as There is the messaging list with described message with identical calculations mark, then in really described second memory element Using the corresponding service processing result of messaging list with described message with identical calculations mark, disappear to described Breath carries out Business Processing;
Described first writing unit, specifically for adding described messaging list by described message, and using described This service processing result of processing unit updates the corresponding industry of messaging list described in described second memory element Business result;
Described second writing unit, specifically for depositing the described messaging list write first comprising described message Storage unit.
Device described in 11. according to Claim 8,9 or 10 it is characterised in that described processing unit, It is additionally operable to for described message and its corresponding service processing result to put into write queue;
Described first writing unit and described second writing unit are write using the content execution in said write queue The operation entering.
12. devices according to claim 11 it is characterised in that described first writing unit, specifically For from said write queue take out a cache node, comprise in one cache node message and its Corresponding service processing result, the content in described cache node is write described second memory element, by institute State the message in cache node and put into duplicate removal storage write queue;
Described second writing unit, deposits specifically for storing to take out message in write queue and write first from duplicate removal Storage unit.
13. devices according to claim 11 are it is characterised in that this device also includes:
4th judging unit, for when the judged result of described first judging unit is no, judging said write Queue and described go to re-write in queue whether there is described message, if it is, not to described message Processed;Otherwise, the operation that the described second judging unit execution of triggering judges.
14. devices according to claim 11 are it is characterised in that described first writing unit is by institute Before stating content described second memory element of write in cache node, it is additionally operable to judge that described duplicate removal storage is write Whether there is the message that the comprised message with described cache node has identical calculations mark in enqueue, If it is, described cache node is put back to said write queue, wait retrieves;Otherwise, continue to hold The content being about in described cache node writes the operation of described second memory element.
CN201510413095.4A 2015-07-15 2015-07-15 Message processing method and device for streaming computing system Active CN106354722B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510413095.4A CN106354722B (en) 2015-07-15 2015-07-15 Message processing method and device for streaming computing system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510413095.4A CN106354722B (en) 2015-07-15 2015-07-15 Message processing method and device for streaming computing system

Publications (2)

Publication Number Publication Date
CN106354722A true CN106354722A (en) 2017-01-25
CN106354722B CN106354722B (en) 2019-12-24

Family

ID=57842506

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510413095.4A Active CN106354722B (en) 2015-07-15 2015-07-15 Message processing method and device for streaming computing system

Country Status (1)

Country Link
CN (1) CN106354722B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107943840A (en) * 2017-10-30 2018-04-20 深圳前海微众银行股份有限公司 Data processing method, system and computer-readable recording medium
CN110457305A (en) * 2019-08-13 2019-11-15 腾讯科技(深圳)有限公司 Data duplicate removal method, device, equipment and medium

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103763378A (en) * 2014-01-24 2014-04-30 中国联合网络通信集团有限公司 Task processing method and system and nodes based on distributive type calculation system
CN104615777A (en) * 2015-02-27 2015-05-13 浪潮集团有限公司 Method and device for real-time data processing based on stream-oriented calculation engine
CN104683488A (en) * 2015-03-31 2015-06-03 百度在线网络技术(北京)有限公司 Flow-type calculation system as well as dispatching method and dispatching device of flow-type calculation system

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103763378A (en) * 2014-01-24 2014-04-30 中国联合网络通信集团有限公司 Task processing method and system and nodes based on distributive type calculation system
CN104615777A (en) * 2015-02-27 2015-05-13 浪潮集团有限公司 Method and device for real-time data processing based on stream-oriented calculation engine
CN104683488A (en) * 2015-03-31 2015-06-03 百度在线网络技术(北京)有限公司 Flow-type calculation system as well as dispatching method and dispatching device of flow-type calculation system

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107943840A (en) * 2017-10-30 2018-04-20 深圳前海微众银行股份有限公司 Data processing method, system and computer-readable recording medium
CN107943840B (en) * 2017-10-30 2022-01-11 深圳前海微众银行股份有限公司 Data processing method, system and computer readable storage medium
CN110457305A (en) * 2019-08-13 2019-11-15 腾讯科技(深圳)有限公司 Data duplicate removal method, device, equipment and medium

Also Published As

Publication number Publication date
CN106354722B (en) 2019-12-24

Similar Documents

Publication Publication Date Title
JP6743141B2 (en) Service processing method and device
CN107395665A (en) A kind of block chain service handling and business common recognition method and device
US9665460B2 (en) Detection of abnormal resource usage in a data center
CN104063293B (en) A kind of data back up method and stream calculation system
CN109725899B (en) Data stream processing method and device
CN104035836B (en) Automatic disaster-tolerant recovery method and system in cluster searching platform
CN109298990A (en) Log storing method, device, computer equipment and storage medium
KR101574451B1 (en) Imparting durability to a transactional memory system
KR102024005B1 (en) Method for processing transactions using blockchain, and transaction management server using the same
CN103294558B (en) A kind of MapReduce dispatching method supporting dynamic trust evaluation
CN104156197A (en) Microprocessor and method for operating microprocessor
CN107729018A (en) Cyclic vector method and apparatus
CN108108248A (en) A kind of CPU+GPU cluster management methods, device and equipment for realizing target detection
CN106201561B (en) The upgrade method and equipment of distributed caching cluster
CN110213128A (en) Serve port detection method, electronic equipment and computer storage medium
CN109325548A (en) Image processing method, device, electronic equipment and storage medium
CN107766188A (en) Internal-memory detection method and device in train control system
CN110737504A (en) deep learning model training fault-tolerant method, system, terminal and storage medium
CN110324211A (en) A kind of data capture method and device
CN106354722A (en) Message processing method and device for stream-oriented computation system
CN108369547A (en) The tail portion of daily record in persistence main memory
CN103634167B (en) Security configuration check method and system for target hosts in cloud environment
CN108984779A (en) Distributed file system snapshot rollback metadata processing method, device and equipment
US11210757B2 (en) GPU packet aggregation system
JP6847112B2 (en) How and devices to process data after node reboot

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant