CN106209663B - Flow segmentation and system - Google Patents

Flow segmentation and system Download PDF

Info

Publication number
CN106209663B
CN106209663B CN201610547429.1A CN201610547429A CN106209663B CN 106209663 B CN106209663 B CN 106209663B CN 201610547429 A CN201610547429 A CN 201610547429A CN 106209663 B CN106209663 B CN 106209663B
Authority
CN
China
Prior art keywords
flow
cut
according
cutting
assignment
Prior art date
Application number
CN201610547429.1A
Other languages
Chinese (zh)
Other versions
CN106209663A (en
Inventor
何萌鑫
张云锋
周盛
潘柏宇
王冀
Original Assignee
优酷网络技术(北京)有限公司
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 优酷网络技术(北京)有限公司 filed Critical 优酷网络技术(北京)有限公司
Priority to CN201610547429.1A priority Critical patent/CN106209663B/en
Publication of CN106209663A publication Critical patent/CN106209663A/en
Application granted granted Critical
Publication of CN106209663B publication Critical patent/CN106209663B/en

Links

Abstract

The present invention relates to Internet technical field, a kind of flow segmentation and system are disclosed.The method includes the steps: flow cutting request is received, the information of shunt volume to be cut in the request is obtained;It is selected according to the information of the shunt volume to be cut and searches random tree;The cutting condition value that the shunt volume to be cut is determined according to the information of the shunt volume to be cut carries out lookup matching in the lookup random tree;It is matched as a result, the assignment of traffic mode of the shunt volume to be cut is determined according to the branch to match with the cutting condition value, according to shunt volume to be cut described in the assignment of traffic mode treatment according to the lookup.The flow cutting for being implemented in combination with high efficient and flexible that technical solution of the present invention passes through tree-like lookup and stochastic searching.

Description

Flow segmentation and system

Technical field

The present invention relates to Internet technical field, in particular to a kind of flow segmentation and system.

Background technique

Internet sites need first to carry out small-scale test assessment after new function either new strategy is developed and finished Just meeting full flow is online, for the superiority and inferiority for sufficiently assessing new function or new strategy as far as possible, is often required in real network environment It is tested, the efficiency and flexibility of testing assessment have the influence of essence to system operation.Such as the demand of Internet advertising Fang Pingtai (Demand-Side Platform, DSP) system, exactly one system for needing dynamic to adjust at any time, is opened in system Need ceaselessly online some operational functions and demand, influence of some demands to whole system non-in hair and operational process Chang great influences the reliability and stability of system, singly can not if directly online may be problem on unexpected line occur The loss of income and the mistake of data are caused, user experience is also very bad.

Effectively to be assessed internet system, for generalling use stream at present i.e. by online critical function/strategy It measures cutting method and carries out real network assessment test.Common flow cutting appraisal procedure is A-B test, and concrete mode is complete It samples out in flow two parts of small flows, walks the B branch of new function/strategy A branch and old function/strategy respectively, pass through comparison The difference of each index under this two parts of flows can effectively evaluate new function/strategy superiority and inferiority, and then determine new function/plan Slightly whether formal online and full flow is run.

More specifically, main flow segmentation is random chance cutting either hash cutting in the prior art.With Machine probability slit mode is then to choose one by the size of this value by calculating a random value for each ad traffic A processing node, to achieve the purpose that cutting flow.Hash cutting be by for each ad traffic uniquely indicate into Row hash operation chooses a processing node in processing node listing according to operation result to handle current flow, to reach The purpose of flow cutting.

However, either random chance cutting or hash cutting, actual flow processing is all very coarse cutting side The characteristics of method, these cutting methods ignore flow itself, using flow as the single mode cutting of indiscriminate data at Reason, is a kind of cutting method that resource is extensive.Although existing flow segmentation is easily achieved because simple and original, Existing method is no longer satisfied the demand of traffic classification user.Furthermore what existing flow slit mode was all planar as cuts The mode of dividing can only realize the flow cutting of fixed level, and the flexible configuration that this slit mode is difficult, readable and maintainability is all It is excessively poor, it configures a set of effective cutting method for meeting actual demand and needs manual compiling largely cumbersome and duplicate with confidence Breath.

Summary of the invention

In view of the drawbacks of the prior art, the object of the present invention is to provide a kind of flow segmentation and systems, with flexibly high Realize flow cutting in effect ground.

According to an aspect of the invention, there is provided a kind of flow segmentation, comprising steps of

Flow cutting request is received, the information of shunt volume to be cut in the request is obtained;

It is selected according to the information of the shunt volume to be cut and searches random tree;

The cutting condition value that the shunt volume to be cut is determined according to the information of the shunt volume to be cut is searched at random described Lookup matching is carried out in tree;

It is matched as a result, being determined according to the branch to match with the cutting condition value described to be slit according to the lookup The assignment of traffic mode of flow, according to shunt volume to be cut described in the assignment of traffic mode treatment.

Preferably, the method also includes steps:

Preset configuration file is read, at least one is generated according to the configuration file and searches random tree.

Preferably, include: in the configuration file

At least one node passes through level connection between the node and constitutes at least one tree structure;

The lookup random tree is generated by the tree structure.

Preferably, include: in the assignment of traffic mode

The assignment of traffic ratio that the processing node of current branch is configured to.

Preferably, the shunt volume to be cut according to the assignment of traffic mode treatment includes:

The assignment of traffic ratio determined according to the assignment of traffic mode is to be cut as described in the processing node random process Shunt volume.

In another aspect of this invention, it goes back while providing a kind of flow cutting system, comprising:

Data obtaining module obtains the information of shunt volume to be cut in the request for receiving flow cutting request;

Selected module is set, is selected according to the information of the shunt volume to be cut and searches random tree;

Matching module is searched, for determining the cutting condition of the shunt volume to be cut according to the information of the shunt volume to be cut Value, carries out lookup matching in the lookup random tree;

Flow processing module, for according to it is described search it is matched as a result, according to matching with the cutting condition value Branch determines the assignment of traffic mode of the shunt volume to be cut, according to shunt volume to be cut described in the assignment of traffic mode treatment.

Preferably, the system also includes:

Configuration module generates at least one lookup according to the configuration file for reading preset configuration file Random tree.

Preferably, the configuration module includes:

Node configuration module passes through level connection composition at least one between the node for configuring at least one node A tree structure;

Generation module is set, for generating the lookup random tree by the tree structure.

Preferably, the flow processing module further include:

Ratio obtains module, the assignment of traffic ratio that the processing node for obtaining current branch is configured to.

Preferably, the flow processing module further include:

Random process module, the assignment of traffic ratio for being determined according to the assignment of traffic mode is by the processing node Shunt volume to be cut described in random process.

The embodiment of the invention provides a kind of flow segmentation and systems, pass through the combination of tree-like lookup and stochastic searching Realize the flow cutting of high efficient and flexible.Wherein, filtering and analysis for flow own characteristic are completed by tree-like lookup, The characteristics of this lookup is precisely to match, and scalability and maintainability are all very strong;And stochastic searching is mainly used for the cutting of flow, It needs the reasonable cutting of flow to be used to complete different functions, stochastic searching can ignore the own characteristic of flow and realize a view The cutting of colleague.Therefore the technical solution of the embodiment of the present invention realizes dynamic rationally under the premise of fully considering discharge characteristic Flow planning.

Detailed description of the invention

Fig. 1 is the flow diagram of flow segmentation in one embodiment of the invention;

Fig. 2 is to configure the schematic diagram for searching random tree in a preferred embodiment of the invention by XML file;

Fig. 3 is the structural schematic diagram of the lookup random tree in another preferred embodiment of the present invention with many levels;

Fig. 4 is the module diagram of flow cutting system in another embodiment of the present invention.

Specific embodiment

In order to make the objectives, technical solutions and advantages of the present invention clearer, With reference to embodiment and join According to attached drawing, the present invention is described in more detail.It should be understood that these descriptions are merely illustrative, and it is not intended to limit this hair Bright range.In addition, in the following description, descriptions of well-known structures and technologies are omitted, to avoid this is unnecessarily obscured The concept of invention.

With in internet video ads flow be continuously increased and the continuous development and maturation in video ads market, Flow plans the critical bottleneck for having become web station system performance, and distributing rationally for floating resources receives more and more attention, How by flow reasonably be consideration required for each website critical issue.For the flow to internet system It makes rational planning for, needs to adjust flow processing strategie often, in addition system new function is continually developed, to the effect of flow cutting Rate and flexibility have high requirement.Main flow segmentation is random chance cutting either hash in the prior art All flow demands are considered as indiscriminate request of data to calculate cutting strategy value by cutting, although which decision machine System is simple, is easily achieved, but correspondingly, flexibility is very poor, is unable to satisfy the multi-level cutting demand of diversification, and can not move State extension.

For the reasonable utilization for realizing flow by the flow cutting of high efficient and flexible, the present invention provides a kind of completely new flows The technical solution of cutting.Specifically, embodiment of the invention discloses a kind of flow segmentation and systems, can be with flexible configuration Segmentation rules simultaneously realize that dynamic adjusts.As shown in Figure 1, in one embodiment of the invention, flow segmentation comprising steps of

S1 receives flow cutting request, obtains the information of shunt volume to be cut in the request;

S2 is selected according to the information of the shunt volume to be cut and is searched random tree;

S3 determines the cutting condition value of the shunt volume to be cut according to the information of the shunt volume to be cut, in the lookup Lookup matching is carried out in random tree;

S4, according to it is described search it is matched as a result, according to the branch that matches with the cutting condition value determine it is described to The assignment of traffic mode of cutting flow, according to shunt volume to be cut described in the assignment of traffic mode treatment.

In the technical solution of the embodiment of the present invention, automatic distribution and the place of flow cutting are realized by searching for random tree Reason, wherein searching the selection of random tree and the specific allocation model of flow is determined according to the various characteristic informations of flow , thus can guarantee that cutting scheme is best suitable for the actual features and demand of flow, it realizes efficiently careful and reasonably flows Measure cutting configuration.

In addition, in a preferred embodiment of the invention, the flow segmentation further comprises the steps of: reading and sets in advance Fixed configuration file generates at least one according to the configuration file and searches random tree.That is, in the embodiment of the present invention Simultaneously dynamic generation can freely be configured by searching random tree, this allows the technical solution of the embodiment of the present invention for the tool of flow Body situation and dynamic adjusts, realize more flexible demand Flow cutting planning.In addition, each configuration file can preferably match Set one or more lookup random trees;Random tree is searched to determine by least one node for the structure that has levels.It is i.e. described to match Setting includes at least one node in file, and multiple nodes are contacted by level and constitute at least one tree structure.It is highly preferred that institute Stating configuration file is XML file.Specifically, searching random tree includes at least one cutting node, and each cutting node is configured with At least two attributes, name and value, wherein name indicates that the title of branch representated by current cutting node, the title are also The unique identification of current cutting node, value indicate the percentage of flow assigned by current cutting node.Ordinary circumstance Under, 100% should be less than or equal to by searching the sum of value value of all cutting nodes in random tree for one.Further, cutting section Specific slit mode, each direction node are configured by least one cutting condition node (direction) under point Including conditional name (type) and at least one cutting value (value) node.Cutting value node include condition value and one or Multiple processing nodes (chain), each processing node includes id, title (name) and the cutting flow-rate ratio for handling node (chance)。

XML configuration file in another preferred embodiment of the present invention is as shown in Figure 2.In Fig. 2, used in stochastic searching tree AbTest is configured with a cutting node, entitled test1, and the flow percentage value being assigned to is 30;More specifically, Test1 cutting node include one according to region area come the cutting condition node of cutting;The cutting condition node is cut by three Score value (value) node is constituted, and the first cutting value node condition value is a1, including a processing node, processing node i d are Test1_1, entitled chain1, (corresponding flow is assigned randomly at this responsible 30% cutting flow-rate ratio according to this ratio Node completion processing is managed, same as below);Second cutting value node condition value is a2, including two processing nodes, two processing nodes Id is respectively test1_2 and test1_3, and title is respectively chain2 and chain1, is each responsible for 30% cutting flow-rate ratio;The Three cutting value node condition values are a3, including a processing node, which is test1_4, entitled chain3, are born The cutting flow-rate ratio of duty 30%;4th cutting value node condition value be a4, further comprise one according to type b cutting come cutting Slitting part node;The cutting condition node is made of two cutting value nodes, i.e. the 5th cutting value node condition value is b1, including One processing node, the processing node i d are test1_5, entitled chain4, are responsible for 30% cutting flow-rate ratio;6th cutting Value node condition value is b2, including a processing node, which is test1_6, entitled chain5, is responsible for 30% Cutting flow-rate ratio.By the file configuration of the XML of Fig. 2, a lookup random tree test1, the lookup random tree can be generated Flow cutting is mainly configured according to region, tetra- regions a1, a2, a3, a4 are each to have 1~2 processing node by oneself, walks 5 respectively Chain of Responsibility chain1, chain2, chain3, chain4, chain5 handle flow, wherein the region a4 is according further still to class Type b configures flow cutting, have two types of b1, b2 by a respective processing node come by corresponding Chain of Responsibility chain4, Chain5 handles flow.

As can be seen that the technical solution of the embodiment of the present invention can be realized by abundant flexibly configure by default item The free cutting of the flow of part, so as to realize the distribution according to need flow of high efficient and flexible.Further, the present invention another In preferred embodiment, configuration file can be adjusted dynamically at any time, for example additions and deletions lookup random tree, additions and deletions are all kinds of in XML file The technical solution of node and/or the related setting of modification etc., the preferred embodiment of the present invention can be at any time according to configuration text adjusted Part is read out, generates and the dynamic adjustment for completing flow cutting such as search random tree and select newest allocation model.

Normally, the cutting number of plies searched in random tree is fixed, one and only one cutting condition of same layer; Under same cutting value node, if the processing node of access has multiple, the sum of chance values of all processing nodes are necessary Less than or equal to 100;One request is only possible to one stochastic searching tree of selection and carries out lookup Chain of Responsibility, because of an XML configuration text It is configured with multiple lookup random trees in part, must be selected before processing request and random tree is searched using which.But in the present invention Other preferred embodiments in, not the considered critical cutting number of plies and/or cutting conditional number, meet configuration rationally, demand it is bright Really, it under the premise of not influencing system treatment effeciency, may be implemented more to enrich free cutting by more combinations Configuration.

For example, in a preferred embodiment of the invention, searching random tree can also be with more levels, more multi-configuration side The multiple-limb tree structure of formula.As shown in figure 3, A and B are indicated in the lookup random tree of another preferred embodiment of the invention The title of cutting condition, the expressions such as a1, b1 be cutting condition value, the subsequent number of cutting condition value, indicate flow symbol In the case where closing this condition value, how many probability enters sub-process below.The entirety of the lookup random tree of Fig. 3 is meant that: Under the conditions of A cutting, meets in the flow of A=a1 30% and walk No. 1 Chain of Responsibility, 30% walks No. 2 Chain of Responsibilities;Meeting A=a2's Under the conditions of cutting, if under conditions of current traffic request has B=b1,30% flow walks No. 3 Chain of Responsibilities, 30% flow is walked No. 4 Chain of Responsibilities, if under conditions of current traffic request has B=b2,30% flow walks No. 5 Chain of Responsibilities.In this way A variety of conditional combinations distribution of cutting condition may be implemented in configuration mode, and the strategy of flow cutting also may be implemented more thin The greasy specific requirements for being more in line with flow.

Through the invention in each embodiment in technical solution, the process of entire flow cutting combine it is tree-like lookup and The characteristics of stochastic searching.Wherein, filtering and analysis for flow own characteristic are completed by tree-like lookup, this lookup Feature is precisely to match, and scalability and maintainability are all very strong.Stochastic searching is mainly used for the cutting processing of flow, needs stream Reasonable cutting is measured to be used to complete different functions, is to need to ignore the own characteristic of flow to complete one and treat as in this layer The cutting of benevolence.Therefore the technical solution in the embodiment of the present invention may be implemented precisely to match and balance efficient assignment of traffic.

Further as shown in figure 4, correspondingly with the above method, the embodiment of the present invention also provides a kind of flow simultaneously Cutting system 1, comprising:

Data obtaining module obtains the information of shunt volume to be cut in the request for receiving flow cutting request;

Selected module is set, is selected according to the information of the shunt volume to be cut and searches random tree;

Matching module is searched, for determining the cutting condition of the shunt volume to be cut according to the information of the shunt volume to be cut Value, carries out lookup matching in the lookup random tree;

Flow processing module, for according to it is described search it is matched as a result, according to matching with the cutting condition value Branch determines the assignment of traffic mode of the shunt volume to be cut, according to shunt volume to be cut described in the assignment of traffic mode treatment.

Wherein, in embodiments of the present invention, independent equipment can be set to realize above-mentioned flow cutting system, with automatic The overall process of flow cutting is completed, the flow cutting system receives traffic requests, is determining flow processing node and flow point After mode, transfer to specific flow processing node that traffic requests are responded and handled by assignment of traffic mode.

Preferably, the system also includes:

Configuration module generates at least one lookup according to the configuration file for reading preset configuration file Random tree.

Preferably, the configuration module includes:

Node configuration module passes through level connection composition at least one between the node for configuring at least one node A tree structure;

Generation module is set, for generating the lookup random tree by the tree structure.

Preferably, the flow processing module further include:

Ratio obtains module, the assignment of traffic ratio that the processing node for obtaining current branch is configured to.

Preferably, the flow processing module further include:

Random process module, the assignment of traffic ratio for being determined according to the assignment of traffic mode is by the processing node Shunt volume to be cut described in random process.

Preferably, the flow cutting system in the embodiment of the present invention is relatively independent processing equipment, such as cluster, service Device, PC or laptop etc.;It is also possible to relatively independent functional unit, such as individual chips, system on chip (System On Chip, SoC) or IP kernel etc., corresponding function is realized after loading by processing equipment.In practical applications, in above system Each module can be by central processing unit (Central Processing Unit, CPU), the microprocessor in system equipment (Micro ProcessorUnit, MPU), digital signal processor (Digital Signal Processor, DSP) or scene Programmable gate array (FieldProgrammable GateArray, FPGA) etc. is realized;File data (including but not limited to flows Amount request, configuration file, lookup random tree, the corresponding audio-video document of traffic requests etc.) preferably protected by storing equipment It deposits and is supplied to above-mentioned processing equipment/unit.

In addition, in a preferred embodiment of the invention, the flow cutting system can be distributed computer System, the Distributed Computer System include at least one processing node, and the flow cutting system can be implemented as at least one A processing node;The Distributed Computer System can also include at least one server, to form centralization or area The distributed system of domain centralization, the flow cutting system can be implemented as at least one described server.Wherein, described point Cloth computer system preferably can be formed into the cloud system based on computer network.

The term " flow " referred in embodiments of the present invention may include the various network data transmissions for calculating equipment, than Such as text file, image file, audio file, video file, hand over using data and various visible and/or sightless data Mutual etc. network transmission.In view of the rapid development of Internet application and mobile Internet, the flow of the embodiment of the present invention is excellent Needle selection is to internet traffic and/or mobile Internet flow.Traffic requests may be from various known internet terminals, preferably For PC terminal and/or mobile terminal.It is highly preferred that " flow " mentioned by the embodiment of the present invention refers to is by internet site The video ads flow that PC terminal and/or mobile terminal provide.

In more preferred embodiment of the present invention, there is the comprehensive stream in internet including being formed comprising flow cutting system The electronic system of amount processing and integrated flow rate administrative mechanism.Wherein, electronic system preferably includes such as client or server First equipment is connected to the second equipment of such as client or server;First equipment can by communication path (such as without Line or cable network) it is communicated with the second equipment.Preferably, the communication mode between the first equipment and the second equipment includes but unlimited In: the first equipment initiates traffic requests to the second equipment, and the second equipment is completed according to the operating condition of flow cutting system to this The processing of traffic requests.Flow cutting system can be formed in the first equipment and/or the second equipment of the electronic system, Equipment/system can be independent independently of first equipment and the formation of the second equipment.

In one example, the first equipment can be any one in the various calculating equipment with certain processing capacity Kind, such as smart phone, smart television, personal digital assistant (PDA), notebook computer, PC, wearable is set tablet device Standby, cart-mounted computing device or other multifunctional entertainment equipment.First equipment can be coupled directly or indirectly to communication path It is communicated with the second equipment, or can be independent equipment.But relevant technical staff in the field is it should also be appreciated that the first equipment It can also be different types of equipment;For example, the first equipment is also possible to for assisting providing the equipment of computing capability, such as machine top Box, TV box, game station, HUD, intelligent telephone set etc..Herein, above-mentioned first equipment and multimedia enumerate are not answered It is interpreted as to the exhaustive of the specific embodiment of the invention.

Second equipment can be any one in various centralized or distributive computing facility, file or data processing equipments Kind.For example, the second equipment can be network server, desktop computer, plate, laptop computer, grid computing resource, void Quasi-ization computer resource, cloud computing resources, reciprocity distributive computing facility etc., or a combination thereof.In another example, second Equipment can be the arbitrary equipment for receiving and storing network data, such as disk array, cloud storage equipment, file-sharing system System, distributed memory system etc..Second equipment can concentrate in single room, be distributed in different rooms, be distributed in not Geographical location together is embedded within telecommunication network.Second equipment can have for coupling with communication path and set with first The component of standby communication.

Communication path can be various networks, such as may include wireless communication, wire communication, light, ultrasonic wave or combinations thereof, Satellite communication, cellular communication, bluetooth, Infrared Data Association's standard (IrDA), Wireless Fidelity (WiFi) and Worldwide Interoperability for Microwave access are mutual Connection (WiMAX) is the example for the wireless communication that can be included in communication path.Ethernet, Digital Subscriber Line (DSL), optical fiber It is that may include logical to family (FTTH), high-definition multimedia interface (HDMI) cable and plain old telephone service (POTS) Believe the example of the wire communication in path.In addition, communication path can pass through many network topologies and distance.For example, communication lines Diameter may include be directly connected to, personal area network (PAN), local area network (LAN), Metropolitan Area Network (MAN) (MAN), wide area network (WAN) or combinations thereof.

In an embodiment of the invention, flow segmentation or system can be in the arbitrary equipments of above-mentioned electronic system It executes processing and/or display includes intermediate treatment process (for example configures, receives and the details such as request, search, matching, being randomly assigned) Various data processing operations inside.

Further, one is enumerated for realizing the useful computer system of the disclosed embodiments, the department of computer science System may be constructed any appliance in above-mentioned electronic system or electronic system comprising one or more processors, and can be with It further comprise electronic display unit (for showing figure, text, audio-video and other data), main memory is (for example, random Access memory RAM), memory device (for example, hard disk drive), removable memory part (for example, can be removed storage driving Device, can be removed memory module, tape drive, CD drive, in the calculating for being wherein stored with computer software and/or data Machine readable medium), user interface facilities (for example, keyboard, touch screen, keyboard, pointer device) and communication interface are (for example, adjust Modulator-demodulator, network interface (such as Ethernet card), communication port or PCMCIA slot and card).Communication interface permission is calculating Software and data are transmitted between machine system and external equipment.Computer system further includes connection inside and outside device to realize that data are handed over The mutual communications infrastructure (for example, communication bus, crossover sub or network).

In addition, Miscellaneous Documents that traffic requests are targeted (including but not limited to text file, image file, audio file, Video file and application interaction data etc.) and configuration file etc. can be stored in the first equipment and/or the second equipment is local, or It is stored in any storage equipment of the electronic system;In another embodiment, input data (including but not limited to flows Measure request, user information, configuration information etc.) it can also be from input source (including but not limited to PC terminal, mobile terminal, server Deng) provide, output data (including but not limited to lookup random tree, lookup matching result, assignment of traffic mode and traffic requests institute Respective file etc.) other output source (including but not limited to caching, temporary storage, intermediary of network equipment, text can also be supplied to Part server, http-server, ftp server, P2P equipment, data server, streaming media server etc.) carry out present and/or Storage.Input/output data can be transmitted/received wirelessly or by wireline interface, and may include uncompressed/compression Data content.In one embodiment, the wired or wireless communication of the data content in electronic system may include below Communication in one or more: Ethernet, phone (for example, POTS), cable, power line and fibre system, and/or including long-term Evolution (TD-LTE or FDD-LTE) system, CDMA (CDMA or CDMA2000) communication system, frequency division multiple access (FDMA) system System, time division multiple acess (TDMA) system (such as GSM/GPRS (General Packet Radio Service)/EDGE (data GSM ring of enhancing Border), TETRA (terrestrial trunked radio) mobile telephone system), wideband code division multiple access (WCDMA) system, IEEE802.11 system The wireless system of one or more of system, orthogonal frequency division multiple access (OFDM) system etc..In one embodiment, input/output It source can be in wireless network (for example, internet, local area network (LAN), wide area network (WAN), personal area network (PAN), campus wireless network (CAN), Metropolitan Area Network (MAN) (MAN) etc.) on send.In one embodiment, input/output source may come from system based on cloud, Server, computer equipment, mobile device etc..

In one embodiment, it can be encrypted, press before reaching input node from the file data that input source inputs Contracting and/or digital signature, for example, via encryption unit, compression unit and/or signature coding unit processing;And reach input section It can directly transmit after point or be transmitted again after correspondingly being decrypted, decompressing and/or being verified signature, for example, single via decryption Member, decompression unit and/or verifying signature unit processing.In one embodiment, output node file data to be output can Be initial data, can also be before be encrypted, compression and/or digital signature data, can also be from receive data in solve The data obtained after close, decompression and/or verifying signature.The file data for being output to output source from output node reaches output Can also be encrypted before source, compress and/or digital signature, include herein directly forward be encrypted, compression and/or number label Name data, or through output node it is secondary encryption, compression and/or digital signature data, reach output source after again through correspondingly Definitive document household register is obtained after decryption, decompression and/or verifying signature.In one embodiment, output source can wirelessly or Output data is received from output node by wireline interface.

In one embodiment, the file data from input source can be from communication network, computer network, other calculating Machine readable storage medium storing program for executing or any other suitable data provide means to provide.In one embodiment, input source is come from File data can via radio frequency interface (4G, 3G communication etc.), computer network wired or wireless connection (such as with Too net, WLAN etc.), the various not Tongfangs of other wireline interfaces (USB etc.) or other wireless interfaces (bluetooth etc.) etc Formula is coupled to receiving device.

In certain embodiments, it may make up above-mentioned department of computer science in the symmetric multi-processors computer with multi-core processor System;According to some embodiments, core processing component may include assembly line (pipeline) processing equipment, is coupled to by bus Each processor;In one embodiment, coprocessor can be coupled to chipset core logic and primary processor by bus.? In some embodiments, one of coprocessor or primary processor or it is multiple can be used the symmetric multi-processors with multicore come count parallel Calculate gradient.Memory can be coupled to chipset core logic, and memory can for example comprising main memory, hard disk drive and Removable media etc..

In one embodiment, realize that the sequence of computer instructions of each step of method above can be realized to be stored in Machine readable instructions in the part of main memory.In addition, can also be established in main memory continuous buffer, register, pond, The forms such as storehouse and queue.But in other examples, the buffering and storage for realizing the sequence of the above method can also be in main process tasks Device, another memory are realized in coprocessor.

In certain embodiments, realize that the sequence of the above method can be realized with software, and in other embodiments they It can be realized with hardware.In a software embodiment, sequence can be deposited by the instruction executable to processor or processor by calling Instruction in computer-readable medium is stored up to realize.Computer-readable medium can be any computer readable storage medium, packet Include magnetic, semiconductor or optical memory.

Flow segmentation technique described in the embodiment of the present invention can be realized with various hardware configurations.For example, search and/or with Machine processing function can be integrated into chipset, and by entity hardware or the equipment with firmware is completed.Alternatively, distribution can be used Processor carrys out parallel processing operation.As another embodiment, the available general purpose including multi-core processor of parallel processing function Processor is realized.

It can be the form of signal via the information that communication interface is transmitted, such as electricity, electromagnetism, light or via communication link energy Enough by the received other signals of communication interface, the communication link carries signal and line or cable, optical fiber, phone can be used Line, cellular phone link, radio frequency (RF) link and/or other communication channels are realized.The described meter of the embodiment of the present invention Calculation machine program instruction can be loaded in computer, programmable data processing system or processing equipment, so that executing on it Sequence of operations to generate computer implemented process.

Term " computer program medium ", " computer usable medium ", " computer-readable medium " and " computer program production Product " are used to generically refer to media, and such as main memory, removable memory part, is mounted in hard disk drive auxiliary storage Hard disk and signal.These computer program products are for providing the component of software to computer system.Computer-readable Jie Matter allows computer system to read data, instruction, message or message package and other computer-readable from computer-readable medium Information.For example, computer-readable medium may include nonvolatile memory, such as floppy disk, ROM, flash memory, disc driver are deposited Reservoir, CD-ROM and other persistent storage systems.For example, it for transmitting such as data and calculating between computer systems The information of machine instruction is useful.Computer program instructions can be stored in computer-readable medium, computer-readable Jie Matter can instruct computer, other programmable data processing systems or other equipment to work in a specific manner, so that by depositing The function that the instruction generation of storage in computer-readable medium is specified in Fig. 1 flow chart and/or Fig. 4 block diagram block including implementation/ The manufacture of the instruction of behavior.

In addition, computer-readable information can be included connecing in such as network link and/or network by computer-readable medium In the state medium of mouth, including cable network or wireless network, allow computer to read such computer-readable information.Meter Calculation machine program (i.e. computer control logic) is stored in main memory and/or auxiliary storage.Computer program can also be through It is received by communication interface.Such computer program makes computer system be able to carry out spy as in this discussion when being run Sign.Specifically, computer program is when being run, so that multi-core processor executes the feature of computer system.Such calculating The controller of machine program representation computer system.

Technical solution based on the embodiment of the present invention can have plurality of application scenes, for example, in exploitation Internet advertising Party in request's platform (Demand-Side Platform, DSP) system when, need ceaselessly online some operational new functions And new demand, influence of some demands to whole system is very big, if directly online it is possible that on unexpected line Problem, or even the loss of income and the mistake of data are caused, and very deleterious effect is caused for the experience of user.And it uses It is online that function/demand branch may be implemented in the technical solution of the flow cutting of the embodiment of the present invention, if want online one very Important function, a part of flow for going cutting seldom by using the flow segmentation technique scheme of the embodiment of the present invention is to newly In the version of function, so that whether on the one hand can sufficiently detect the function normal, on the other hand for other peripheral systems and Service does not influence particularly again, all flows is all cut into after only need to will not causing problem determining new function new On feature release, Efficient Development and the publication of dsp system ensure that.

In addition, also frequently occurring some specific demands of client in dsp system, for example client completes advertisement in this system Setting and dispensing when, advertisement is launched by way of bidding, and in this process, needs constantly to attempt algorithm model To determine reasonable charge and putting mode.The prior art realizes that the demand needs to do certain some specific crowd algorithm reality It tests, such as the male of 300,000 or more annual income, is directed to this partial discharge and is carried out independently in fact using specific algorithm newly developed It tests, but because realizing relative closure, is difficult to obtain true staining effect.And by using the flow cutting of the embodiment of the present invention Technical solution successfully can directly cut out qualified flow, and real time algorithm experiment is carried out in complete platform, Extraordinary experiment effect can be obtained, the realization data of truthfulness and integrality are obtained.

The embodiment of the invention provides a kind of flow segmentation and systems, pass through the combination of tree-like lookup and stochastic searching Realize the flow cutting of high efficient and flexible.Wherein, filtering and analysis for flow own characteristic are completed by tree-like lookup, The characteristics of this lookup is precisely to match, and scalability and maintainability are all very strong;And stochastic searching is mainly used for the cutting of flow, It needs the reasonable cutting of flow to be used to complete different functions, stochastic searching can ignore the own characteristic of flow and realize a view The cutting of colleague.Therefore the technical solution of the embodiment of the present invention realizes dynamic rationally under the premise of fully considering discharge characteristic Flow planning, can satisfy each flow user demand, floating resources configuration can be optimized, improve flow utilization rate and stream Measure cashability.

It should be understood that above-mentioned specific embodiment of the invention is used only for exemplary illustration or explains of the invention Principle, but not to limit the present invention.Therefore, that is done without departing from the spirit and scope of the present invention is any Modification, equivalent replacement, improvement etc., should all be included in the protection scope of the present invention.In addition, appended claims purport of the present invention Covering the whole variations fallen into attached claim scope and boundary or this range and the equivalent form on boundary and is repairing Change example.

Claims (6)

1. a kind of flow segmentation, which is characterized in that the method includes the steps:
Flow cutting request is received, the information of shunt volume to be cut in the request is obtained;
It is selected according to the information of the shunt volume to be cut and searches random tree;
The cutting condition value that the shunt volume to be cut is determined according to the information of the shunt volume to be cut, in the lookup random tree Carry out lookup matching;
It is matched as a result, determining the shunt volume to be cut according to the branch to match with the cutting condition value according to the lookup Assignment of traffic mode, according to shunt volume to be cut described in the assignment of traffic mode treatment, wherein the assignment of traffic mode packet Include the assignment of traffic ratio that the processing node of current branch is configured to;It is described according to the assignment of traffic mode treatment to Cutting flow include: according to the assignment of traffic mode determine assignment of traffic ratio as described in the processing node random process Shunt volume to be cut.
2. the method according to claim 1, wherein the method also includes steps:
Preset configuration file is read, at least one is generated according to the configuration file and searches random tree.
3. according to the method described in claim 2, it is characterized in that, including: in the configuration file
At least one node passes through level connection between the node and constitutes at least one tree structure;
The lookup random tree is generated by the tree structure.
4. a kind of flow cutting system, which is characterized in that the system comprises:
Data obtaining module obtains the information of shunt volume to be cut in the request for receiving flow cutting request;
Selected module is set, is selected according to the information of the shunt volume to be cut and searches random tree;
Matching module is searched, for determining the cutting condition value of the shunt volume to be cut according to the information of the shunt volume to be cut, Lookup matching is carried out in the lookup random tree;
Flow processing module, for matched as a result, according to the branch to match with the cutting condition value according to the lookup The assignment of traffic mode for determining the shunt volume to be cut, according to shunt volume to be cut described in the assignment of traffic mode treatment;
The flow processing module includes that ratio obtains module and random process module;
The ratio obtains module, the assignment of traffic ratio that the processing node for obtaining current branch is configured to;
The random process module, the assignment of traffic ratio for being determined according to the assignment of traffic mode is by the processing node Shunt volume to be cut described in random process.
5. system according to claim 4, which is characterized in that the system also includes:
Configuration module generates at least one according to the configuration file and searches at random for reading preset configuration file Tree.
6. system according to claim 5, which is characterized in that the configuration module includes:
Node configuration module constitutes at least one by level connection between the node and sets for configuring at least one node Shape structure;
Generation module is set, for generating the lookup random tree by the tree structure.
CN201610547429.1A 2016-07-12 2016-07-12 Flow segmentation and system CN106209663B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610547429.1A CN106209663B (en) 2016-07-12 2016-07-12 Flow segmentation and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610547429.1A CN106209663B (en) 2016-07-12 2016-07-12 Flow segmentation and system

Publications (2)

Publication Number Publication Date
CN106209663A CN106209663A (en) 2016-12-07
CN106209663B true CN106209663B (en) 2019-06-18

Family

ID=57477129

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610547429.1A CN106209663B (en) 2016-07-12 2016-07-12 Flow segmentation and system

Country Status (1)

Country Link
CN (1) CN106209663B (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1633111A (en) * 2005-01-14 2005-06-29 中国科学院计算技术研究所 High-speed network traffic flow classification method
CN101119321A (en) * 2007-09-29 2008-02-06 杭州华三通信技术有限公司 Network flux classification processing method and apparatus
CN104601486A (en) * 2013-10-30 2015-05-06 阿里巴巴集团控股有限公司 Method and device for shunt of network flow

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1633111A (en) * 2005-01-14 2005-06-29 中国科学院计算技术研究所 High-speed network traffic flow classification method
CN101119321A (en) * 2007-09-29 2008-02-06 杭州华三通信技术有限公司 Network flux classification processing method and apparatus
CN104601486A (en) * 2013-10-30 2015-05-06 阿里巴巴集团控股有限公司 Method and device for shunt of network flow

Also Published As

Publication number Publication date
CN106209663A (en) 2016-12-07

Similar Documents

Publication Publication Date Title
Hogan et al. Nist cloud computing standards roadmap
Weldon The future X network: a Bell Labs perspective
Rusitschka et al. Smart grid data cloud: A model for utilizing cloud computing in the smart grid domain
US20070087756A1 (en) Multifactorial optimization system and method
Hearn et al. From value chain to value creating ecology: Implications for creative industries development policy
CN105024865B (en) Cloud joint services
Kang et al. Enabling localized peer-to-peer electricity trading among plug-in hybrid electric vehicles using consortium blockchains
US9135068B2 (en) System and method for dynamic rescheduling of multiple varying resources with user social mapping
Simmhan et al. An analysis of security and privacy issues in smart grid software architectures on clouds
US20120246065A1 (en) Techniques for offering context to service providers utilizing incentives
Youssef Exploring cloud computing services and applications
WO2009020501A1 (en) Unwired enterprise platform
CN102483732A (en) Enterprise smart grid and demand management platform and methods for application development and management
Gai et al. A survey on FinTech
WO2010123553A1 (en) Mobile grid computing
CN103688244B (en) Method and apparatus for providing end to end privacy distributed computing
CN105247930A (en) Systems and methods for establishing wireless connections based on access conditions
US20130297698A1 (en) Method and Apparatus for Utilizing Context and Profile Information in Content Searching and Recommendation
Albakri et al. Security risk assessment framework for cloud computing environments
Yousefpour et al. All one needs to know about fog computing and related edge computing paradigms: A complete survey
Sun et al. Survey of service description languages and their issues in cloud computing
Kang et al. An interactive trust model for application market of the internet of things
CN101777792A (en) Safety communication system and method of electric network
US9117191B2 (en) Automatic device inventory management for different types of devices
Al Noor et al. Cellcloud: A novel cost effective formation of mobile cloud based on bidding incentives

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination
CB02 Change of applicant information
GR01 Patent grant