[go: up one dir, main page]

CN109117321B - A kind of full link application moving method of cloud platform based on label figure - Google Patents

A kind of full link application moving method of cloud platform based on label figure Download PDF

Info

Publication number
CN109117321B
CN109117321B CN201810846524.0A CN201810846524A CN109117321B CN 109117321 B CN109117321 B CN 109117321B CN 201810846524 A CN201810846524 A CN 201810846524A CN 109117321 B CN109117321 B CN 109117321B
Authority
CN
China
Prior art keywords
application
link
node
full
abnormal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201810846524.0A
Other languages
Chinese (zh)
Other versions
CN109117321A (en
Inventor
吕晨
张海杰
吕蕾
刘弘
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.)
Pulian Intelligent Data Construction Technology Beijing Co ltd
Original Assignee
Shandong Normal University
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 Shandong Normal University filed Critical Shandong Normal University
Priority to CN201810846524.0A priority Critical patent/CN109117321B/en
Publication of CN109117321A publication Critical patent/CN109117321A/en
Application granted granted Critical
Publication of CN109117321B publication Critical patent/CN109117321B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/07—Responding to the occurrence of a fault, e.g. fault tolerance
    • G06F11/16—Error detection or correction of the data by redundancy in hardware
    • G06F11/20—Error detection or correction of the data by redundancy in hardware using active fault-masking, e.g. by switching out faulty elements or by switching in spare elements
    • G06F11/202—Error detection or correction of the data by redundancy in hardware using active fault-masking, e.g. by switching out faulty elements or by switching in spare elements where processing functionality is redundant
    • G06F11/2023—Failover techniques
    • G06F11/203—Failover techniques using migration
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/3055—Monitoring arrangements for monitoring the status of the computing system or of the computing system component, e.g. monitoring if the computing system is on, off, available, not available
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/34—Recording or statistical evaluation of computer activity, e.g. of down time, of input/output operation ; Recording or statistical evaluation of user activity, e.g. usability assessment
    • G06F11/3447—Performance evaluation by modeling

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Computer Hardware Design (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

本发明公开了一种基于标记图的云平台全链路应用迁移方法，该方法对微应用数据进行标记图建模，然后根据模型进行全链路可通性判定，生成状态转移图，利用该图对全链路应用全局运行状态进行观测和计算并执行异常检测，定位异常应用及应用源，进而求取异常链路子图并对异常链路子图进行替换，有效完成全链路应用迁移。

The invention discloses a cloud platform full-link application migration method based on a marker graph. The method models micro-application data on a marker graph, and then determines the connectivity of the whole link according to the model to generate a state transition graph. The graph observes and calculates the global running state of the full-link application, performs anomaly detection, locates the abnormal application and application source, then obtains the abnormal link subgraph and replaces the abnormal link subgraph, effectively completing the full-link application migration .

Description

A kind of full link application moving method of cloud platform based on label figure
Technical field
The invention belongs to software developments and automated maintenance field, specifically, being to be related to a kind of cloud based on label figure The full link application moving method of platform.
Background technique
Currently, the other cloud platform application distribution of data-oriented centre grade and operating status show new variation tendency: one Aspect, using no longer binding with specific hardware resource, the parameter for influencing application state increases sharply and is closely connected each other, increases The complexity of monitored item and triggering alarm rule；On the other hand, after service application presses micro- Applications construct mode cloud, micro- application Increasing number, while running boundary dynamic change, the distribution of application and operation range increase, cause to have between Distributed Application Complicated calling dependence, the full link for constituting the call chain formation that different starting points is applied to clearing end application are answered With.
Conventional needle stands in the position of " part monitoring ", it is difficult to answer to " point-to-point " monitoring method of hardware and middleware Variation compound to the multi-parameter of above-mentioned application state simultaneously feeds back the global operating status that chain is applied under distributed structure/architecture.Therefore, base In the application moving method of " point-to-point " monitoring abnormal problem diagnosis and stale link can not be carried out based on above-mentioned global operating status Road analysis, finally will lead to the transfer ability deficiency of the full link application for data center's grade, and the business of full link application needs It asks variation to be unable to get quick response and completes real-time migration.
Summary of the invention
The present invention to solve the above-mentioned problems, proposes a kind of full link application of the cloud platform based on label figure migration side Figure modeling is marked to micro- application data in method, this method, then according to model carry out full link can the general character determine, generate state Transfer figure is observed and calculates and execute abnormality detection to the global operating status of full link application using the figure, and positioning is abnormal Using and application source, and then seek anomaly link subgraph and it be replaced, efficiently accomplish full link application migration.
To achieve the goals above, the present invention adopts the following technical scheme:
A kind of full link application moving method of cloud platform based on label figure, includes the following steps:
Step (1): the micro- of input applies data, builds to micro- label figure for carrying out operating status and dependence using data Mould, building label graph model；
Step (2): full link is calculated according to the invocation pattern and state parameter applied in label graph model and is applied in reality Global operating status in operational process；
Step (3): carrying out abnormality detection the global operating status of full link application, if abnormal, execute step (4), no Then, step (2) are executed；
Step (4): abnormal application source is positioned according to abnormality detection result, seeks anomaly link subgraph and normal link replacement Subgraph replaces anomaly link subgraph with normal link replacement subgraph, completes full link application migration.
Further, the method for the step 1 specifically:
Step (1-1): it is modeled using data dependence relation:
The mapping relations of building label graph structure model and application data: the node in label graph model indicates to apply, mark Remember that the side in graph structure model indicates matched using call relation based on application interface parameter；
Step (1-2): global state parameter model is applied:
The mapping relations of building label figure state parameter and application operating status: each node is associated with shape in label graph model State parameter list records global running state parameter of the node in full link application chain；
Step (1-3): building label graph model: the mapping relations established using step (1-1), step (1-2), building mark Note figure G=(M, E, L)；Wherein, M indicates the set of application, and E indicates that the interface parameters of application matches set, and L indicates the complete of application Office's running state parameter.
Further, the method for the step 2 specifically:
Step (2-1), judge to apply in full link it is complete can the general character, search from starting point application node msIt is answered to clearing end With node mtQuan Ketong application in link；
Step (2-2) is successively accessed and is answered from starting point application node to clearing end using BFS (breadth first traversal) algorithm With the Quan Ketong application in the link of node, the label figure comprising call relation between application and Quan Ketong application can be led to entirely is generated, It generates from starting point application node msTo clearing end application node mtFull link state shift figure；
Step (2-3) is calculated using aggregate function from starting point application node msTo clearing end application node mtFull link The full link comprising multiple micro- application nodes in state transition diagram applies the global operating status in actual moving process, institute The actual observed value that the global operating status value for calculating and obtaining is known as the complete global operating status of link application is stated, is referred to as observed Value.
Further, the step (2-1) carries out full link passability judgement to application, is determined using counter mechanism The complete of application can the general character, specially step are as follows:
Step (2-1-1) determines starting point application node msTo clearing end application node mtFull link application；
Step (2-1-2) is each application build counter in full link, sets the initial value of counting as each application The number of interface input parameter；
Step (2-1-3) determines the interface input parameter of each application in full link: successively inquiry application section Point msEach interface inputs parameter, if the m inquiredsAn interface input parameter can be by node msForerunner Quan Ketong It using output, then enables Counter Value subtract 1, inquires msTotal interface input parameter, judge msCounter Value, if it is zero, when Front nodal point msIt is judged as Quan Ketong node, otherwise msIt is not Quan Ketong node；Full link is successively judged according to the method for this step In each application it is complete can the general character, until having handled from starting point using msTo clearing end application mtFull link application.
Further, the method that the step (3) executes abnormality detection to the global operating status of full link application are as follows:
The global operating status of full link application is modeled using temporal model, autoregression mould is utilized according to modeling data Type seeks full link and applies predicted value at the accordingly observation moment, the then distinctiveness ratio of predicted value and observation, if distinctiveness ratio is big In given threshold, then the complete global operating status of link application is abnormal.
Further, the specific steps of the method that the global operating status of full link application is carried out abnormality detection are as follows:
Step (3-1) applies global operating status time series modeling to full link: by the resulting full link application of step (2) Global operating status L (mn, j) and in the observation that the preceding n period accordingly observes the moment it is modeled as time series data LT+v L2T+v
..., LnT+v, wherein T is state observation periodicity, and at the time of v is that current period is observed, n is observation wheel Number, and have 1≤v≤T；
Step (3-2) measures the global operating status distinctiveness ratio of full link application, utilizes i rank according to time series modeling data Autoregression model seeks the predicted value that the complete global operating status of link application accordingly observes the moment in current (n-th) period L'nT+v, the complete global operating status of link application is accordingly observed different between the predicted value at moment and observation in current period Angle value is Q=| LnT+v-L'nT+v|。
Step (3-3), to full link application abnormal test, given threshold λ makes the following judgment: H0:Q>λ,H1:Q<λ；
Wherein λ is the threshold value of setting, if H0It sets up, then detects using abnormal；If H1It sets up, then application is not detected It is abnormal.
Step (3) method carried out abnormality detection to the global operating status of full link application can be used for full chain The global operating status of single micro- application in road executes abnormality detection, and detection method is consistent.
Further, the method that abnormal application source is positioned according to abnormality detection result are as follows:
If starting point application node msTo terminal applies node mtEnd-to-end full link application to be detected as full link abnormal Using then: initialization stack first enables node mtStacking, stack top node mv=mt, and from mvStart backtracking search mvWhole entirely may be used Logical predecessor node mp；The method detection node m that the global operating status of full link application is carried out abnormality detection using step (3)pEntirely Whether office's operating status is abnormal: if mpFor exception, then m is enabledpStacking and stack top node updates are mv=mpAnd iteration executes aforementioned behaviour Make, otherwise mvFor anomaly source application node.
Further, anomaly link subgraph and normal link replacement subgraph are sought, is replaced with normal link replacement subgraph migration Anomaly link subgraph is changed, the method specific steps of full link application migration are completed are as follows:
It seeks anomaly link subgraph: enabling the anomaly source application node m soughtvImmediate successor node mwFor Quan Ketong node, It generates from starting point application node msStart through anomaly source application node mvTo its immediate successor application node mwState transition diagram G'；Defining anomaly link subgraph is from starting point application node msStart the transition through abnormal nodes to its immediate successor node to turn Figure is moved, G' is the anomaly link subgraph sought.The link subgraph is state transition diagram.
The replacement of anomaly link subgraph: m is enabledwFor abnormal source node mvAll immediate successor nodes, from mwIt is applied to starting point Node msAlong the Quan Ketong predecessor node of state transition diagram recursive backtracking search present node to generate normal link replacement subgraph G " completes full link application migration with normal link replacement subgraph G " migration replacement anomaly link subgraph.
Further, the method for normal link replacement subgraph G " is generated specifically:
Stack is initialized, successively by abnormal source node mvAll immediate successor nodes be set as present node, grasped as follows Make: present node m is abnormal source node mvOne of immediate successor node, recursive backtracking search for the complete of present node m can Logical predecessor node mp, judge whether to meet mp∈ G' enables m if NOpStacking；If so, not stacking；
Judge whether stack top element is ms, if ms, successively accessed using BFS (breadth first traversal) algorithm from starting point Application node is applied through application node in stack to the Quan Ketong in clearing end application node link, is generated from starting point application node msStart through stack interior joint to mwNormal link replace subgraph G "=(M ", E ", L ").
A kind of full link application migratory system of cloud platform based on label figure, comprising:
It is micro- to apply data input module, building is modeled to micro- label figure for carrying out operating status and dependence using data Mark the module of graph model；
The complete global operating status computing module of link application；
The module of abnormality detection is executed to the global operating status of full link application；
The module for positioning abnormal application source seeks the module of anomaly link subgraph and normal link replacement subgraph, will be abnormal Link subgraph replaces with the transferring module of normal link replacement subgraph.
It is described micro- to be modeled using data input module, to micro- label figure for carrying out operating status and dependence using data Building marks the module of graph model, the global operating status computing module of full link application, to the global operating status of full link application The module for executing the module of abnormality detection, positioning abnormal application source, seeks anomaly link subgraph and normal link replaces subgraph The transferring module that anomaly link subgraph replaces with normal link replacement subgraph is sequentially connected by module.
Compared with prior art, the invention has the benefit that
(1) present invention applies data to micro- by a kind of method of the full link application migration of cloud platform based on label figure Figure modeling is marked, using model built carry out full link can the general character determine, generate full link state transfer figure, utilize the figure The global operating status of full link application is observed and predicted value is combined to carry out abnormal state detection, positioning is abnormal to apply and answer With source, and then seeks anomaly link subgraph and realize full link application ground effective mobility.
(2) method of the invention is proposed based on label figure to micro- application dependence Formal Modeling, theoretically can be more preferable Ground describes micro- full link property relied on using intrinsic call under cloud distributed structure/architecture.Meanwhile this method passes through on label figure State transfer search technology can solve the critical issue that the complete global running state information of link application calculates, and overcome traditional " point To point " local monitor mode the shortcomings that can not feeding back under cloud distributed structure/architecture micro- application and its full link call chain operating status, It can be realized the abnormality detection to the application of full link and effectively and correctly repair anomaly link, enhance moving for full link application Shifting ability.
Detailed description of the invention
The accompanying drawings constituting a part of this application is used to provide further understanding of the present application, and the application's shows Meaning property embodiment and its explanation do not constitute the restriction to the application for explaining the application.
Fig. 1 is full link application moving method flow chart of the invention；
Fig. 2 is the label graph model schematic diagram of the embodiment of the present invention.
Specific embodiment:
The invention will be further described with embodiment with reference to the accompanying drawing.
It is noted that described further below be all exemplary, it is intended to provide further instruction to the application.Unless another It indicates, all technical and scientific terms used herein has usual with the application person of an ordinary skill in the technical field The identical meanings of understanding.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root According to the illustrative embodiments of the application.As used herein, unless the context clearly indicates otherwise, otherwise singular Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
Relevant technical terms involved in the present invention are illustrated first:
Application of the present invention refers to micro- application.
The complete global operating status of link application: the operating status of all micro- applications is according to cumulative, minimum in full link application The calculated polymerizing value of the rule such as value or maximum value reflects the overall operation shape that full link application includes all micro- applications State, such as total response time, the such as total network throughput are complete, and the global operating status of link application is eventually equal in full link application The global operating status of terminal node.
Link subgraph: utilize BFS (breadth first traversal) algorithm from starting point application node msTo clearing end application node mt Successively access label figure in Quan Ketong application, it is generated comprising can lead to full application and its call relation directed acyclic graph and Its subgraph.The directed acyclic graph includes from msTo mtFull link application.
Normal link subgraph: from starting point application node msTo clearing end application node mtThe global operation of full link application When state is detected as normal, respective links subgraph and its contained any directed acyclic subgraph are normal link subgraph.
Anomaly link subgraph: from starting point application node msTo clearing end application node mtThe global operation of full link application When state is detected as abnormal, from starting point application node m in respective links subgraphsStarted through anomaly source application node mv To its immediate successor application node mwThe directed acyclic subgraph of composition is defined as anomaly link subgraph.
The global running state parameter of application: for indicating and recording the global operating status of current application node (as rung Between seasonable, cpu busy percentage etc.), polymerization can be carried out by step (2-3) to be calculated.
The interface parameters of application matches set: given two are applied maAnd mb, work as maOutput parameter type and mbOne When input parameter type is identical, claim to apply maInterface parameters with apply mbInterface parameters matching, mark all applications in figure The collection that interface parameters matching is constituted is collectively referred to as the interface parameters matching set of application.
The complete of node can the general character: given label figure, once from starting point application node msTo clearing end application node mt's In the search process of full link application, for the arbitrary node m being accessed in figurei, miQuan Ketong is and if only if miInput parameter All by miThe output of Quan Ketong predecessor node, then miMeet node it is complete can the general character, then miFor Quan Ketong node.It is searching for Under original state, starting endpoint msIt is defaulted as Quan Ketong node.
Full link state transfer figure: BFS (breadth first traversal) algorithm based on figure is successively accessed from msTo mtScheme in label In Quan Ketong node, until in figure accessed node include node mtUntil.What final traversal generated only includes Quan Ketong node Directed acyclic subgraph be from msTo mtFull link state shift figure, which, which reflects full link, applies and called mutually The process of state propagation and polymerization in journey.
Following embodiments are a kind of typical embodiment of the application, as shown,
Implementation method of the invention is described in detail below with reference to Fig. 1.It is worth noting that, in this flow chart, The S ∈ Ps, wherein Ps is stack PnormalStack；T < the Tm, wherein Tm is end time Terminal.
A kind of full link application moving method of cloud platform based on label figure, includes the following steps:
Step (1): the micro- of input applies data, builds to micro- label figure for carrying out operating status and dependence using data Mould, building label graph model；
Step (2): the complete global fortune of link application is calculated according to the invocation pattern and state parameter applied in label graph model Row state；
Step (3): carrying out abnormality detection the global operating status of full link application, if abnormal, execute step (4), no Then, step (2) are executed；
Step (4): abnormal application source is positioned according to abnormality detection result, seeks anomaly link subgraph and normal link replacement Subgraph replaces anomaly link subgraph with normal link replacement subgraph, completes full link application migration.
In embodiments of the present invention, the step (1) inputs micro- using data first, is then marked figure and builds Mould.The modeling method that figure is marked includes the following steps:
Step (1-1) is modeled using data dependence relation: building label graph structure mode-applies data mapping relations, mark Note graph model interior joint indicates application, and side indicates matched using call relation based on application interface parameter.
Step (1-2) applies global state parameter model: building label figure state parameter-is closed using operating status mapping System marks each node in graph model to be associated with a state parameter table, records the overall situation of the node in full link application chain Running state parameter L.
Step (1-3) building label graph model: the mapping relations established using step (1-1), step (1-2), building mark Note figure G=(M, E, L).In label figure, node collection M indicates set of applications, Wherein, k is node mkMark, I and O respectively indicate mkThe input parameter type and output parameter type of corresponding application.Directed edge Collecting E indicates that application parameter matching set, the set meet:Wherein, muAnd mvIt is respectively The head node and tail node on side, muThe m for beingvDirect precursor, mvIt is muImmediate successor, muCorresponding application matches mvIt is corresponding to answer With.ekLabelMeet: (1)(2)In addition, each node is associated in label figure One global Running Status Table works as prosthomere under conditions of the global operating status of its known predecessor node for particularly pointing out The global operating status L (m of pointn, j) and to describe with mnUsing the status information of j-th of type of the full link application for terminal.
As shown in Fig. 2, being illustrated using the operating status of response time as j-th of type, it is known that node m1, m2, m3, m4, m5, m6Response time is respectively 10ms, 15ms, 12ms, 8ms, 16ms and 14ms.Thus each node reality etc. can be calculated To the response time, wherein t is indicated comprising to apply mnFor the m of terminaln-1The global state information of a application, f expression do not include To apply mnFor the m of terminaln-1The global state information of a application.With node m6For, it is known that node m3And m5With node m6For Terminal, because there are four types of situations, i.e. node m for this type j value3、m5Value t, t respectively, t, f, f, t and f, f.It is taken for the first It is worth situation, node m3With node m5Respectively when value t, t, node m6Response time be 30ms (m6+m5), due to node m3And m5 It is connected in parallel to node m6, therefore only take the longest response time.For second situation, node m3、m5Respectively when value t, f, section Point m6Response time be 26ms (m3+m6), and so on.
In embodiments of the present invention, described according to built label graph model, to the full link overall operation state of cloud platform Calculate and includes the following steps:
The full link of step (2-1) can be determined by node: after establishing label figure, give starting point msTo end point mt's One full link application, need to generate from msTo mtAll related application states can be by node along full link, the state propagated Transfer figure, with the application of determination full link end to end state shift it is traversed all using invocation pattern.It is called according to application There are a variety of graph structures such as sequence, convergence, branch in the label figure of Model Establishment, it is different from the definition of the path accessibility of node, From msTo mtNode m in full link entirely can the general character be defined as the interface parameters of m all by forerunner's Quan Ketong node m of mpIt is defeated Out, it may be assumed that m.I=mp.O。
Quan Ketong node determination method is as follows: it is assumed that starting endpoint msFor initial Quan Ketong node.From msTo mtSearch It in the process can the general character using the complete of counter mechanism predicate node.It is each node building counter in calculating, it is initial sets it Value is the input number of parameters # (m.I) of node.If there is parameter that can be exported by forerunner's Quan Ketong node in search process, meter is enabled Number device value subtracts one, i.e. # (m.I) --.As Counter Value # (m.I)=0, since all parameters are exported, node m is determined For Quan Ketong.
The full link state transfer figure of step (2-2) generates: BFS (breadth first traversal) algorithm based on figure successively access from msTo mtThe Quan Ketong node determined in the step (2-1) in label figure, method particularly includes: 1. accessed node m firsts, and will Its access flag, which is set to, to be accessed, i.e. visited [msG={ M, E } is schemed in]=1, initialization, wherein M=ms, E is empty set；② Then vertex m is successively accessedsImmediate successor Quan Ketong nodeThen enableE isUpdate figure G={ M, E }；3. and then successively accessingIn immediate successor m The Quan Ketong node of visited [m]=0, and by 2. middle policy update figure G={ M, E }；And so on, until accessing section in figure Point includes node mtUntil.The only subgraph G comprising Quan Ketong node that final traversal generates is from msTo mtFull link state Transfer figure.
The complete global operating status of link application of step (2-3) is observed and is calculated: being generated full link state by step (2-2) and is turned Move Quan Ketong node m in figurenThe global operating status of full link application be defined as from msTo mnQuan Ketong predecessor node current The global state of the observation operating status value of moment v shifts L (mn,j).Consider a variety of graph structures in full link state transfer figure (as previously described), the computation rule that can define the global operating status of full link application of correlation graph structure are respectively as follows: L sequence L mergesL bifurcated (mk, j) and=F1 (L(m,j),Qk,j)(1≤k≤n).Wherein, mnFor the Quan Ketong terminal note under corresponding graph structure, j is Status Type label, Qi,j For the jth class operating status of node i, F1It indicates aggregate function set { ∑, min, ∏, max }.It is specific to calculate, according to from msIt arrives mnThe full link state transfer figure graph structure integrated use above three formula that includes carry out polymerization full link be calculated answering With global operating status L (mn,j).In the present invention, the global state of the full link application of j-th of type is defined as L (mt,j).It needs It is to be noted, that the global operating status of the full link application of the present embodiment is equal to it and terminates end node m according to above-mentioned computation rulet Global operating status.It can be used as the termination end segment during intermediate search due to individual node any in full link application Point, step (2-3) can also be used for the calculating of the global operating status of any individual node application in full link.
In embodiments of the present invention, described that abnormality detection is executed to the global operating status of full link application, using timing Model models the global operating status of full link application in the observation that the preceding n period accordingly observes the moment, according to modeling Data using autoregression model seek full link and apply accordingly to observe the predicted value at moment in current (n-th) period, then sentence Disconnected current (n-th) period accordingly observes the distinctiveness ratio of moment predicted value and observation, if distinctiveness ratio is greater than given threshold, full chain Application global operating status in road is abnormal, the specific steps are as follows:
Step (3-1) applies global operating status time series modeling to full link: by the resulting full link application of step (2) Global operating status L (mn, j) and in the observation that n period accordingly observes the moment it is modeled as time series data LT+v, L2T+v..., LnT+v, wherein T is state observation periodicity, and at the time of v is that current period is observed, n is observation wheel number, and has 1≤v ≤T；
The global operating status distinctiveness ratio measurement of the full link application of step (3-2): it is based on the full link application of preceding i periodic knot Observation L of the observation of global operating status to current periodnT+VIt is predicted, utilizes i rank certainly according to time series modeling data Regression model obtains predicted value L'nT+v=aiL(n-i)T+v+ai-1L(n-i-1)T+v+......+a2L(n-2)T+v+a1L(n-1)T+v+a0, wherein a0It is constant term, a1,…,aiIt is model parameter.And then based on the predicted value and sight for measuring the complete global operating status of link application Different angle value between measured value is Q=| LnT+v-L'nT+v|。
The full link application abnormal test of step (3-3): it is based on it is assumed hereinafter that examining to application execution abnormality detection:
H0:Q>λ,H1:Q<λ；
Wherein λ is the threshold value of setting, if H0It sets up, then detects using abnormal；If H1It sets up, then application is not detected It is abnormal, and go to step (2-3) and carry out full link state calculating.
In embodiments of the present invention, the step (4) carries out according to testing result using abnormity diagnosis and impact analysis, Full link application migration scheme is provided to include the following steps:
Step (4-1) anomaly source application positioning, from msTo mtEnd-to-end full link application be detected as full link and answer extremely With then initialization stack PabnormalStack first, enables node mtStacking, at this time stack top node mv=mt, and from mvStart to recall Search for its whole Quan Ketong predecessor node mp；Secondly detection mpWhether global operating status is abnormal: if mpFor exception, then m is enabledpEnter Stack, and stack top node updates are mv=mpAnd iteration executes aforementioned operation, otherwise mvFor anomaly source application node, above-mentioned iteration mistake Journey terminates.
Step (4-2) seeks anomaly link subgraph, enables the abnormal source node m sought in step (4-1)vImmediate successor be Quan Ketong node is searched for based on step (2-2), is generated from starting point application node msStart through anomaly source application node mvArrive it Immediate successor application node mwState transition diagram G'；Defining anomaly link subgraph is from starting point application node msStart through different Figure is shifted in the transition of Chang Jiedian to its immediate successor node, and G' is the anomaly link subgraph sought.The link subgraph is State transition diagram.
The replacement of step (4-3) anomaly link subgraph, enables mwFor abnormal source node mvAll immediate successor node (direct shadows Ring node), from mwTo msQuan Ketong predecessor node along state transition diagram recursive backtracking search present node m is to generate replacement Figure: if recalling the predecessor node m of present node m in search processp∈ G'(anomaly link subgraph), then select other forerunners of m Node backtracking search；Otherwise, m is enabledpStacking PnormalStack.Work as mp(present node is unmatched in state transition diagram when for sky Drive node), above-mentioned trace-back process terminates.
Judge whether PnormalStack stack top element is ms, if ms, then new slave m is generated using step (2-1)sIt opens Begin through PnormalStack stack interior joint to mwNormal link replace subgraph G "=(M ", E ", L "), with replacement from msStart through Anomaly source mvTo mwAnomaly link subgraph G'=(M', E', L'), i.e., original full link subgraph G=(M, E, L) migration be G ∧= (M ∧, E ∧, L ∧), wherein M ∧=(M ∪ M' ∪ M ")-(M ∩ M'), E ∧=(E ∪ E' ∪ E ")-(E ∩ E'), L ∧= L(n+1)T+v.If current time is less than specified end time Terminal, the observation of subsequent time, the sight of the present embodiment setting are carried out The time interval of survey is 1s, then t=t+1, goes to step (2-3) and executes full link state calculating, executes the monitoring and observation of a new round And migration.
The foregoing is merely preferred embodiment of the present application, are not intended to limit this application, for the skill of this field For art personnel, various changes and changes are possible in this application.Within the spirit and principles of this application, made any to repair Change, equivalent replacement, improvement etc., should be included within the scope of protection of this application.
Above-mentioned, although the foregoing specific embodiments of the present invention is described with reference to the accompanying drawings, not protects model to the present invention The limitation enclosed, those skilled in the art should understand that, based on the technical solutions of the present invention, those skilled in the art are not Need to make the creative labor the various modifications or changes that can be made still within protection scope of the present invention.

Claims (10)

1.一种基于标记图的云平台全链路应用迁移方法，其特征在于，包括如下步骤：1. a cloud platform full-link application migration method based on marker graph, is characterized in that, comprises the steps: 步骤(1)：输入的微应用数据，对微应用数据进行运行状态和依赖关系的标记图建模，构建标记图模型；Step (1): with the input micro-application data, carry out labeling graph modeling of the running state and dependencies of the micro-application data, and construct a labeling graph model; 步骤(2)：根据标记图模型中应用的调用模式和状态参数计算全链路应用在实际运行过程中的全局运行状态；Step (2): Calculate the global running state of the full-link application in the actual running process according to the calling mode and state parameters of the application in the markup graph model; 步骤(3)：对全链路应用全局运行状态进行异常检测，如果异常，执行步骤(4)，否则，执行步骤(2)；Step (3): perform abnormality detection on the global operating state of the full-link application, if abnormal, perform step (4), otherwise, perform step (2); 步骤(4)：根据异常检测结果定位异常应用源，求取异常链路子图和正常链路替换子图，用正常链路替换子图替换异常链路子图，完成全链路应用迁移。Step (4): locate the abnormal application source according to the abnormal detection result, obtain the abnormal link subgraph and the normal link replacement subgraph, replace the abnormal link subgraph with the normal link replacement subgraph, and complete the full-link application migration. 2.如权利要求1所述的一种基于标记图的云平台全链路应用迁移方法，其特征在于：所述步骤1的方法具体为：2. The method for migrating cloud platform full-link applications based on a marker graph as claimed in claim 1, characterized in that: the method of the step 1 is specifically: 步骤(1-1)：应用数据依赖关系建模：Step (1-1): Apply data dependency modeling: 构建标记图结构模型与应用数据的映射关系：标记图模型中的节点表示应用，标记图结构模型中的边表示基于应用接口参数匹配的应用调用关系；Build the mapping relationship between the labelled graph structure model and application data: the nodes in the labelled graph model represent applications, and the edges in the labelled graph structure model represent the application invocation relationship based on the matching of application interface parameters; 步骤(1-2)：应用全局状态参数建模：Step (1-2): Apply global state parameter modeling: 构建标记图状态参数与应用运行状态的映射关系：标记图模型中每个节点关联状态参数表，记录该节点在全链路应用链条中的全局运行状态参数；Build the mapping relationship between the marker graph state parameters and the application running state: each node in the marker graph model is associated with a state parameter table, and records the global running state parameters of the node in the full-link application chain; 步骤(1-3)：构建标记图模型：采用步骤(1-1)、步骤(1-2)建立的映射关系，构建标记图G＝(M,E,L)；其中，M表示应用的集合，E表示应用的接口参数匹配集合，L表示应用的全局运行状态参数。Step (1-3): Constructing a labeled graph model: Using the mapping relationship established in steps (1-1) and (1-2), construct a labeled graph G=(M, E, L); Set, E represents the interface parameter matching set of the application, and L represents the global running state parameter of the application. 3.如权利要求1所述的一种基于标记图的云平台全链路应用迁移方法，其特征在于：所述步骤2的方法具体为：3. a kind of cloud platform full-link application migration method based on mark graph as claimed in claim 1, is characterized in that: the method of described step 2 is specifically: 步骤(2-1)、判断全链路中应用的全可通性，查找从起始端应用节点ms到终止端应用节点mt的链路中的全可通应用；Step (2-1), judge the fully accessible application of the application in the full link, and search for the fully accessible application in the link from the starting end application node m s to the terminating end application node m t ; 步骤(2-2)、利用BFS(广度优先遍历)算法依次访问从起始端应用节点到终止端应用节点的链路中的全可通应用，生成包含全可通应用及全可通应用间调用关系的标记图，即生成从起始端应用节点ms到终止端应用节点mt的全链路状态转移图；Step (2-2), use the BFS (Breadth First Traversal) algorithm to sequentially access the all-communicable applications in the link from the starting-end application node to the terminating-end application node, and generate the all-communicable application and the all-communicable application call The labeling graph of the relationship, that is, generating a full-link state transition graph from the application node m s at the starting end to the application node m t at the terminating end; 步骤(2-3)、采用聚合函数计算从起始端应用节点ms到终止端应用节点mt的全链路状态转移图中的节点应用的全链路应用在实际运行过程中的全局运行状态。Step (2-3), using an aggregate function to calculate the global operating state of the full-link application of the node application in the full-link state transition diagram from the start-end application node m s to the end-end application node m t in the actual operation process . 4.如权利要求3所述的一种基于标记图的云平台全链路应用迁移方法，其特征在于：所述步骤(2-1)对应用进行全链路可通过性判定，采用计数器机制判定应用的全可通性，具体为步骤为：4. The method for migrating cloud platform full-link applications based on a marker graph as claimed in claim 3, wherein in the step (2-1), the full-link passability judgment of the application is performed, and a counter mechanism is adopted. To determine the full accessibility of an application, the specific steps are: 步骤(2-1-1)、确定起始端应用节点ms到终止端应用节点mt的全链路应用；Step (2-1-1), determine the full-link application from the starting end application node ms to the terminating end application node m t ; 步骤(2-1-2)、为全链路中每个应用构建计数器，设定计数的初始值为每个应用的接口输入参数的数目；Step (2-1-2), build a counter for each application in the full link, and set the initial value of the count to the number of interface input parameters of each application; 步骤(2-1-3)、对全链路中的每个应用的接口输入参数进行判定：依次查询应用节点ms每个接口输入参数，如果查询到的ms的一个接口输入参数能够被节点ms的前驱全可通应用输出，则令计数器值减1，查询ms的所有接口输入参数，判断ms的计数器值，如果为零，当前节点ms被判定为全可通节点，否则ms不是全可通节点；按照本步骤的方法依次判断全链路中的每个应用的全可通性，直到处理完从起始端应用ms到终止端应用mt的全链路应用。Step ( 2-1-3 ), determine the interface input parameters of each application in the entire link: query each interface input parameter of the application node ms in turn, if an interface input parameter of the queried ms can be If the output of the precursor of node ms is fully accessible, the counter value is decremented by 1, and all interface input parameters of ms are queried to determine the counter value of ms. If it is zero, the current node ms is determined as a fully accessible node. Otherwise, ms is not an all-connectable node; according to the method in this step, the all-connectivity of each application in the full link is judged in turn, until the full-link application from the starting end application ms to the terminating end application m t is processed . . 5.如权利要求1所述的一种基于标记图的云平台全链路应用迁移方法，其特征在于：5. a kind of cloud platform full-link application migration method based on marker graph as claimed in claim 1, is characterized in that: 进一步的，所述步骤(3)对全链路应用全局运行状态执行异常检测的方法为：采用时序模型对全链路应用全局运行状态进行建模，根据建模数据利用自回归模型求取全链路应用在相应观测时刻的预测值，然后预测值和观测值的相异度，若相异度大于设定阈值，则全链路应用全局运行状态异常。Further, the method for performing abnormality detection on the global operating state of the full-link application in the step (3) is as follows: using a time sequence model to model the global operating state of the full-link application, and using an autoregressive model to obtain the global operating status according to the modeling data. The predicted value of the link application at the corresponding observation time, and then the dissimilarity between the predicted value and the observed value. If the dissimilarity is greater than the set threshold, the global running state of the full-link application is abnormal. 6.如权利要求5所述的一种基于标记图的云平台全链路应用迁移方法，其特征在于：所述对全链路应用全局运行状态进行异常检测的方法的具体步骤为：6. A marker graph-based cloud platform full-link application migration method as claimed in claim 5, characterized in that: the specific steps of the method for performing anomaly detection on the global operating state of the full-link application are: 步骤(3-1)、对全链路应用全局运行状态时序建模：将步骤(2)所得的全链路应用全局运行状态L(mn,j)在n个周期相应观测时刻的观测值建模为时序数据LT+v，L2T+v，...，LnT+v，其中，T是状态观测周期数，v(1≤v≤T)是当前周期进行观测的时刻，n为观测轮数，L(mn,j)描述以mn应用为终点的全链路应用的第j个类型的状态信息；Step (3-1), modeling the global operating state of the full link application: the observed value of the global operating state L(m n ,j) obtained in step (2) at the corresponding observation time of n cycles Modeled as time series data L T+v , L 2T+v , ..., L nT+v , where T is the number of state observation periods, v(1≤v≤T) is the moment of observation in the current period, n is the number of observation rounds, L(m n ,j) describes the j-th type of state information of the full-link application with the m n application as the end point; 步骤(3-2)、对全链路应用全局运行状态相异度测量，利用i阶自回归模型求取全链路应用全局运行状态在对应观测时刻的预测值L'nT+v，全链路应用全局运行状态的预测值和观测值之间的相异度值为Q＝|LnT+v-L'nT+v|；Step (3-2), measure the dissimilarity of the global operating state of the full-link application, and use the i-order autoregressive model to obtain the predicted value L' nT+v of the global operating state of the full-link application at the corresponding observation time, and the full chain The dissimilarity between the predicted value and the observed value of the global operating state of the road application is Q=|L nT+v -L' nT+v |; 步骤(3-3)、对全链路应用异常检验，设定阈值λ，进行如下判断：H0:Q>λ,H1:Q＜λ；Step (3-3), apply abnormality check to the whole link, set a threshold λ, and judge as follows: H 0 : Q>λ, H 1 : Q<λ; 其中λ为设定的阈值，若H0成立，则检测出应用异常；若H1成立，则没有检测出应用异常。Among them, λ is the set threshold. If H 0 is established, application abnormality is detected; if H 1 is established, application abnormality is not detected. 7.如权利要求1所述的一种基于标记图的云平台全链路应用迁移方法，其特征在于：所述根据异常检测结果定位异常应用源的方法具体为：7. The marker graph-based cloud platform full-link application migration method according to claim 1, wherein the method for locating an abnormal application source according to an abnormal detection result is specifically: 若起始端应用节点ms到终端应用节点mt的端到端全链路应用被检测为全链路异常应用，则：首先初始化栈，令节点mt入栈，栈顶节点mv＝mt，并从mv开始回溯搜索mv的全部全可通前驱节点mp；采用步骤(3)对全链路应用全局运行状态进行异常检测的方法检测节点mp全局运行状态是否异常：若mp为异常，则令mp入栈且栈顶节点更新为mv＝mp并迭代执行前述操作，否则mv为异常源应用节点。If the end-to-end full-link application from the start-end application node m s to the terminal application node m t is detected as a full-link abnormal application, then: firstly initialize the stack, push the node m t into the stack, and the top node m v =m t , and start backtracking from mv to search for all the all-connectable precursor nodes mp of mv ; adopt the method of step (3) to perform abnormal detection on the global operating state of the full-link application to detect whether the global operating state of node mp is abnormal: if mp is an exception, then let mp be pushed into the stack and the top node of the stack is updated to m v =m p and the foregoing operations are iteratively performed, otherwise m v is the exception source application node. 8.如权利要求1所述的一种基于标记图的云平台全链路应用迁移方法，其特征在于：求取异常链路子图和正常链路替换子图，用正常链路替换子图迁移替换异常链路子图，完成全链路应用迁移的方法具体为：8. a kind of cloud platform full-link application migration method based on marked graph as claimed in claim 1, it is characterized in that: seek abnormal link subgraph and normal link replacement subgraph, replace subgraph with normal link The migration replaces the abnormal link subgraph, and the method to complete the full-link application migration is as follows: 求取异常链路子图：求取的异常源应用节点mv的直接后继节点mw为全可通节点，生成从起始端应用节点ms开始经异常源应用节点mv到其直接后继应用节点mw的状态转移图G'；G'即为求取的异常链路子图；Obtain the abnormal link subgraph: The direct successor node m w of the obtained abnormal source application node m v is a fully accessible node, and the generation starts from the starting end application node m s and passes through the abnormal source application node m v to its direct successor application. The state transition graph G' of node m w ; G' is the obtained abnormal link subgraph; 异常链路子图替换：令mw为异常源节点mv的所有直接后继节点，从mw到起始端应用节点ms沿状态转移图递归回溯搜索当前节点的全可通前驱节点以生成正常链路替换子图G"，用正常链路替换子图G"迁移替换异常链路子图，完成全链路应用迁移。Abnormal link subgraph replacement: let m w be all direct successor nodes of abnormal source node m v , from m w to the starting end application node m s recursively backtrack along the state transition graph to search for all accessible predecessor nodes of the current node to generate normal The link replaces the subgraph G", and the normal link replaces the subgraph G" to migrate and replace the abnormal link subgraph to complete the full-link application migration. 9.如权利要求1所述的一种基于标记图的云平台全链路应用迁移方法，其特征在于：生成正常链路替换子图G"的方法具体为：9. a kind of cloud platform full-link application migration method based on marked graph as claimed in claim 1, is characterized in that: the method for generating normal link replacement subgraph G " is specifically: 初始化栈；依次将异常源节点mv的所有直接后继节点设为当前节点，进行如下操作：当前节点m为异常源节点mv的其中一个直接后继节点，递归回溯搜索当前节点m的全可通前驱节点mp，判断是否满足mp∈G'，如果为否，令mp入栈；如果是，不入栈；Initialize the stack; set all direct successor nodes of the abnormal source node mv as the current node in turn, and perform the following operations: the current node m is one of the direct successor nodes of the abnormal source node mv, and recursively backtrack to search for all accessible nodes of the current node m. Predecessor node mp , judge whether mp ∈ G' is satisfied, if it is no, let mp be pushed into the stack; if so, do not push it into the stack; 判断栈顶元素是否为ms，若为ms，利用BFS(广度优先遍历)算法依次访问从起始端应用节点经栈中应用节点到终止端应用节点的链路中的全可通应用，生成从起始端应用节点ms开始经栈中节点到mw的正常链路替换子图G”＝(M”,E”,L”)，其中，M”为用于替换异常链路的正常应用节点集，E”为用于替换异常链路的正常应用接口参数匹配集合，L”为用于替换异常链路的正常应用全局运行状态参数。Determine whether the top element of the stack is m s , if it is m s , use the BFS (breadth-first traversal) algorithm to sequentially access the universally accessible applications in the link from the starting application node through the application node in the stack to the terminating application node, and generate Starting from the application node m s at the starting end, the subgraph G”=(M”, E”, L”) is replaced by the normal link from the node in the stack to m w , where M” is the normal application for replacing the abnormal link Node set, E" is the normal application interface parameter matching set used to replace the abnormal link, L" is the normal application global running state parameter used to replace the abnormal link. 10.一种基于标记图的云平台全链路应用迁移系统，其特征在于，包括：10. A cloud platform full-link application migration system based on a marker graph, characterized in that, comprising: 微应用数据输入模块，对微应用数据进行运行状态和依赖关系的标记图建模构建标记图模型的模块；The micro-application data input module is a module for constructing the label-graph model for the label-graph modeling of the running state and dependencies of the micro-application data; 全链路应用全局运行状态计算模块；The global operating state calculation module of the full-link application; 对全链路应用全局运行状态执行异常检测的模块；A module that performs anomaly detection on the global operating state of the full-link application; 定位异常应用源的模块，求取异常链路子图和正常链路替换子图的模块，将异常链路子图替换为正常链路替换子图的迁移模块；The module that locates the abnormal application source, obtains the module of the abnormal link subgraph and the normal link replacement subgraph, and replaces the abnormal link subgraph with the migration module of the normal link replacement subgraph; 所述微应用数据输入模块、对微应用数据进行运行状态和依赖关系的标记图建模构建标记图模型的模块、全链路应用全局运行状态计算模块、对全链路应用全局运行状态执行异常检测的模块、定位异常应用源的模块，求取异常链路子图和正常链路替换子图的模块，将异常链路子图替换为正常链路替换子图的迁移模块依次连接。The micro-application data input module, the label-graph modeling module for the micro-application data to perform the operation state and dependency relationship and constructing the label-graph model, the full-link application global operation state calculation module, the abnormal execution of the full-link application global operation state The detection module, the module for locating the abnormal application source, the module for obtaining the abnormal link subgraph and the normal link replacement subgraph, and the migration module for replacing the abnormal link subgraph with the normal link replacement subgraph are connected in sequence.
CN201810846524.0A 2018-07-27 2018-07-27 A kind of full link application moving method of cloud platform based on label figure Active CN109117321B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810846524.0A CN109117321B (en) 2018-07-27 2018-07-27 A kind of full link application moving method of cloud platform based on label figure

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810846524.0A CN109117321B (en) 2018-07-27 2018-07-27 A kind of full link application moving method of cloud platform based on label figure

Publications (2)

Publication Number Publication Date
CN109117321A CN109117321A (en) 2019-01-01
CN109117321B true CN109117321B (en) 2019-05-28

Family

ID=64863493

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810846524.0A Active CN109117321B (en) 2018-07-27 2018-07-27 A kind of full link application moving method of cloud platform based on label figure

Country Status (1)

Country Link
CN (1) CN109117321B (en)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109617811B (en) * 2019-01-29 2020-06-30 电子科技大学 A fast migration method for mobile applications in cloud network
CN112383446B (en) * 2020-10-30 2023-03-28 银盛支付服务股份有限公司 Method and system for searching abnormal link in full link
CN115119241B (en) * 2022-06-21 2024-05-14 电子科技大学 A method for detecting link failure in wireless sensor networks
CN117149887A (en) * 2023-11-01 2023-12-01 建信金融科技有限责任公司 Abnormality processing method, abnormality processing device, electronic equipment and computer readable medium

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2419696B (en) * 2004-10-29 2008-07-16 Hewlett Packard Development Co Communication link fault tolerance in a supercomputer
CN102255777B (en) * 2011-08-31 2016-12-21 重庆中天重邮通信技术有限公司 A kind of signaling data acquisition full-link delay correction method
CN106301843A (en) * 2015-05-28 2017-01-04 亿阳信通股份有限公司 A kind of cloud platform safeguards system and method
CN107193967A (en) * 2017-05-25 2017-09-22 南开大学 A kind of multi-source heterogeneous industry field big data handles full link solution

Also Published As

Publication number Publication date
CN109117321A (en) 2019-01-01

Similar Documents

Publication Publication Date Title
CN109117321B (en) A kind of full link application moving method of cloud platform based on label figure
Sheng et al. Identifying influential nodes in complex networks based on global and local structure
CN106970788B (en) A kind of object dependency relationship discovery method and system based on tense
CN109828925B (en) Software reliability measurement method based on software network structure characteristics
CN110808883B (en) A Search Method for Optical Fiber Communication Transmission Path
CN107066256A (en) A kind of object based on tense changes the modeling method of model
CN106027399B (en) A kind of communication network key link recognition
CN108363660B (en) A test program generation method and device
CN115622903B (en) Telecommunication network node importance calculation method based on network structure
CN117114102A (en) Transformer fault diagnosis method based on Bayesian network and fault tree
CN114357703A (en) A system architecture verification evaluation method and device
CN113467421B (en) Methods for obtaining microservice health status indicators and methods for diagnosing microservice abnormalities
Nassif et al. A regression model with mamdani fuzzy inference system for early software effort estimation based on use case diagrams
CN115756929A (en) Abnormal root cause positioning method and system based on dynamic service dependency graph
Huang et al. A case-based knowledge graph with reinforcement learning for intelligent design approach of complex product
CN119597521A (en) Microservice platform service failure recovery method, system and storage medium
Teixeira et al. Graph pattern mining and learning through user-defined relations
CN119440993A (en) Software system dependency defect tracking method and system based on graph database
CN114371998A (en) An Architecture Testing Method Based on Constrained Parallel Breadth-First Search
CN105184373A (en) Bayesian network structure learning method and system and reliability model construction method
CN117291266A (en) Equipment fault reasoning method and system based on extended FMECA analysis method
CN109445843B (en) A software class importance measurement method based on class multi-layer network
CN109637638B (en) A process model building and repairing method with loop-back structure
Zeginis et al. Event pattern discovery for cross-layer adaptation of multi-cloud applications
CN106408155A (en) Reliability evaluating and preconceived fault set searching method based on related circuit set

Legal Events

Date Code Title Description
PB01 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
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20231213

Address after: Room 1501, 1st Floor, Building 10, No. 17 Houyongkang Hutong, Dongcheng District, Beijing, 100010

Patentee after: Pulian Intelligent Data Construction Technology (Beijing) Co.,Ltd.

Address before: 250014 No. 88, Wenhua East Road, Lixia District, Shandong, Ji'nan

Patentee before: SHANDONG NORMAL University