CN103607300A

CN103607300A - Method and system for risk processing within cloud application operation period

Info

Publication number: CN103607300A
Application number: CN201310566811.3A
Authority: CN
Inventors: 许力; 毛军; 马云存
Original assignee: Neusoft Corp
Current assignee: Neusoft Corp
Priority date: 2013-11-12
Filing date: 2013-11-12
Publication date: 2014-02-26
Anticipated expiration: 2033-11-12
Also published as: CN103607300B

Abstract

The invention discloses a method and system for risk processing within a cloud application operation period. The method comprises: first indicator data of a first monitoring indicator related to cloud application operation are obtained in the cloud application operation period; if the first indicator data meet a preset warning condition, all first association resources influencing changing of the first indicator data and/or all second association resources influencing changing of second indicator data are determined in a cloud computing environment, wherein the second indicator data are indicator data of a second monitoring indicator influencing changing of the first indicator data in the cloud computing environment; an adjustment task for first association resource adjustment and/or an adjustment task for second association resource adjustment are/is determined, wherein the adjustment task indicates an adjustment way enabling the first indicator data not to meet the preset warning condition; and one adjustment task is selected from all the adjustment tasks and the association resources corresponding to the selected adjustment task are adjusted by using the selected adjustment task.

Description

A kind of risk processing method and system of cloud application run-time

Technical field

The present invention relates to communication technical field, relate in particular to a kind of risk processing method and system of cloud application run-time.　

Background technology

Cloud data center has built network virtualization layer and the cloud service layer of functions such as having calculating, storage on physical data center, can more intelligent hommization more equipment be managed, transfer more dynamically data resource, and use to pay user with the mode that need get.This method of service of cloud data center significantly promotes the utilization ratio of Liao Yun data center resource and the convenience that resource is used, but meanwhile the complexity of ，Yun data center also increases greatly, and has increased the risk management difficulty of cloud application run-time.　

Traditional cloud data center is to the risk management measures of cloud application run-time mainly: monitoring resource index definition max-thresholds and/or minimum threshold values for being associated with cloud application operation just trigger warning strategies after exceeding predetermined threshold value scope.Illustrate: suppose that moving relevant correlated resources to application APP-A comprises server A, server B, switch C, database D, application server E etc., for ensureing that the operation risk of APP-A within the runtime is found at any time and processed at any time, need and/or minimum threshold values maximum to the monitor control index definition of above-mentioned all correlated resources, and corresponding triggering warning strategies, particularly, suppose that monitoring resource index comprises: the cpu busy percentage of server A, application server E enlivens Thread Count etc., for server A, if the cpu busy percentage of server A be greater than 90% and the duration reach 20 minutes, trigger the not enough alarm of computational resource, if the cpu busy percentage of server A be less than 5% and the duration reach 24 hours, trigger the alarm of computational resource residue.　

After alarm is triggered, traditional cloud data center can pass through mail, the modes such as note or web administration end notify the current risk of O&M personnel correlated resources (such as, server A) and carry risk content indication information (such as, the cpu busy percentage of server A be greater than 90% and the duration reach 20 minutes), then by manually current risk being investigated to processing, and in investigation process, the experience that risk need to rely on O&M personnel is eliminated in How to choose valid function, but, the mode of this artificial investigation risk, make investigation speed slow, investigation difficulty is large, can not remove fast and accurately the operation risk of cloud application run-time.　

Summary of the invention

In view of this, the main purpose of the embodiment of the present invention is to provide a kind of risk processing method and system of cloud application run-time, to realize the object of the operation risk of removing fast and accurately cloud application run-time.　

For achieving the above object, the embodiment of the present invention provides a kind of risk processing method of cloud application run-time, comprising:

In cloud application run-time, obtain the first achievement data of the first supervision index relevant with cloud application operation;

If described the first achievement data meets default alarm conditions, in cloud computing environment, determine each first correlated resources that described the first achievement data of impact changes and/or affect each second correlated resources that the second achievement data changes, described the second achievement data is the second achievement data that monitors index that affects in cloud computing environment that described the first achievement data changes;

Determine the adjustment task that described the first correlated resources is adjusted and/or the adjustment task that described the second correlated resources is adjusted, described adjustment task is to make described the first achievement data not meet the adjustment mode of described default alarm conditions;

From each adjustment task of determining, choose an adjustment task, and utilize choose the adjustment of adjustment task with described in choose correlated resources corresponding to adjustment task.　

Preferably, in said method, described definite adjustment task that described the first correlated resources is adjusted and/or the adjustment task that described the second correlated resources is adjusted, specifically comprise:

Determining makes described the first achievement data not meet the adjustment aim of described default alarm conditions;

When described adjustment aim is while reducing described the first achievement data, obtain the adjustment task for reducing described the first achievement data that described the first correlated resources is adjusted, and/or, obtain the adjustment task for reducing described the first achievement data that described the second correlated resources is adjusted;

When described adjustment aim is while raising described the first achievement data, obtain the adjustment task for described the first achievement data that raises that described the first correlated resources is adjusted, and/or, obtain the adjustment task for described the first achievement data that raises that described the second correlated resources is adjusted.　

Preferably, in said method, the adjustment task that described the first correlated resources is adjusted is the first adjustment task, and the adjustment task that described the second correlated resources is adjusted is the second adjustment task, describedly from each adjustment task of determining, choose an adjustment task, specifically comprise:

Calculate the described first first condition of adjusting task and carry out the second condition execution probability that probability and/or described second is adjusted task, wherein, under the condition that described first condition execution probability is is the first achievement data in described the first supervision index, make described the first achievement data not meet the execution probability of the adjustment task of described default alarm conditions, it is to monitor that described first index is that the first achievement data and described second monitors that index is to make described the first achievement data not meet the execution probability of the adjustment task of described default alarm conditions under the condition of the second achievement data that described second condition is carried out probability,

From the described all conditions that calculates, carry out to choose probability and maximumly carry out probability, and choose described maximum adjustment task corresponding to probability of carrying out, with utilize choose the adjustment of adjustment task with described in choose correlated resources corresponding to adjustment task.　

Preferably, in said method, utilization choose the adjustment of adjustment task with described in choose after correlated resources corresponding to adjustment task, also comprise:

Again obtain the first achievement data of described the first supervision index, again if the first achievement data obtaining described still meets described default alarm conditions, reduce the described maximum probable value of carrying out probability, continue to carry out from the described all conditions calculating and carry out probability and choose the maximum step of carrying out probability, until described the first achievement data does not meet described default alarm conditions.　

Preferably, said method also comprises: build in advance monitor control index related reasoning model;

The method for building up of described monitor control index related reasoning model, specifically comprises:

Determine at least one first supervision index relevant with cloud application operation;

Determine that impact described first monitors each correlated resources that each correlated resources that the data of index change and/or the data that affect the second supervision index change, described second monitors that index is in cloud computing environment, to affect described first to monitor the supervision index that the data of index change;

According to the adjustment task of definite described each correlated resources of adjustment aim of the achievement data of described the first supervision index, described adjustment aim is the achievement data of described the first supervision index of rising or the achievement data that reduces described the first supervision index;

Probability distribution table is set, and described probability distribution table comprises the preset value of effective execution probability of first condition probability, second condition probability and described adjustment task;

Wherein, described effective execution probability is to make the achievement data of the first supervision index not meet the probable value of default alarm conditions after carrying out described adjustment task, described first condition probability is effectively to carry out the probability that probability is preset value described in when described first monitors that index is in the first preset data interval, described second condition probability be when described effective execution probability is preset value described in the probability of the second supervision index in the second preset data interval.　

Preferably, in said method, described calculating described first is adjusted the first condition of task and is carried out the second condition execution probability that probability and/or described second is adjusted task, specifically comprises:

Utilize described the first achievement data to upgrade the current data of the first supervision index in described monitor control index related reasoning model, according to the current data after described renewal, inquire about the probability distribution table in described monitor control index related reasoning model, and the first condition probability inquiring is carried out to probability as the described first first condition of adjusting task;

And/or,

Utilize described the first achievement data to upgrade the current data of the first supervision index in described monitor control index related reasoning model, and utilize described the second achievement data to upgrade the current data of the second supervision index in described monitor control index related reasoning model; According to the current data after described renewal, inquire about the probability distribution table in described monitor control index related reasoning model, and the preset value of the second condition probability obtaining according to inquiry and described effective execution probability calculates the second condition execution probability of described the second adjustment task.　

The embodiment of the present invention also provides a kind of risk treatment system of cloud application run-time, comprising:

Achievement data acquisition module, for obtaining the first achievement data of the first supervision index relevant with cloud application operation in cloud application run-time;

Correlated resources determination module, for when described the first achievement data meets default alarm conditions, in cloud computing environment, determine each first correlated resources that described the first achievement data of impact changes and/or affect each second correlated resources that the second achievement data changes, described the second achievement data is the second achievement data that monitors index that affects in cloud computing environment that described the first achievement data changes;

Adjustment task determination module, for determining the adjustment task that described the first correlated resources is adjusted and/or the adjustment task that described the second correlated resources is adjusted, described adjustment task is to make described the first achievement data not meet the adjustment mode of described default alarm conditions;

Adjustment task is chosen module, for choosing an adjustment task from each adjustment task of determining;

Risk processing module, for utilize choose the adjustment of adjustment task with described in choose correlated resources corresponding to adjustment task.　

Preferably, in said system, described adjustment task determination module, specifically comprises:

Adjustment aim determining unit, makes described the first achievement data not meet the adjustment aim of described default alarm conditions for determining;

First adjusts task determining unit, be used for when described adjustment aim is described the first achievement data of reduction, obtain the adjustment task for reducing described the first achievement data that described the first correlated resources is adjusted, and/or, obtain the adjustment task for reducing described the first achievement data that described the second correlated resources is adjusted;

Second adjusts task determining unit, be used for when described adjustment aim is described the first achievement data of rising, obtain the adjustment task for described the first achievement data that raises that described the first correlated resources is adjusted, and/or, obtain the adjustment task for described the first achievement data that raises that described the second correlated resources is adjusted.　

Preferably, in said system, the adjustment task that described the first correlated resources is adjusted is the first adjustment task, and the adjustment task that described the second correlated resources is adjusted is the second adjustment task, and described adjustment task is chosen module, specifically comprises:

Carry out probability calculation unit, for calculating the described first first condition of adjusting task, carry out the second condition execution probability that probability and/or described second is adjusted task, wherein, under the condition that described first condition execution probability is is the first achievement data in described the first supervision index, make described the first achievement data not meet the execution probability of the adjustment task of described default alarm conditions, it is to monitor that described first index is that the first achievement data and described second monitors that index is to make described the first achievement data not meet the execution probability of the adjustment task of described default alarm conditions under the condition of the second achievement data that described second condition is carried out probability,

Adjustment task is chosen unit, for carry out probability from the described all conditions calculating, choose the maximum probability of carrying out, and choose described maximum adjustment task corresponding to probability of carrying out, with utilize choose the adjustment of adjustment task with described in choose correlated resources corresponding to adjustment task.　

Preferably, in said system,

Described risk processing module, also in utilization, choose the adjustment of adjustment task with described in choose after correlated resources corresponding to adjustment task, again obtain the first achievement data of described the first supervision index, again if the first achievement data obtaining described still meets described default alarm conditions, reduce the described maximum probable value of carrying out probability, continue to carry out from the described all conditions calculating and carry out probability and choose the maximum step of carrying out probability, until described the first achievement data does not meet described default alarm conditions.　

Preferably, said system also comprises: model building module, for building in advance monitor control index related reasoning model;

Described model building module, specifically comprises:

Monitor index determining unit, for determining at least one first supervision index relevant with cloud application operation;

Correlated resources determining unit, for determining that impact described first monitors each correlated resources that each correlated resources that the data of index change and/or the data that affect the second supervision index change, described second monitors that index is in cloud computing environment, to affect described first to monitor the supervision index that the data of index change;

Adjustment task determining unit, be used for according to the adjustment task of definite described each correlated resources of adjustment aim of the achievement data of described the first supervision index, described adjustment aim is the achievement data of described the first supervision index of rising or the achievement data that reduces described the first supervision index;

Probability tables setting unit, for probability distribution table is set, described probability distribution table comprises the preset value of effective execution probability of first condition probability, second condition probability and described adjustment task; Wherein, described effective execution probability is to make the achievement data of the first supervision index not meet the probable value of default alarm conditions after carrying out described adjustment task, described first condition probability is effectively to carry out the probability that probability is preset value described in when described first monitors that index is in the first preset data interval, described second condition probability be when described effective execution probability is preset value described in the probability of the second supervision index in the second preset data interval.　

Preferably, in said system,

Described execution probability calculation unit, specifically for utilizing described the first achievement data to upgrade the current data of the first supervision index in described monitor control index related reasoning model, according to the current data after described renewal, inquire about the probability distribution table in described monitor control index related reasoning model, and the first condition probability inquiring is carried out to probability as the described first first condition of adjusting task; And/or, utilize described the first achievement data to upgrade the current data of the first supervision index in described monitor control index related reasoning model, and utilize described the second achievement data to upgrade the current data of the second supervision index in described monitor control index related reasoning model; According to the current data after described renewal, inquire about the probability distribution table in described monitor control index related reasoning model, and the preset value of the second condition probability obtaining according to inquiry and described effective execution probability calculates the second condition execution probability of described the second adjustment task.　

Risk processing method and the system of the cloud application run-time that the embodiment of the present invention provides, after the satisfied default alarm conditions of the first achievement data of the first supervision index of obtaining, in cloud computing environment, determine each first correlated resources that described the first achievement data of impact changes and/or affect each second correlated resources that the second achievement data changes, described the second achievement data is the second achievement data that monitors index that affects in cloud computing environment that described the first achievement data changes; Then, determine the adjustment task that described the first correlated resources is adjusted and/or the adjustment task that described the second correlated resources is adjusted, finally from each adjustment task of determining, choose an adjustment task, and utilize choose the adjustment of adjustment task with described in choose correlated resources corresponding to adjustment task, to make the first achievement data no longer meet, preset alarm conditions after correlated resources is adjusted.The embodiment of the present invention can be adjusted automatically to meeting the monitor control index of default alarm conditions, thereby automatically realized risk investigation, overcome the defects such as the investigation speed that artificial investigation brings is slow, investigation difficulty is large, realized the object of removing fast and accurately the operation risk of cloud application run-time.　

Accompanying drawing explanation

In order to be illustrated more clearly in the embodiment of the present invention or technical scheme of the prior art, to the accompanying drawing of required use in embodiment or description of the Prior Art be briefly described below, apparently, accompanying drawing in the following describes is some embodiments of the present invention, for those of ordinary skills, do not paying under the prerequisite of creative work, can also obtain according to these accompanying drawings other accompanying drawing.　

Fig. 1 is one of the schematic flow sheet of the risk processing method of embodiment of the present invention cloud application run-time;

Fig. 2 be embodiment of the present invention cloud application run-time risk processing method schematic flow sheet two;

Fig. 3 is the node definition schematic diagram of embodiment of the present invention monitor control index related reasoning model;

Fig. 4 is the first directed acyclic graph of embodiment of the present invention monitor control index related reasoning model;

Fig. 5 is the second directed acyclic graph of embodiment of the present invention monitor control index related reasoning model;

Fig. 6 is the schematic flow sheet of embodiment of the present invention related reasoning method for establishing model;

Fig. 7 is a kind of structural representation of the risk treatment system of embodiment of the present invention cloud application run-time;

Fig. 8 is the another kind of structural representation of the risk treatment system of embodiment of the present invention cloud application run-time.　

Embodiment

For making object, technical scheme and the advantage of the embodiment of the present invention clearer, below in conjunction with the accompanying drawing in the embodiment of the present invention, technical scheme in the embodiment of the present invention is clearly and completely described, obviously, described embodiment is the present invention's part embodiment, rather than whole embodiment.Embodiment based in the present invention, those of ordinary skills, not making the every other embodiment obtaining under creative work prerequisite, belong to the scope of protection of the invention.　

In the cloud computing environment of cloud application operation, the resource relevant to cloud application running status is except physical server, switch, database and intermediate equipment, also have virtual server (VM server), virtual switch etc., and virtual resource also can dynamically increase according to the actual requirements or reduce, mapping relations between virtual resource and physical resource also can change dynamically, etc.Based on this, when breaking down, the virtual unit in physical equipment or cloud computing environment or cloud application itself etc. all will affect the normal operation of cloud application, so, risk processing method and the device of the cloud application run-time that the invention process provides, mainly when abnormal operating condition appears in cloud application run-time, in time the related resource in cloud computing environment is adjusted, to realize the object of automatic processing operation risk.For realizing this object, with regard to each embodiment of the present invention, be specifically introduced below.　

Embodiment mono-

Referring to Fig. 1, the schematic flow sheet of the risk processing method of the cloud application run-time providing for the embodiment of the present invention one, specifically comprises:

Step 101: the first achievement data that obtains the first supervision index relevant with cloud application operation in cloud application run-time.　

Wherein, described first monitor that index can be service request response time or user's monitor control index that the O&M personnel such as number or service request average handling time are concerned about online.In addition, can preset a plurality of the first monitor control indexs, and in cloud application run-time, the achievement data of these a plurality of the first supervision indexs be gathered.　

Step 102: if described the first achievement data meets default alarm conditions, in cloud computing environment, determine each first correlated resources that described the first achievement data of impact changes and/or affect each second correlated resources that the second achievement data changes, described the second achievement data is the second achievement data that monitors index that affects in cloud computing environment that described the first achievement data changes.　

In embodiments of the present invention, the resource in cloud computing environment is divided into two class resources, a class is the first correlated resources, and another kind of is the second correlated resources.The variation of described the first correlated resources will directly cause described the first achievement data to change; The variation of described the second correlated resources will directly cause described the second achievement data to change and the variation of described the second achievement data directly causes again described the first achievement data to change, i.e. the variation of described the second correlated resources causes described the first achievement data to change indirectly.So, in actual applications, affect the correlated resources that the first achievement data changes and can only comprise the first correlated resources, or only comprise the second correlated resources, or comprise the first correlated resources and the second correlated resources simultaneously.　

Wherein, described the first correlated resources or described the second correlated resources can be: physical host, fictitious host computer, network, operating system, etc.; Described second monitors that index can be: VM cpu busy percentage or disk I/O (input and output) flow, or VM memory usage, or network bandwidth utilization rate, or network I/O delay, etc.　

Step 103: determine the adjustment task that described the first correlated resources is adjusted and/or the adjustment task that described the second correlated resources is adjusted, described adjustment task is to make described the first achievement data not meet the adjustment mode of described default alarm conditions.　

In step 103, suppose that " fictitious host computer " is the first correlated resources or the second correlated resources of having determined, the adjustment task of fictitious host computer being adjusted can be: migration virtual machine or lifting virtual machine CPU quota, or restart virtual machine etc.　

Step 104: from each adjustment task of determining, choose an adjustment task, and utilize choose the adjustment of adjustment task with described in choose correlated resources corresponding to adjustment task.　

In order to understand more easily the embodiment of the present invention one, illustrate below:

Suppose that current the first supervision index getting is: the service request response time, the first corresponding achievement data is: 10 seconds, default alarm conditions were: the service request response time is greater than 5 seconds and can triggers alarm.Because current the first achievement data obtaining has met default alarm conditions, so triggering alarm, after triggering alarm, when only exist make that the service request response time changes one or more second while monitoring index, if the second supervision index of determining is VM cpu busy percentage (can also comprise other the second supervision index), further determine that making the second correlated resources that VM cpu busy percentage changes is fictitious host computer, in order to make the service request response time be less than 5 seconds, just need fictitious host computer to adjust (for example: promote virtual machine CPU quota), the object of adjusting is that VM cpu busy percentage is changed, final purpose is after VM cpu busy percentage changes, make the service request response time be less than 5 seconds.　

Embodiment bis-

Referring to Fig. 2, the schematic flow sheet of the risk processing method of the cloud application run-time providing for the embodiment of the present invention two, specifically comprises:

Step 201: the first achievement data that obtains the first supervision index relevant with cloud application operation in cloud application run-time.　

Step 202: if described the first achievement data meets default alarm conditions, in cloud computing environment, determine each first correlated resources that described the first achievement data of impact changes and/or affect each second correlated resources that the second achievement data changes, described the second achievement data is the second achievement data that monitors index that affects in cloud computing environment that described the first achievement data changes.　

Step 203: determine and make described the first achievement data not meet the adjustment aim of described default alarm conditions.　

Step 204: when described adjustment aim is while reducing described the first achievement data, obtain the adjustment task for reducing described the first achievement data that described the first correlated resources is adjusted, and/or, obtain the adjustment task for reducing described the first achievement data that described the second correlated resources is adjusted, perform step 206.　

Step 205: when described adjustment aim is while raising described the first achievement data, obtain the adjustment task for described the first achievement data that raises that described the first correlated resources is adjusted, and/or, obtain the adjustment task for described the first achievement data that raises that described the second correlated resources is adjusted.　

If the definite correlated resources of step 202 is only the first correlated resources, in step 204 or step 205, determine the adjustment task to the first correlated resources; If the definite correlated resources of step 202 is only the second correlated resources, in step 204 or step 205, determine the adjustment task to the second correlated resources; If the definite correlated resources of step 202 not only comprises the first correlated resources but also comprise the second correlated resources, in step 204 or step 205, determine respectively the adjustment task to the first correlated resources and the second correlated resources.　

Step 206: the first condition that calculates the first adjustment task is carried out the second condition execution probability that probability and/or second is adjusted task, wherein, the adjustment task that described the first correlated resources is adjusted is the first adjustment task, and the adjustment task that described the second correlated resources is adjusted is the second adjustment task.　

Wherein, it is to monitor that described first index is under the condition of the first achievement data, to make described the first achievement data not meet the execution probability of the adjustment task of described default alarm conditions that described first condition is carried out probability, and it is to monitor that described first index is that the first achievement data and described second monitors that index is to make described the first achievement data not meet the execution probability of the adjustment task of described default alarm conditions under the condition of the second achievement data that described second condition is carried out probability.　

Step 207: from the described all conditions that calculates, carry out to choose probability and maximumly carry out probability, and choose described maximum adjustment task corresponding to probability of carrying out, with utilize choose the adjustment of adjustment task with described in choose correlated resources corresponding to adjustment task.　

Wherein, adjustment task corresponding to described maximum execution probability may be the adjustment task that described the first correlated resources is adjusted, and may be also the adjustment task that described the second correlated resources is adjusted.　

Step 208: after correlated resources is adjusted, again obtain the first achievement data of described the first supervision index.　

Step 209: whether the first achievement data again obtaining described in judgement meets described default alarm conditions, if so, performs step 210, if not, performs step 211;

Step 210: reduce the described maximum probable value of carrying out probability, continue execution step 707.　

Step 211: process ends.　

Step 207 to 210 in, after utilizing adjustment task corresponding to maximum execution probability to adjust correlated resources, the adjustment effect that may not occur expection, the first achievement data that is described the first supervision index still meets described default alarm conditions, now the maximum that calculates is carried out probability reduce certain probable value (such as: maximum when carrying out probability and being 80%, 80% minimizing 20% is obtained to 60%, 60% maximum that not all condition is carried out in probability is carried out probability), then, from all condition execution probability, obtain again the maximum probability of carrying out of another one, continue to utilize this maximum to carry out adjustment task corresponding to probability and carry out the adjustment of corresponding correlated resources, so circulation, until the first achievement data does not meet described default alarm conditions.　

For the first condition in calculation procedure 206 is carried out probability and second condition execution probability, further comprising the steps of before carrying out the embodiment of the present invention: to build in advance monitor control index related reasoning model.Lower mask body place of matchmakers states the method for building up of monitor control index related reasoning model, is mainly divided into following three steps:

The first step: each node in monitor control index related reasoning model is defined

1, first monitor index node

The operation system index definition of the cloud application run-time that O&M personnel are paid close attention to is the first supervision index node, such as: service request response time, the online number of user etc.In addition, each first supervision index all has its corresponding risk warning strategies, utilizes the first supervision index currency to judge whether to meet default alarm conditions, if so, triggers alarm and automatically processes to carry out follow-up risk.　

After having defined each the first supervision index node, further the codomain of each node is defined, for certain the first supervision index node, suppose that its node codomain is { 1,2,3 }, wherein, a preset data interval (achievement data the is interval) mapping mutually of each thresholding in node codomain and this first supervision index.For example: the node codomain of supposing the service request response time is { 1,2,3 }, the thresholding 1 in definable node codomain is corresponding with the interval 0-5s of preset data, thresholding 2 in defined node codomain is corresponding with the interval 5-10s of preset data, and the thresholding 3 in defined node codomain is corresponding with the interval 10-20s of default value.It should be noted that, the definition of node codomain is not limited to above-mentioned form, can also the thresholding in node codomain be carried out suitable minimizing or be increased etc.　

2, second monitor index node

By the supervision index definition of cloud application run-time correlated resources, be the second supervision index node, such as: VM cpu busy percentage, disk I/O flow, VM memory usage, network bandwidth utilization rate, network I/O postpones, etc.　

After having defined each the second supervision index node, further the codomain of each node is defined, for certain the second supervision index node, suppose that its node codomain is { 1,2,3 }, wherein, a preset data interval (achievement data the is interval) mapping mutually of each thresholding in node codomain and this second supervision index.For example: the node codomain of supposing VM cpu busy percentage is { 1,2,3 }, the thresholding 1 in definable node codomain is corresponding with the interval 0%-20% of preset data, thresholding 2 in defined node codomain is corresponding with the interval 20%-50% of preset data, and the thresholding 3 in defined node codomain is corresponding with the interval 50%-100% of default value.It should be noted that, the definition of node codomain is not limited to above-mentioned form, can also the thresholding in node codomain be carried out suitable minimizing or be increased etc.　

3, adjust task node

To be defined as adjustment task node based on the predefined cloud environment control task of expertise (the adjustment task to the first correlated resources or the second correlated resources), such as: increase clustered node, promote virtual machine density, etc.Particularly, according to the data adjustment aim of the first supervision index (reducing or the first supervision index that raises), the adjustment task corresponding with adjustment aim is set; In addition, the data adjustment aim of each the first supervision index, can be to there being one or more adjustment tasks.　

After having defined each adjustment task node, further the codomain of each node is defined, for certain, adjust task node, suppose that its node codomain is for { T, F }, after adjustment task is carried out in T representative, effectively (after carrying out adjustment task, removed current alarm, such as making the data of current the first supervision index of obtaining no longer meet default alarm conditions), invalid after F representative execution adjustment task (after carrying out adjustment task, not removing yet current alarm, such as making the data of current the first supervision index of obtaining still meet default alarm conditions).　

Second step: the directed acyclic graph (DAG) of drawing monitor control index related reasoning model

After node definition completes (referring to the node definition schematic diagram of monitor control index related reasoning model as shown in Figure 3), according to each internodal correlation, the node with correlation is connected into a directed acyclic graph (DAG) by associated order with directed edge, wherein, out-degree is that the node of O is the first supervision index node, in-degree be 0 node for adjusting task node, it is the second supervision index node that existing out-degree has again the node of in-degree.　

For example: first monitors that index node is exemplified as " service request response time ", actual realization is also replaceable is the operation system index node that any O&M personnel are concerned about, such as: " online user number ", " service request average handling time ", etc., these nodes can certainly be defined as respectively to the first different supervision indexs; Second monitors that index node is exemplified as " VM cpu busy percentage ", actual realization can increase or replace the system resource index nodes such as any software relevant to the first monitor control index node, equipment, such as: " disk I/O flow ", " VM memory usage ", " network bandwidth utilization rate ", " network I/O delay ", etc.; Adjust task node and be exemplified as " increase clustered node " and " promoting virtual machine density ", actual realize changeable or replace anyly can affect the adjustment task node that the second supervision achievement data changes, such as: " optimized network topology ", " minimizing clustered node ", " promote internal memory quota ", " promote network bandwidth quota ", etc.These nodes that are mutually related are connected with directed edge, with this, reflect internodal relevance.　

Illustrate: the first directed acyclic graph of monitor control index related reasoning model shown in Figure 4, in Fig. 4, definition " increase clustered node " and " promoting virtual machine density " are the adjustment tasks to the second correlated resources, this adjustment task will affect the data variation of the second supervision index " VM memory usage ", and second monitors that index " VM memory usage " will affect the data variation of the first supervision index " service request response time ".In addition, also there is the directed edge connected mode shown in Fig. 5 in the embodiment of the present invention, the second directed acyclic graph of monitor control index related reasoning model shown in Figure 5, in Fig. 5, definition " increase clustered node " and " promoting virtual machine density " are the adjustment tasks to the first correlated resources, and this adjustment task will directly affect the first supervision index data variation of " service request response time ".Certainly, also there is the simultaneous directed edge connected mode of connected mode shown in Fig. 4 and Fig. 5 in the embodiment of the present invention.　

The 3rd step: conditional probability distribution table is set

After directed acyclic graph (DAG) has defined, in figure, each node all needs, according to expertise data, a conditional probability distribution table is set.Mainly be divided into following two kinds of situations:

Situation one: referring to Fig. 4, based on there being following relation between the first supervision index A, the second supervision index B and the task of adjustment C: C-> B-> A, carrying out C will cause the achievement data of B to change, the achievement data of B changes causing the achievement data of A to change, and establishes P(A|B) be that A is the probability of the first designated value (thresholding in the node codomain of the first supervision index) when B is the second designated value (thresholding in the node codomain of the second supervision index); If P(B|C) for B when C is the 3rd designated value (effectively carrying out probability or invalid execution probability) be the probability of the second designated value (thresholding in the node codomain of the second supervision index); P(C=T) represent effectively to carry out the preset value of probability, P(C=F) represent the preset value of invalid execution probability.Wherein, described effective execution probability makes the achievement data of the first supervision index not meet the probable value of default alarm conditions after referring to and carrying out adjustment task, and described invalid execution probability makes the achievement data of the first supervision index still meet the probable value of default alarm conditions after referring to and carrying out adjustment task.　

For example: referring to table 1 to table 3, the conditional probability distribution table that table 1 is A, the conditional probability distribution table that table 2 is B, the predetermined probabilities distribution table that table 3 is C.　

Ｂ	Ｐ（Ａ＝1｜Ｂ）	Ｐ（Ａ＝2｜Ｂ）	Ｐ（Ａ＝3｜Ｂ）
				1	0．5	0．3	0．2
2	0．1	0．6	0．3
				3	0．1	0．2	0．7

Table 1 A node conditional probability table

Ｃ	Ｐ（Ｂ＝1｜Ｃ）	Ｐ（Ｂ＝2｜Ｃ）	Ｐ（Ｂ＝3｜Ｃ）
				Ｔ	0．3	0．6	0．1
Ｆ	0．1	0．2	0．7

Table 2 B node conditional probability table

Ｐ（Ｃ＝Ｔ）	Ｐ（Ｃ＝Ｆ）
		0．6	0．4

Table 3 C node predetermined probabilities table

Situation two: referring to Fig. 5, based on there being following relation between the first supervision index A, adjustment task C: C-> A, carry out C by causing the achievement data of A to change, establish P(C=T|A) be that C is the probability of the 3rd designated value (effectively carrying out probability or invalid execution probability) when A is the first designated value (thresholding in the node codomain of the first supervision index).　

For example: referring to table 4 to table 5, the conditional probability distribution table that table 4 is C, the predetermined probabilities distribution table that table 5 is C.　

Ａ	Ｐ（Ｃ＝Ｔ｜Ａ）	Ｐ（Ｃ＝Ｆ｜Ａ）
			1	0．8	0．2
2	0．6	0．4
			3	0．5	0．5

Table 4 C node conditional probability table

Ｐ（Ｃ＝Ｔ）	Ｐ（Ｃ＝Ｆ）
		0．6	0．4

Table 5 C node predetermined probabilities table

Model establishment step based on above-mentioned introduction, the schematic flow sheet of related reasoning method for establishing model shown in Figure 6, the method for building up of described monitor control index related reasoning model, specifically comprises:

Step 601: determine at least one first supervision index relevant with cloud application operation.　

Step 602: determine that impact described first monitors each correlated resources that each correlated resources that the data of index change and/or the data that affect the second supervision index change, described second monitors that index is to affect described first in cloud computing environment to monitor the supervision index that the data of index change.　

Step 603: according to the adjustment task of definite described each correlated resources of adjustment aim of the achievement data of described the first supervision index, described adjustment aim is the achievement data of described the first supervision index of rising or the achievement data that reduces described the first supervision index.　

Step 604: probability distribution table is set, and described probability distribution table comprises the preset value P(C=T of effective execution probability of first condition probability P (C=T|A), second condition probability P (B|C=T) and described adjustment task).　

Wherein, described effective execution probability P (C=T) is to make the achievement data of the first supervision index not meet the probable value of default alarm conditions after carrying out described adjustment task, described first condition probability P (C=T|A) is effectively to carry out the probability that probability is preset value described in when described first monitors that index is in the first preset data interval, described second condition probability P (B|C=T) be when described effective execution probability is preset value described in the probability of the second supervision index in the second preset data interval.　

Based on the above-mentioned monitor control index related reasoning model of setting up in advance, the first condition in step 206 carries out probability and second condition is carried out probability calculating in the following manner respectively:

1, described first condition is carried out the computing formula of probability and is: p (C=T|A=y) (1) after having obtained effective execution probability C=T of the first achievement data A=y and the task of adjustment C, just can question blank 4 in required first condition probability as first condition, carry out probability.　

2, the computing formula of described second condition execution probability is as follows:

P (C = T | B = x, A = y) = \frac{P (C = T, B = x, A = y)}{P (B = x, A = y)} - - - (2)

Wherein, p (B=x, A=y)=P (A=y|B=x) P (B=x) (3)

p(C＝T,B＝x,A＝y)＝P(A＝y|B＝x)P(B＝x)P(B＝x|C＝T)P(C＝T) 　　　　　　（4）　

Bring formula (3) and (4) into formula (2), obtain:

p(C＝T|B＝x,A＝y)＝P(B＝x|C＝T)P(C＝T) 　　　　　　　　　　　　　（5）　

Obtaining the first achievement data A=y and the second achievement data B=x, and after the task of adjustment C, because a thresholding in the second achievement data B=x and node codomain (second monitors the node codomain of index node) is shone upon mutually, inquiry probability distribution table, obtain the default execution probability P (C=T) of second condition probability P (B=x|C=T) and corresponding adjustment task, according to above-mentioned formula (5), calculate second condition and carry out probability P (B=x|C=T) P (C=T).　

Known based on above-mentioned formula, " calculate the described first first condition of adjusting task and carry out the second condition execution probability that probability and/or described second is adjusted task " in step 206, can realize in the following manner:

And/or,

The risk processing method of the cloud application run-time that the embodiment of the present invention provides, after the satisfied default alarm conditions of the first achievement data of the first supervision index of obtaining, in cloud computing environment, determine each first correlated resources that described the first achievement data of impact changes and/or affect each second correlated resources that the second achievement data changes, described the second achievement data is the second achievement data that monitors index that affects in cloud computing environment that described the first achievement data changes; Then, determine the adjustment task that described the first correlated resources is adjusted and/or the adjustment task that described the second correlated resources is adjusted, finally from each adjustment task of determining, choose an adjustment task, and utilize choose the adjustment of adjustment task with described in choose correlated resources corresponding to adjustment task, to make the first achievement data no longer meet, preset alarm conditions after correlated resources is adjusted.The embodiment of the present invention can be adjusted automatically to meeting the monitor control index of default alarm conditions, thereby automatically realized risk investigation, overcome the defects such as the investigation speed that artificial investigation brings is slow, investigation difficulty is large, realized the object of removing fast and accurately the operation risk of cloud application run-time.　

Embodiment tri-

Referring to Fig. 7, the structural representation of the risk treatment system of the cloud application run-time providing for the embodiment of the present invention three, is characterized in that, comprising:

Achievement data acquisition module 1, for obtaining the first achievement data of the first supervision index relevant with cloud application operation in cloud application run-time;

Correlated resources determination module 2, for when described the first achievement data meets default alarm conditions, in cloud computing environment, determine each first correlated resources that described the first achievement data of impact changes and/or affect each second correlated resources that the second achievement data changes, described the second achievement data is the second achievement data that monitors index that affects in cloud computing environment that described the first achievement data changes;

Adjustment task determination module 3, for determining the adjustment task that described the first correlated resources is adjusted and/or the adjustment task that described the second correlated resources is adjusted, described adjustment task is to make described the first achievement data not meet the adjustment mode of described default alarm conditions;

Adjustment task is chosen module 4, for choosing an adjustment task from each adjustment task of determining;

Risk processing module 5, for utilize choose the adjustment of adjustment task with described in choose correlated resources corresponding to adjustment task.　

Wherein, described adjustment task determination module 3, specifically comprises:

Wherein, described adjustment task is chosen module 4, specifically comprises:

Carry out probability calculation unit, for calculating the described first first condition of adjusting task, carry out the second condition execution probability that probability and/or described second is adjusted task, wherein, the adjustment task that described the first correlated resources is adjusted is the first adjustment task, the adjustment task that described the second correlated resources is adjusted is the second adjustment task, under the condition that described first condition execution probability is is the first achievement data in described the first supervision index, make described the first achievement data not meet the execution probability of the adjustment task of described default alarm conditions, it is to monitor that described first index is that the first achievement data and described second monitors that index is to make described the first achievement data not meet the execution probability of the adjustment task of described default alarm conditions under the condition of the second achievement data that described second condition is carried out probability,

Wherein, described risk processing module, also in utilization, choose the adjustment of adjustment task with described in choose after correlated resources corresponding to adjustment task, again obtain the first achievement data of described the first supervision index, again if the first achievement data obtaining described still meets described default alarm conditions, reduce the described maximum probable value of carrying out probability, continue to carry out from the described all conditions calculating and carry out probability and choose the maximum step of carrying out probability, until described the first achievement data does not meet described default alarm conditions.　

In addition, described system also comprises: model building module 5, for building in advance monitor control index related reasoning model; Described model building module 5, specifically comprises:

Probability tables setting unit, for probability distribution table is set, described probability distribution table comprises the preset value of effective execution probability of first condition probability, second condition probability and described adjustment task;

Wherein, described execution probability calculation unit, specifically for described execution probability calculation unit, specifically for utilizing described the first achievement data to upgrade the current data of the first supervision index in described monitor control index related reasoning model, according to the current data after described renewal, inquire about the probability distribution table in described monitor control index related reasoning model, and the first condition probability inquiring is carried out to probability as the described first first condition of adjusting task; And/or, utilize described the first achievement data to upgrade the current data of the first supervision index in described monitor control index related reasoning model, and utilize described the second achievement data to upgrade the current data of the second supervision index in described monitor control index related reasoning model; According to the current data after described renewal, inquire about the probability distribution table in described monitor control index related reasoning model, and the preset value of the second condition probability obtaining according to inquiry and described effective execution probability calculates the second condition execution probability of described the second adjustment task.　

The risk treatment system of the cloud application run-time that the embodiment of the present invention provides, after the satisfied default alarm conditions of the first achievement data of the first supervision index of obtaining, in cloud computing environment, determine each first correlated resources that described the first achievement data of impact changes and/or affect each second correlated resources that the second achievement data changes, described the second achievement data is the second achievement data that monitors index that affects in cloud computing environment that described the first achievement data changes; Then, determine the adjustment task that described the first correlated resources is adjusted and/or the adjustment task that described the second correlated resources is adjusted, finally from each adjustment task of determining, choose an adjustment task, and utilize choose the adjustment of adjustment task with described in choose correlated resources corresponding to adjustment task, to make the first achievement data no longer meet, preset alarm conditions after correlated resources is adjusted.The embodiment of the present invention can be adjusted automatically to meeting the monitor control index of default alarm conditions, thereby automatically realized risk investigation, overcome the defects such as the investigation speed that artificial investigation brings is slow, investigation difficulty is large, realized the object of removing fast and accurately the operation risk of cloud application run-time.　

The system configuration schematic diagram providing based on above-mentioned Fig. 7, the embodiment of the present invention also provides another system configuration schematic diagram, the structural representation of the risk treatment system of cloud application run-time shown in Figure 8.　

For the effectively application of supervision cloud and cloud environment running status, automatically find risk and process in time risk, by this system model, divide application layer risk analysis subsystem and environment layer ADMINISTRATION SUBSYSTEM, two subsystems communicate by message mechanism.　

Wherein, application layer risk analysis subsystem comprises following equipment:

1, application monitor: be deployed in cloud application.　

Cloud application can provide one or more services, in cloud application run-time, regularly gather the achievement data that first of each service monitors index, and the achievement data of each the first supervision index is kept in cloud working knowledge storehouse, for the inquiry of application risk analyzer, use.　

2, working knowledge storehouse: be deployed in cloud application.　

Be responsible for the achievement data of each the first supervision index of regularly collection of storage application monitor.　

3, application risk analyzer: be deployed in cloud application.　　

Application risk analyzer comprises the following module in embodiment tri-and can realize the function of these modules: achievement data acquisition module 1, correlated resources determination module 2, adjustment task determination module 3, adjustment task are chosen module 4.Wherein, achievement data acquisition module 1 is for obtaining the first achievement data from working knowledge storehouse, correlated resources determination module 2 can be communicated by letter with task processor, and the second achievement data sending to receive task processor, to utilize the second achievement data obtaining to determine the second correlated resources.In addition, application risk analyzer is also responsible for safeguarding the monitor control index related reasoning model of cloud application run-time.　

Wherein, environment layer ADMINISTRATION SUBSYSTEM comprises following equipment:

1, environmental monitor: be deployed in cloud computing environment.　

Environmental monitor can be a physical server or virtual server, for regularly gathering second of each resource of cloud computing environment (physical host, fictitious host computer, network, operating system etc.), monitor index, and the achievement data of the second supervision index is kept in cloud environment knowledge base, for task processor, inquire about.　

2, cloud environment knowledge base: be deployed in cloud application.　

Be responsible for the achievement data of each the second supervision index of regularly collection of storage environment monitor.　

2, task processor: be deployed in cloud computing environment.　

From cloud environment knowledge base, obtain each second achievement data relevant to current abnormal the first achievement data getting; Communicate by letter with application risk analyzer, and send to risk analyzer each second achievement data obtaining; The adjustment task that receives the transmission of application risk analyzer is chosen the adjustment task that module 4 is chosen.　

3, environmental control: be deployed in cloud computing environment.　

Receive the environment control command (as: migration virtual machine, promote virtual machine CPU quota, restart virtual machine etc.) of the task of adjusting for realizing correlated resources that task processor sends, can realize embodiment tri-risk processing modules 5 function.　

As seen through the above description of the embodiments, those skilled in the art can be well understood to the mode that all or part of step in above-described embodiment method can add essential general hardware platform by software and realizes.Understanding based on such, the part that technical scheme of the present invention contributes to prior art in essence in other words can embody with the form of software product, this computer software product can be stored in storage medium, as ROM/RAM, magnetic disc, CD etc., comprise that some instructions are with so that a computer equipment (can be personal computer, server, or such as network communication equipments such as media gateway, etc.) method described in some part of each embodiment of the present invention or embodiment carried out.　

It should be noted that, in this specification, each embodiment adopts the mode of going forward one by one to describe, and each embodiment stresses is the difference with other embodiment, between each embodiment identical similar part mutually referring to.For the disclosed system of embodiment, because it corresponds to the method disclosed in Example, so description is fairly simple, relevant part partly illustrates referring to method.　

Also it should be noted that, in this article, relational terms such as the first and second grades is only used for an entity or operation to separate with another entity or operating space, and not necessarily requires or imply and between these entities or operation, have the relation of any this reality or sequentially.And, term " comprises ", " comprising " or its any other variant are intended to contain comprising of nonexcludability, thereby the process, method, article or the equipment that make to comprise a series of key elements not only comprise those key elements, but also comprise other key elements of clearly not listing, or be also included as the intrinsic key element of this process, method, article or equipment.The in the situation that of more restrictions not, the key element being limited by statement " comprising ... ", and be not precluded within process, method, article or the equipment that comprises described key element and also have other identical element.　

Above-mentioned explanation to the disclosed embodiments, makes professional and technical personnel in the field can realize or use the present invention.To the multiple modification of these embodiment, will be apparent for those skilled in the art, General Principle as defined herein can, in the situation that not departing from the spirit or scope of the present invention, realize in other embodiments.Therefore, the present invention will can not be restricted to these embodiment shown in this article, but will meet the widest scope consistent with principle disclosed herein and features of novelty.　

Claims

1. a risk processing method for cloud application run-time, is characterized in that, comprising:

2. method according to claim 1, is characterized in that, described definite adjustment task that described the first correlated resources is adjusted and/or the adjustment task that described the second correlated resources is adjusted, specifically comprise:

3. method according to claim 1 and 2, it is characterized in that, the adjustment task that described the first correlated resources is adjusted is the first adjustment task, the adjustment task that described the second correlated resources is adjusted is the second adjustment task, describedly from each adjustment task of determining, choose an adjustment task, specifically comprise:

4. method according to claim 3, is characterized in that, utilization choose the adjustment of adjustment task with described in choose after correlated resources corresponding to adjustment task, also comprise:

5. method according to claim 3, is characterized in that, described method also comprises: build in advance monitor control index related reasoning model;

6. method according to claim 5, is characterized in that, described calculating described first is adjusted the first condition of task and carried out the second condition execution probability that probability and/or described second is adjusted task, specifically comprises:

And/or,

7. a risk treatment system for cloud application run-time, is characterized in that, comprising:

8. system according to claim 7, is characterized in that, described adjustment task determination module, specifically comprises:

9. according to the system described in claim 7 or 8, it is characterized in that, the adjustment task that described the first correlated resources is adjusted is the first adjustment task, and the adjustment task that described the second correlated resources is adjusted is the second adjustment task, described adjustment task is chosen module, specifically comprises:

10. system according to claim 9, is characterized in that,

11. systems according to claim 9, is characterized in that, described system also comprises: model building module, for building in advance monitor control index related reasoning model;

Described model building module, specifically comprises:

12. systems according to claim 11, is characterized in that,