CN103473100B - A kind of Web service replacement method based on dependence - Google Patents

Abstract

The present invention relates to a kind of Web service replacement method based on dependence, belong to computer software technical field, method includes: when a certain Web service can not meet the change of external environment condition or user's request, identifies deployment scenario, version information and the direct precursor service in composite services thereof and the immediate successor service of problem service；Describe, in conjunction with from problem service composite, the dependence extracted in file, select suitably service according to different situations and be replaced；When problem service other dispose and other versions the most not in the presence of, the immediate successor set of the service in similar services set comprises all follow-up service of problem service, calculate the service interaction relation set up between they and the problem service similarity in input and output, constraints and the service of specific area inner question and similar services, select the service mated most to be replaced.

Description

A kind of Web service replacement method based on dependence

Technical field

The present invention relates to the replacement method of a kind of Web service, particularly relate in a kind of large-scale complex software systems based on dependence The Web service replacement method of relation, belongs to computer software technical field.

Background technology

Web service be by ISP issue under Networks Environment autonomous, stateless, there is standalone feature Computational entity.Service-oriented computing utilizes Web service as essential structure unit, supports the one-tenth quick, low of Distributed Application This combined type exploitation.Based on Service-Oriented Architecture Based (Service-Oriented Architecture, SOA), software i.e. services (SaaS) With cloud computing, enterprise bought software and safeguard infrastructure to be changed into by internet to lease in a large number and consumption service service.So And, along with service-oriented software systems scale is increasing, its problem such as availability, reliability is increasingly difficult to ensure that.When When Web service can not tackle the change of external environment condition or user's request, in order to ensure the correct execution of composite services, need immediately Replace this Web service.

Through the retrieval of prior art is found, Chinese Patent Application No. 201110058676.2, describe a kind of excellent based on service The Web service replacement method of first level.The service choosing importance in architecture strong is first replaced, and prioritizing selection is waited Select the service that service quality in services set is high, by the way of replacing part Web service, repair the composite service of inefficacy.This Method does not accounts for the restriction relation between Web service, directly replaces and may produce the Web service with its Constrained relation Raw impact.

In service oriented computing environment, call mutually between the service of distributed, different grain size, different field, all deposit In complementary contact.Chinese Patent Application No. 201210159395.0, describes and associates between service under a kind of open network The method for building up of relation, according to the execution route of composite services, anabolic process and service implementation process, analyzes composite services not With the incidence relation in step, set up substep relationship maps model, build association service resource pool.Chinese Patent Application No. 201110168027.8, describe a kind of web service composition method based on dependence, it is proposed that describe Web service The model of data dependence information, during Service Matching, utilizes the dependence of Web service that Web service combination is converted Become dynamic programming problems, and find out the service combining method of optimum.Chinese Patent Application No. 201010289364.8, describes one Planting Web application and dispose constraint automatic testing method, the deployment to Web application deployment scheme model retrains automatic examinations, After detecting that deployment constraint is run counter to by current Web application, it is automatically positioned and point out user to carry out the amendment of deployment scheme.

In sum, for dependence, the dependence between Web service is applied to by current existing relevant programme mostly During Services Composition, and Web service replacement method research approach based on dependence is less.

Summary of the invention

It is an object of the invention to: overcome the deficiencies in the prior art, it is provided that a kind of Web service replacement side based on dependence Method, it is ensured that the correct execution of composite services under cloud computing environment.

The technology of the present invention solution includes: a kind of Web service replacement method based on dependence, it is achieved step is:

Step 1, when a certain Web service can not meet the change of external environment condition or user's request in a certain position, first identifies The deployment scenario of described a certain Web service, version information and the forerunner's set of service in composite services thereof that go wrong and Follow-up set of service；

Step 2, extracts input/output, constraints that the described a certain Web service that goes wrong describes in file and calls Relation, the input and output of problem Web service described in following steps, constraints and follow-up set of service are all to be walked by this Suddenly obtain；

Step 3, disposes if the described a certain Web service gone wrong exists in other positions, checks and be deployed in other positions Whether Web service meets external environment condition or user's request, meets, then by this problem Web service if there is a Web service Replace with the described Web service meeting external environment condition or user's needs, jump to step 8, otherwise perform step 4；

Step 4, if there is other versions in the described a certain Web service gone wrong, using above-mentioned every other version as similar Set, compares each Web service in similar set with the described a certain Web service gone wrong, i.e. performs step Rapid 6；If the described a certain Web service gone wrong does not exists other versions, then perform step 5；

Step 5, if other of this described a certain Web service gone wrong dispose and other versions do not exist, sentences the most one by one Whether the immediate successor set of service of other Web services disconnected comprises all follow-up service of problem Web service, will comprise problem All other the follow-up Web services of Web service put similar set under；

Step 6, calculates each similar Web service and the similarity of problem Web service in similar set, obtains input and output Matrix M_{in_out}, constraint matrix M_constraint, service interaction matrix M_correlation；

Step 7, according to the input and output matrix M in step 6_{in_out}, constraint matrix M_constraint, service interaction matrix M_correlationCalculate dependence matrix M, Web service is ranked up, it is recommended that go out to replace go wrong described a certain Web service；

Step 8, intercepts the use request that user sends to problem service, replaces this problem Web service.

Direct precursor service described in step 1 and immediate successor service be: assumes that having dependence is < s₁,s₂>、<s₂,s₃>, table Show Web service s₁Depend on Web service s₂, Web service s₂Depend on Web service s₃, then s₁It is s₂Direct precursor, s₃It is s₂Immediate successor.

Extracting the method that the described a certain Web service composite gone wrong describes information in described step 2 is: by SCA Composite file extended, the service that<service>tag representation assembly can be provided by, its level attribute represents this Level residing for service；<function>label defines, with the form of operation, the function side that service provides in terms of interface and behavior two Method；The service that<reference>tag representation assembly is relied on, its relation attribute represents that the service quoted is interlayer or layer In；<function>tag representation needs to be relied on service provides the input parameter of operation, output parameter requirement, and refType is for depending on Rely type, have sequential relationship and mandatory relationship.

The described a certain Web service that this goes wrong is replaced with described this by described step 3 and meets external environment condition or user's need Whether the concrete grammar of the Web service wanted is: compare the QoS of Web service and meet and use go wrong described a certain Tenant's requirement of Web service, select meeting property best one.

The input and output matrix of similar Web service in described step 6 represents such as formula (1), io_ijRepresent Web service i and Web The similarity of input and output, 0 < io between service j_ij<1；Computational methods are as follows:

The matching degree of the input and output between two Web services is calculated by formula (2), essentially equal finger Web service The output parameter of i is equal to the input parameter of Web service j, io_ij=1；Comprise and refer to that the output parameter set of Web service i belongs to Input parameter sets in Web service j；The output parameter set that part comprises finger Web service i comprises the defeated of Web service j Enter parameter sets；The unequal output parameter set of finger Web service i is wrapped the most mutually with the input parameter sets of Web service j Relation io contained_ij=0；When matching degree is for comprising or partly comprising situation, need to calculate similarity io_ijSuch as formula (3), d table Show the distance between two Web services；H represents the height of two Web services child's Web service together；μ, ν are respectively Represent the weight shared by d and h, 0 < μ < 1,0 < ν < 1, μ+ν=1；

$M_{in\_out}=\left[\begin{array}{cccc}{\mathrm{io}}_{11}& {\mathrm{io}}_{12}& \mathrm{..}& {\mathrm{io}}_{1n}\\ {\mathrm{io}}_{21}& {\mathrm{io}}_{22}& \mathrm{..}& {\mathrm{io}}_{2n}\\ \mathrm{..}& \mathrm{..}& \mathrm{..}& \mathrm{..}\\ {\mathrm{io}}_{n1}& {\mathrm{io}}_{n2}& \mathrm{..}& {\mathrm{io}}_{nn}\end{array}\right]---\left(1\right)$

$similarity=e^{-\mathrm{\&mu;}d}\left(\frac{e^{vh}-e^{-vh}}{e^{vh}+e^{-vh}}\right)---\left(3\right).$

Constraint matrix M in described step 6_constraintRepresent such as formula (4), con_ijRepresent Web service i and Web service j Between the similarity of constraints, 0 < con_ij< 1, computational methods are as follows:

Constrained matching degree between two Web services is calculated by formula (5), the pact of essentially equal finger Web service i Bundle condition is equal to the constraints of Web service j, con_ij=1；Comprise and refer to that the constraints set of Web service i belongs to The constraints set of Web service j；Part comprises the constraint referring to that the constraints set of Web service i comprises Web service j Set of circumstances；The constraints set of unequal finger Web service i and the constraints set of Web service j；Wrap the most mutually Relation con contained_ij=0；When matching degree is for comprising or partly comprising situation, need to calculate similarity con_ijSuch as formula (6), d Represent the distance between two Web services；H represents the height of two Web services child's Web service together；μ,ν Represent the weight shared by d and h respectively；

$M_{constra\mathrm{int}}=\left[\begin{array}{cccc}{\mathrm{con}}_{11}& {\mathrm{con}}_{12}& \mathrm{..}& {\mathrm{con}}_{1n}\\ {\mathrm{con}}_{21}& {\mathrm{con}}_{22}& \mathrm{..}& {\mathrm{con}}_{2n}\\ \mathrm{..}& \mathrm{..}& \mathrm{..}& \mathrm{..}\\ {\mathrm{con}}_{n1}& {\mathrm{con}}_{n2}& \mathrm{..}& {\mathrm{con}}_{nn}\end{array}\right]---\left(4\right)$

$similarity=e^{-\mathrm{\&mu;}d}\left(\frac{e^{vh}-e^{-vh}}{e^{vh}+e^{-vh}}\right)---\left(6\right)$

Service interaction matrix M in described step 6_correlationValue be multiplied by dependency degree obtain by relying on type weights.

In described step 7, dependence matrix M computing formula is:

$M=\left[\begin{array}{cccc}{d}_{11}& d_{12}& \mathrm{..}& d_{1n}\\ d_{21}& d_{22}& \mathrm{..}& d_{2n}\\ \mathrm{..}& \mathrm{..}& \mathrm{..}& \mathrm{..}\\ d_{n1}& d_{n2}& \mathrm{..}& d_{nn}\end{array}\right]=M_{in\_out}*M_{constra\mathrm{int}}*M_{correlation}---\left(7\right).$

Present invention advantage compared with prior art and good effect are as follows:

(1) other deployment, other versions, other similar services being divided into problem to service are replaced in service by the present invention, solve comprehensively Determine service replacement problem, it is ensured that the correct operation of large-scale complex software systems.

(2) complicated dependence between sing on web of the present invention service, in addition to carrying out the coupling of input and output, also Carry out the Similarity measures of restriction relation, and calculate the business dependence pass set up between business scope inner question service and similar services System, improves the accuracy of replacement.

Accompanying drawing explanation

Fig. 1 is embodiments of the invention flow chart；

Detailed description of the invention

As it is shown in figure 1, the Web service replacement method based on dependence that the embodiment of the present invention is set up is: as a certain Web When service can not meet the change of external environment condition or user's request, identify problem service deployment scenario, version information and Direct precursor service in composite services and immediate successor service；In conjunction with from problem service composite, extraction in file is described Dependence, selects suitably service according to different situations and is replaced；When other of problem service are disposed and other versions In the presence of the most not, the immediate successor set of the service in similar services set comprises all follow-up service of problem service, calculates it With problem service similarity on input and output, constraints, interactive relation, select the service mated most to be replaced.

Implement step as follows:

Step 1, when a certain Web service can not meet the change of external environment condition or user's request, first identifies problem service Deployment scenario, version information and the direct precursor service in composite services thereof and immediate successor service.Assume there is dependence For < s₁,s₂>、<s₂,s₃>, represent Web service s₁Depend on Web service s₂, Web service s₂Depend on Web service s₃, So s₁It is s₂Direct precursor, s₃It is s₂Immediate successor.

Step 2, extracts input/output, constraints and call relation that this problem service composite describes in file.Will The composite file of service component architecture SCA (Service Composite Architecture, SCA) is extended. The service that<service>tag representation assembly can be provided by, its level attribute represents the level residing for this service；<function>marks Sign and in terms of interface and behavior two, define, with the form of operation, the functional method that service provides.<reference>tag representation assembly institute The service relied on, in its relation attribute represents that the service quoted is interlayer or layer；<function>tag representation needs to be depended on Relying service to provide the requirements such as the input parameter of operation, output parameter, refType, for relying on type, has sequential relationship, forces to close System etc..

Step 3, disposes if the service of this problem exists in other positions, checks whether the Web service being deployed in other positions meets External environment condition or user's request, whether the QoS comparing Web service meets tenant's requirement of use problem Web service, selects Meeting property best one.If there is a Web service to meet, then problem service is replaced with this service, jumps to step 8, Otherwise perform step 4.

Step 4, if problem service exists other versions, is similar set by version sets cooperation, is taken by the Web in version set Business compares with problem service, i.e. performs step 6；If problem service does not exist other versions, then perform step 5；

Step 5, if other of this problem service dispose and other versions do not exist, judges the straight of other Web services the most one by one Connecing whether successor set comprises all follow-up service of problem service, if comprising, putting this Web service under similar services set；

Step 6, calculates the input and output matrix M of similar services respectively_{in_out}, constraint matrix M_constraint, service interaction square Battle array M_correlation.The input and output matrix of similar services represents such as formula (1), io_ijRepresent input and output between service i and service j Dependency degree, 0 < io_ij<1；Computational methods are as follows:

The matching degree of the input and output between two services is calculated by formula (2), the essentially equal output ginseng referring to service i Manifold closes out_s_iIt is equal to service input parameter sets in_s of j_j, io_ij=1；Comprise the output parameter set referring to service i out_s_iBelong to input parameter sets in_s of service j_j；Part comprises the output parameter set out_s referring to service i_iComprise service Input parameter sets in_s of j_j；The unequal output parameter set out_s referring to service i_iInput parameter sets with service j in_s_jRelation io comprised the most mutually_ij=0；When matching degree is for comprising or partly comprising situation, need to calculate similarity io_ij, seeing formula (3), d represents the distance between two services；H represents two height servicing child's service together；μ,ν Represent the weight shared by d and h, 0 < μ < 1,0 < ν < 1, μ+ν=1 respectively.

Constraint matrix represents the matching degree of constraints, M between service i and service j_constraintComputational methods are defeated with reference to input Go out the computational methods of matrix.

Service interaction refers to the business dependence set up between specific area inner question service and similar services, service interaction matrix M_correlationValue be multiplied by dependency degree obtain by relying on type weights.

$M_{in\_out}=\left[\begin{array}{cccc}{\mathrm{io}}_{11}& {\mathrm{io}}_{12}& \mathrm{..}& {\mathrm{io}}_{1n}\\ {\mathrm{io}}_{21}& {\mathrm{io}}_{22}& \mathrm{..}& {\mathrm{io}}_{2n}\\ \mathrm{..}& \mathrm{..}& \mathrm{..}& \mathrm{..}\\ {\mathrm{io}}_{n1}& {\mathrm{io}}_{n2}& \mathrm{..}& {\mathrm{io}}_{nn}\end{array}\right]---\left(1\right)$

$similarity=e^{-\mathrm{\&mu;}d}\left(\frac{e^{vh}-e^{-vh}}{e^{vh}+e^{-vh}}\right)---\left(3\right)$

Step 7, calculate dependence matrix M, service is ranked up, it is recommended that go out can replacement problem service Web service.

$M=\left[\begin{array}{cccc}{d}_{11}& d_{12}& \mathrm{..}& d_{1n}\\ d_{21}& d_{22}& \mathrm{..}& d_{2n}\\ \mathrm{..}& \mathrm{..}& \mathrm{..}& \mathrm{..}\\ d_{n1}& d_{n2}& \mathrm{..}& d_{nn}\end{array}\right]=M_{in\_out}*M_{constra\mathrm{int}}*M_{correlation}---\left(4\right).$

Step 8, interception service is asked, and replaces the service of this problem.

Non-elaborated part of the present invention belongs to the known technology of those skilled in the art.

The above, the only detailed description of the invention in the present invention, but protection scope of the present invention is not limited thereto, and any ripe Know the people of this technology in the technical scope that disclosed herein, it will be appreciated that the conversion expected or replacement, all should contain in the present invention Comprise within the scope of, therefore, protection scope of the present invention should be as the criterion with the protection domain of claims.

Claims (8)

1. a Web service replacement method based on dependence, it is characterised in that comprise the following steps:

Step 1, when a certain Web service can not meet the change of external environment condition or user's request in a certain position, first identifies The deployment scenario of described a certain Web service, version information and the forerunner's set of service in composite services thereof that go wrong and Follow-up set of service；

Step 2, extracts input/output, constraints that the described a certain Web service that goes wrong describes in file and calls Relation, in following steps, the input/output argument of problem Web service, constraints are all obtained by this step；

Step 3, disposes if the described a certain Web service gone wrong exists in other positions, checks and be deployed in other positions Whether Web service meets external environment condition or user's request, meets, then by this problem Web service if there is a Web service Replace with the described Web service meeting external environment condition or user's needs, jump to step 8, otherwise perform step 4；

Step 4, if there is other versions in the described a certain Web service gone wrong, using above-mentioned every other version as similar Set, compares each Web service in similar set with the described a certain Web service gone wrong, i.e. performs step Rapid 6；If the described a certain Web service gone wrong does not exists other versions, then perform step 5；

Step 5, if other of this described a certain Web service gone wrong dispose and other versions do not exist, sentences the most one by one Whether the immediate successor set of service of other Web services disconnected comprises all follow-up service of problem Web service, will comprise problem All other the follow-up Web services of Web service put similar set under；

Step 6, calculates each similar Web service and the similarity of problem Web service in similar set, obtains input and output Matrix M_{in_out}, constraint matrix M_constraint, service interaction matrix M_correlation；

Step 7, according to the input and output matrix M in step 6_{in_out}, constraint matrix M_constraint, service interaction matrix M_correlationCalculate dependence matrix M, Web service is ranked up, it is recommended that go out to replace go wrong described a certain Web service；

Step 8, intercepts the use request that user sends to problem service, replaces this problem Web service.

Web service replacement method based on dependence the most according to claim 1, it is characterised in that: step 1 institute State direct precursor service and immediate successor service is: assume that having dependence is < s₁,s₂>、<s₂,s₃>, represent Web service s₁ Depend on Web service s₂, Web service s₂Depend on Web service s₃, then s₁It is s₂Direct precursor, s₃It is s₂Direct Follow-up.

Web service replacement method based on dependence the most according to claim 1, it is characterised in that: described step The side of input/output, constraints and call relation that the described a certain Web service gone wrong describes in file is extracted in 2 Method is: extended by the composite file of SCA, the service that<service>tag representation assembly can be provided by, its level Attribute represents the level residing for this service；<function>label defines service in terms of interface and behavior two with the form of operation and carries The functional method of confession；The service that<reference>tag representation assembly is relied on, its relation attribute represents that the service quoted is In interlayer or layer；<function>tag representation needs to be relied on service provides the input parameter of operation, output parameter requirement, RefType, for relying on type, has sequential relationship and mandatory relationship.

Web service replacement method based on dependence the most according to claim 1, it is characterised in that: in step 3 The concrete grammar that this problem Web service replaces with the described Web service meeting external environment condition or user's needs is: ratio Whether meet the tenant's requirement using the described a certain Web service gone wrong compared with the QoS of Web service, select meeting property Good one.

Web service replacement method based on dependence the most according to claim 1, it is characterised in that: described step The input and output matrix of similar Web service in 6 represents sees formula (1), io_ijRepresent defeated between Web service i and Web service j Enter the similarity of output, 0 < io_ij<1；Computational methods are as follows:

The matching degree of the input and output between two Web services is calculated by formula (2), essentially equal finger Web service The output parameter set out_s of i_iIt is equal to input parameter sets in_s of Web service j_j, io_ij=1；Comprise finger Web clothes The output parameter set out_s of business i_iBelong to input parameter sets in_s of Web service j_j；Part comprises finger Web service i Output parameter set out_s_iComprise input parameter sets in_s of Web service j_j；The output of unequal finger Web service i Parameter sets out_s_iInput parameter sets in_s with Web service j_jRelation io comprised the most mutually_ij=0；Work as matching degree During for comprising or partly comprising situation, need to calculate similarity io_ijSeeing formula (3), d represents the distance between two Web services； H represents the height of two Web services child's Web service together；μ, ν represent the weight shared by d and h respectively, and 0 < μ < 1, 0 < ν < 1, μ+ν=1；

$M_{in\_out}=\left[\begin{array}{cccc}{\mathrm{io}}_{11}& {\mathrm{io}}_{12}& \mathrm{..}& {\mathrm{io}}_{1n}\\ {\mathrm{io}}_{21}& {\mathrm{io}}_{22}& \mathrm{..}& {\mathrm{io}}_{2n}\\ \mathrm{..}& \mathrm{..}& \mathrm{..}& \mathrm{..}\\ {\mathrm{io}}_{n1}& {\mathrm{io}}_{n2}& \mathrm{..}& {\mathrm{io}}_{nn}\end{array}\right]---\left(1\right)$

$similarity=e^{-\mathrm{\&mu;}d}\left(\frac{e^{\mathrm{\&nu;}h}-e^{-\mathrm{\&nu;}h}}{e^{\mathrm{\&nu;}h}+e^{-\mathrm{\&nu;}h}}\right)---\left(3\right).$

Web service replacement method based on dependence the most according to claim 1, it is characterised in that: described step Constraint matrix M in 6_constraintRepresent and see formula (4), con_ijRepresent constraints between Web service i and Web service j Similarity, 0 < con_ij< 1, computational methods are as follows:

Constraints matching degree between two Web services is calculated by formula (5), essentially equal finger Web service i Constraints set con_s_iIt is equal to the constraints set con_s of Web service j_j, con_ij=1；Comprise finger Web The constraints set con_s of service i_iBelong to the constraints set con_s of Web service j_j；Part comprises finger Web clothes The constraints set con_s of business i_iComprise the constraints set con_s of Web service j_j；Unequal finger Web service i Constraints set con_s_iConstraints set con_s with Web service j_j；Relation con comprised the most mutually_ij=0； When matching degree is for comprising or partly comprising situation, need to calculate similarity con_ijSeeing formula (6), d represents two Web clothes Distance between business；H represents the height of two Web services child's Web service together；μ, ν represent d and h institute respectively The weight accounted for；

$M_{constra\mathrm{int}}=\left[\begin{array}{cccc}{\mathrm{con}}_{11}& {\mathrm{con}}_{12}& \mathrm{..}& {\mathrm{con}}_{1n}\\ {\mathrm{con}}_{21}& {\mathrm{con}}_{22}& \mathrm{..}& {\mathrm{con}}_{2n}\\ \mathrm{..}& \mathrm{..}& \mathrm{..}& \mathrm{..}\\ {\mathrm{con}}_{n1}& {\mathrm{con}}_{n2}& \mathrm{..}& {\mathrm{con}}_{nn}\end{array}\right]---\left(4\right)$

$similarity=e^{-\mathrm{\&mu;}d}\left(\frac{e^{\mathrm{\&nu;}h}-e^{-\mathrm{\&nu;}h}}{e^{\mathrm{\&nu;}h}+e^{-\mathrm{\&nu;}h}}\right)---\left(6\right).$

Web service replacement method based on dependence the most according to claim 1, it is characterised in that: described step Service interaction matrix M in 6_correlationValue be multiplied by dependency degree obtain by relying on type weights.

Web service replacement method based on dependence the most according to claim 1, it is characterised in that: described step In 7, dependence matrix M computing formula is:

$M=\left[\begin{array}{cccc}{d}_{11}& d_{12}& \mathrm{..}& d_{1n}\\ d_{21}& d_{22}& \mathrm{..}& d_{2n}\\ \mathrm{..}& \mathrm{..}& \mathrm{..}& \mathrm{..}\\ d_{n1}& d_{n2}& \mathrm{..}& d_{nn}\end{array}\right]=M_{in\_out}*M_{constra\mathrm{int}}*M_{correlation}---\left(7\right).$