Mass service process fragment indexing and reusing method, equipment and software product

By building directed graphs and index trees, efficient search and sorting of service process fragments is achieved, solving the problem of low efficiency in reuse of service process fragments in traditional technology, and improving the flexibility and practicality of service combinations.

CN118945231AInactive Publication Date: 2024-11-12WUHAN UNIV
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202410959913.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-17
Publication Date
2024-11-12
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The operation unit of traditional service process search is an atomic service, which causes service process fragments to need to be manually programmed when reusing. Most search algorithms only consider a single QoS, resulting in inefficiency and low utilization, making it difficult to search and reuse service processes of different granularity. Fragment.

Method used

Through digital signature technology, the directed graph WS-graph is constructed, the user query conditions are converted into bit strings, SS-Tree+, and SSM-Tree+ trees are constructed, which supports the search of atomic services and service process fragments, and trims branches that do not meet the needs. And obtain service process fragments that meet the needs through the TSM sorting algorithm.

Benefits of technology

It realizes efficient discovery and selection of the best QoS service process segments, improves the search and reuse efficiency of service processes, reduces manual programming needs, and improves the flexibility and practicality of service combinations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118945231A_ABST
    Figure CN118945231A_ABST
Patent Text Reader

Abstract

The invention discloses a massive service process fragment indexing and reusing method, equipment and a software product, and the method comprises the steps: firstly, converting a user query condition and service into a bit string through a digital signature technology, and then constructing a directed graph WS-graph through a signature; secondly, an index SS-Tree is constructed, bit strings and QoS attributes of services are contained in the SS-Tree, so that SS-Tree + is formed, each inner node contains the aggregated QoS attributes, and the bit strings of one node are obtained through calculation of the bit strings of all child nodes of the node; then, each SS-Tree + serves as a leaf node, a new SSM-Tree + tree is constructed, and the search content of the new SSM-Tree + tree supports atomic services and service process fragments at the same time; and finally, based on service process fragment search of SSM-Tree +, cutting a search result, removing branches which do not meet requirements, and comprehensively sorting the remaining service process fragments to finally obtain a plurality of service process fragments or atomic services which meet the requirements and have optimal comprehensive QoS (Quality of Service). According to the method, the service process fragments can be quickly and effectively searched and sorted.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of service computing technology, and relates to a method, device and software product for indexing and reusing massive service process fragments. The method uses a VGI+ mechanism to efficiently find valid fragments in the service process, and combines SSM-Tree+ search and TSM process to return the service process fragment with the best QOS. Background Art

[0002] Due to the popularity of cloud computing technology in recent years, the concept of service has gradually permeated the entire Internet. Through the connection of the Internet, Web services can be used as a communication method between two applications, and it has become an integral part of the development of today's network software, and its role is becoming more and more important. Because of the development of today's technology, there are thousands of Web services based on Simple Access Protocol (SOAP) or Representational State Transfer (REST) ​​throughout the Internet, so the current progress of Web services is largely stimulated by the concept of service reuse. In order to effectively utilize existing services and cultivate a sustainable service ecosystem based on services, many online integrated development environments (OIDEs) have begun to transform concepts into commercial operations. For example, Google's GAE, Sina's SAE, Baidu's BAE, etc.

[0003] However, the operation unit of traditional service flow search is atomic service, which means that before a service is reused in a new service composition, it must be encapsulated as a completely independent service in advance. Therefore, even if a service flow instance fragment in the service group meets the requirements, the service fragment cannot be used due to the deviation of the overall service composition. In addition, software requirements are emerging in an endless stream. Even for automatic service composition (ASC), although its potential dependency chain can barely meet the user's needs for various services, the combined result still needs manual programming to be truly applied in actual scenarios. Due to the high complexity and cost of calculation, most service search and composition algorithms only consider a single QoS, which means that it is possible that the final result of the search and service composition has the global optimal response time, but the throughput is the worst. Therefore, the challenge is how to search and reuse service flow fragments of different granularities to combine into the required service composition.

[0004] Through the fast service process fragment indexing and sorting mechanism, service process fragments can be reused, which speeds up the search of service processes and meets more needs. Most service composition processes only use interfaces containing input and output parameter information, which leads to the entire service process after the composition being regarded as a whole, thus ignoring the service process fragments that can still be searched and reused. In fact, service process fragments of any granularity are reusable. For example, for a simple service process S1→S2→S3, it can only occur in sequence, and the service process fragments S1→S2, S2→S3 and the entire service process can be reused. The purpose of reusing service process fragments is to further utilize existing process fragments, by mining more sub-processes, whether the service process is atomic or composite, and reuse them as deeply as possible to meet the requirements of new service composition. The operating unit of software development is an executable service process fragment. Programmers only need to write a small amount of code to make the entire service run. Especially for automatic service composition algorithms, this will be a more flexible and practical solution. This service composition solution will no longer simply remain at the theoretical level, but will be directly used in commercial products. Summary of the invention

[0005] In order to solve the technical problem of efficiently discovering and selecting service flow fragments with the best QoS to meet different needs, the present invention provides a method, device and software product for indexing and reusing massive service flow fragments.

[0006] The technical solution adopted by the method of the present invention is: a method for indexing and reusing massive service process fragments, comprising the following steps:

[0007] Step 1: Use digital signature technology to convert user query conditions and services into bit strings, and then use the signature to construct a directed graph WS-graph;

[0008] Step 2: Construct the index SS-Tree, include the bit string and the QoS attributes of the service in the SS-Tree, and thus form the SS-Tree+. Each inner node contains the aggregated QoS attributes, and the bit string of a node is calculated from the bit strings of all its child nodes.

[0009] Step 3: Use each SS-Tree+ as a leaf node to construct a new SSM-Tree+ tree, whose search content supports both atomic services and service process fragments;

[0010] Step 4: Based on the service flow fragment search of SSM-Tree+, the search results are trimmed to remove the branches that do not meet the requirements, and the remaining service flow fragments are comprehensively sorted to finally obtain several service flow fragments or atomic services with the best comprehensive QoS that meet the requirements.

[0011] Preferably, in step 1, the directed graph WS-graph is represented by G p =(W,E), where P represents the combination process of the entire service fragment, W = {w1,w2,...,w m} is the vertex set of the directed graph, E={ <w i ,w j >}={ <w i .out→w j .in>} represents the edge set of the directed graph, i, j∈m; the vertex represents the Web service, and the edge represents the execution sequence between the connecting service and the connected service. <w i ,w j > in, w i Indicates connection service, w j Indicates the connected service, w i = <w i .in,w i .out>,w i .in indicates service w i The input parameter, w i .out indicates service w i Output parameters of

[0012] The vertex signature of a service is represented as vSig(w). The signatures of all its input and output parameters are bitwise ORed to obtain vSig(w).in and vSig(w).out respectively. Then the vertex signature is represented as vSig(w):=vSig(w).in*vSig(w).out, where * is a string concatenation operation.

[0013] For connection service i All output parameter signatures are bitwise ORed to obtain For the connected service j All input parameter signatures are also bitwise ORed to obtain Then the signature of the edge between services is expressed as

[0014] Preferably, in step 2, the index SS-Tree, for a given graph G*, each leaf node in the S-Tree corresponds to a vertex signature in G*;

[0015] If and only if there is an edge between v1 and v2 in G*, an edge is introduced between the two leaf node entries d1 and d2 in the S-Tree. In the SS-Tree, use express The edge of

[0016] A superedge between d1 and d2 is established if and only if there is an edge between the child node of d1 and the child node of d2. The signature of the superedge is obtained by performing a "bitwise OR" operation on the edge signatures between all child nodes of d1 and d2.

[0017] Preferably, in step 2, the SS-Tree+ adds a suffix vector to each vertex signature to embed the QoS information into the SS-Tree; the QoS represents several characteristics of the Web service, and all QoS will be reduced to the interval [0,1]; the value of each individual QoS attribute is calculated by the corresponding value of the aggregated component service.

[0018] Preferably, in step 3, the atomic service and the composite service are distinguished by judging whether the SSM-Tree leaf node has a self-edge, wherein each composite service has a self-edge.

[0019] Preferably, in step 4, the service process fragment search based on SSM-Tree+ first determines whether the service process fragment is included in the index. If not, only a bitwise AND operation is needed to match the root node of the SSM-Tree; then, the search result returns a set of SS-Tree root nodes that meet the query conditions, first adds the root node of the SSM-Tree to the stack, loops to take out the top node CurrentNode from the stack, and determines whether the input parameters and output parameters of all child nodes of CurrentNode meet the query conditions. If they meet, determine whether the child node is a leaf node and whether it contains self-edges. If both are satisfied, the child node is added to the Pset set (the Pset set is the root node set of the SS-Tree that meets the query conditions). If only the self-edge is included, the node is stacked and the loop continues until the stack is empty.

[0020] Preferably, in step 4, the search results are pruned to remove branches that do not meet the requirements; the relationship signature is used to locate the process containing the required service flow fragment, wherein the edges between all query condition vertices will be merged by a bitwise OR operation based on the signature, until the matching program that does not meet the vertex signature and edge signature can finally reach the leaf node, and the alternative process in the first step will be deleted; wherein, because each vertex in a service flow fragment search Q corresponds to a single service or even a process of multi-path operation, the dependency relationship between different paths of the process must be consistent with the query condition.

[0021] Preferably, in step 4, comprehensive sorting is performed, and the specific implementation includes the following sub-steps:

[0022] Step 4.1: The TSM sorting algorithm is to calculate the service process fragments SPFs that meet the functional requirements. It includes two steps: calculating the advantage relationship of the SPF fragments and obtaining the advantage set P u SPF i}Intersection AP set, disadvantage set F u SPF i}; Divide different areas to determine the current SPF set skyline and obtain the sorting.

[0023] Defining SPF i The dominant set of the u-th QoS attribute is P u SPF i}, which consists of more than SPF i The same QoS attribute has a smaller value of the service composition, P u SPF j}={SPF j |Q u (SPF j ) u (SPF i ),j∈[1,n]}Q u (SPF i ) indicates SPF i The u-th QoS attribute value, n represents the total number of SPF service flow fragments; define SPF i The inferior set of the u-th QoS attribute is F u SPF i}, which consists of services with a larger value of the same QoS attribute than SPF-i, i.e., F u SPF i}={SPF j |Q u (SPF j )>Q u (SPF i ),j∈[1,n]};

[0024] The intersection of all advantage sets is recorded as AP set, and the intersection of all disadvantage sets is recorded as AI set. For a specific SPF-i, all SPFs in the AP set can dominate it, and all SPFs in the AI ​​set are dominated by it. Calculate the AP set by Calculate the AI ​​set;

[0025] After determining the AP set and AI set of SPF-i, TSM divides different areas according to the following rules:

[0026] if Then SPF-i will be assigned to area A, which is the dominant area.​ Then SPF-i will be assigned to area B, which is the blind spot area. Then SPF-i will be assigned to area C, which is the dominated area;

[0027] A service set S = {s1,s2,s3,…,s m The skyline representation of} consists of services in S that are not dominated by any other service, that is, s k <s i The current skyline is updated according to its location. If SPF-i is in region C, it is dominated by one or more current skyline services, so it does not belong to the skyline and the skyline remains unchanged. If SPF-i is in region B, it is directly added to the current skyline. If SPF-i is in region A, it is directly added to the current skyline, and all SPFs in the AI ​​set of SPF-i are deleted from the skyline. The TSM ranking returned at the end is the skyline of the SPF.

[0028] The technical solution adopted by the device of the present invention is: a massive service process fragment indexing and reuse device, including a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor implements the massive service process fragment indexing and reuse method when executing the program.

[0029] The technical solution adopted by the software product of the present invention is: a software product, including a computer program, and when the computer program is executed by a processor, the method for indexing and reusing massive service process fragments is implemented.

[0030] Compared with the prior art, the beneficial effects of the present invention include:

[0031] (1) The present invention utilizes an improved variable granularity indexing mechanism (VGI+) to effectively solve the utilization and inefficiency issues in the traditional automatic service combination and service selection areas. The reusable granularity is not limited to the entire service flow or atomic service, but to the entire service flow fragment (SPF). The service flow fragments returned by the improved variable granularity indexing mechanism are directly executable fragments that do not require manual programming.

[0032] (2) The present invention uses SSM-Tree to achieve the same effect as the cluster in hash coding, but SSM-Tree has higher balance and can achieve higher performance.

[0033] (3) The present invention can quickly find and reuse service process fragments, and implement indexing of large-scale service process fragments under variable granularity conditions. The present invention constructs a tree in which the information of both functional nodes and non-functional nodes is encoded and stored in leaf nodes. Starting from the root of the tree, each tree layer is recursively searched and sorted until the last leaf node, which can quickly and effectively complete the search and sorting of service process fragments. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The technical solution of the present invention is further described below using embodiments and specific implementation methods. In addition, in the process of describing the technical solution, some drawings are also used. For those skilled in the art, other drawings and the intention of the present invention can also be obtained based on these drawings without paying creative work.

[0035] Figure 1 It is a service flow diagram of an embodiment of the invention;

[0036] Figure 2 It is a signature diagram corresponding to the service process of the embodiment of the invention;

[0037] Figure 3 is a schematic diagram of an embodiment of the invention in which a QoS suffix is ​​embedded in a SS-Tree+ node;

[0038] Figure 4 It is a schematic diagram of assigning a value to a service signature according to an embodiment of the invention;

[0039] Figure 5 It is a schematic diagram of the process of assigning a value to the signature of the connection between S1 and S2 according to an embodiment of the invention;

[0040] Figure 6 It is a schematic diagram of a fragment interface of the same service process in an embodiment of the invention;

[0041] Figure 7 is a schematic diagram of the superior and inferior sets of SPF-1 on each band in the embodiment of the invention;

[0042] Figure 8 It is a schematic diagram of the TSM parallel system of an embodiment of the invention. DETAILED DESCRIPTION

[0043] In order to facilitate ordinary technicians in the field to understand and implement the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the implementation examples described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0044] The present invention is used in the field of service computing, and can quickly find and reuse service process fragments, and realize indexing of large-scale service process fragments under variable granularity conditions. The technology constructs a tree in which the information of functional nodes and non-functional nodes is encoded and stored in leaf nodes. Starting from the root of the tree, each tree layer is recursively searched and sorted until the last leaf node, which can quickly and effectively complete the search and sorting of service process fragments.

[0045] Please see Figure 1 and Figure 2 The present embodiment provides a method for indexing and reusing a large number of service flow fragments, comprising the following steps:

[0046] Step 1: Use digital signature technology to convert user query conditions and services into bit strings, and then use the signature to construct a directed graph WS-graph;

[0047] Definition and construction of WS-graph, the directed graph WS-graph is represented by G p =(W,E), where P represents the combination process of the entire service fragment, W = {w1,w2,...,w m} is the vertex set of the directed graph, E={ <w i ,w j >}={ <w i .out→w j .in>} represents the edge set of the directed graph, i, j∈m; the vertex represents the Web service, and the edge represents the execution sequence between the connecting service and the connected service. <w i ,w j > in, w i Indicates connection service, w j Indicates the connected service, w i = <w i .in,w i .out>,w i .in indicates service w i The input parameter, w i .out indicates service w i Output parameters of

[0048] The vertex signature of a service is represented as vSig(w). The signatures of all its input and output parameters are bitwise ORed to obtain vSig(w).in and vSig(w).out respectively. Then the vertex signature is represented as vSig(w):=vSig(w).in*vSig(w).out, where * is a string concatenation operation.

[0049] For connection service i All output parameter signatures are bitwise ORed to obtain For the connected service j All input parameter signatures are also bitwise ORed to obtain Then the signature of the edge between services is expressed as

[0050] Definition of query satisfaction: For a query Q of a service process fragment, its WS-graph is represented by G q =(W',E'),G q It is used to satisfy the query result Q if and only if there is at least one path from service x to service y belonging to E';

[0051] There exists a path in E' that satisfies w' y .in∈w1.in∪w2.in∪...∪w m .in and w' x .out∈w1.out∪w2.out∪...∪w m .out,(w' x ,w' y )∈W', the search space of the service process fragment is converted into WS-graph, and then the digital signature technology is used to encode the entire WS-graph to complete the search process efficiently.

[0052] The signature operation of WS-graph consists of two parts: the signature of the vertex is represented by vSig(w), and the signature of the edge is represented by Each signature corresponds to a bit string. The input parameter set is represented by w.in and the output parameter set is represented by w.out. The signature operation process for the vertex is:

[0053] (1) The operation on the vertex signature vSig(w) is to first take each input parameter signature and then perform an OR operation: vSig(w).in:=vSig(w).in1|vSig(w).in2|...|vSig(w).in p , then take each output parameter signature and perform an OR operation: vSig(w).out:=vSig(w).out1|vSig(w).out2|...|vSig(w).out q , perform the following operation on the two obtained values: vSig(w):=vSig(w).in*vSig(w).out.

[0054] (2) In order to ensure that the signature is unique during the search and sorting process and to provide unique credentials for the input and output parameters, this paper defines θ different hash functions H i(i=1,2,...,θ), for example, DJBHash and PJWHash are hash functions that have been used. For each hash function, calculate j=(H i (w i .in)MODM), and set vSig(w).in i The j-th value bit in vSig(w).in is '1'. i In the example, the value of θ in M ​​bits is set to '1'. Similarly, the same encoding strategy is used to encode the signature vSig(w).out of the vertex output parameter. Figure 4 The process of assigning a signature is shown.

[0055] Edge Signature Similar to the vertex signature, it is composed of and The relationship between Figure 5 What is shown is Figure 1 The process of assigning the signature of the relationship between service S1 and service S2 in Figure 5 Although service S1 has 4 parameters, the edge signature Only the two input parameters {plan, fee} are encoded because only these two input parameters are relevant to the output of service S3.

[0056] Step 2: Construct the index SS-Tree, include the bit string and the QoS attributes of the service in the SS-Tree, and thus form the SS-Tree+. Each inner node contains the aggregated QoS attributes, and the bit string of a node is calculated from the bit strings of all its child nodes.

[0057] For the query process of a service flow fragment, the query condition can be converted into a query signature graph, namely QS-graph, which is similar to PS-graph (process signature graph) of WS-graph. The key to the problem is how to efficiently find a matching QS-graph in PS-graph. The following is a direct and feasible method:

[0058] (1) For each vertex v belonging to the QS-graph i , find such a list R i = {u i1 ,u i2 ,...u in}, needs to satisfy v i &u ij =v i ,u ij ∈V(PS-graph) and uij ∈R i , where '&' is the bitwise AND operator and V(*) is the vertex set.

[0059] (2) For these lists R i The dependency is calculated, and then the matching QS-graph is found in the PS-graph. In fact, the first step is the classic inclusion query. This query process has certain defects in encoding matching. If a single-level signature method is used, the whole process will be linearly dependent on the number of signatures, so it is very necessary to use a multi-level signature method.

[0060] Given a signature graph G*, the corresponding SS-Tree can be obtained by fusing the hierarchical inclusion relationship between vertex signatures and edge signatures. First, an S-tree is established based on all signatures in G*. Each leaf entry in the S-tree corresponds to a vertex signature in G*, and then all leaf entries are connected based on the structure of G*. More specifically, given two leaf entries d1 and d2 in the S-tree, d1 corresponds to u1, and d2 corresponds to u2. An edge is introduced between v1 and v2 if and only if there is an edge between them in G*; a super edge from d1 to d2 is established if and only if there is at least one edge from the child node of d1 to the child node of d2, and the signature of the super edge is obtained by performing a bitwise "OR" operation on all edge signatures from the child node of d1 to the child node of d2.

[0061] The specific implementation process of building SS-Tree+ with embedded QoS information is as follows:

[0062] It is very inefficient to sort all search results in real time. Most search engines use specific indexing technology to search and sort at the same time. Therefore, the QoS information of Web services must be integrated into SS-Tree first. QoS can be used to represent many characteristics of Web services, such as response time, credibility, etc. Since the feature space of each QoS is different, all QoS will be standardized to the interval [0, 1] for comparison. The normalization formula of positive attributes is used to normalize the QoS with larger values ​​and better quality, such as credibility. The normalization formula of negative attributes is used to normalize the QoS with smaller values ​​and better quality, such as response time and price.

[0063] By adding a suffix vector to each vertex signature, QoS will be embedded in the SS-Tree, thus forming SS-Tree+, process P = {s1,...,s n The QoS vector of} is defined as Q p = {Q'1(P),...,Q' r (P)}, where Q' i(P) The end-to-end estimated value of the i-th QoS attribute;

[0064] The specific QoS process is determined by the QoS of the corresponding composite service and the association structure between services. The individual QoS attribute value of each process can be calculated by aggregating the corresponding values ​​of the composite service. The previous control structure can be reduced or converted into a linear model.

[0065] Step 3: Use each SS-Tree+ as a leaf node to construct a new SSM-Tree+ tree, and search the content to support both atomic services and service process fragments;

[0066] Step 4: Based on the service flow fragment search of SSM-Tree+, the search results are trimmed to remove the branches that do not meet the requirements, and the remaining service flow fragments are comprehensively sorted to finally obtain several service flow fragments or atomic services with the best comprehensive QoS that meet the requirements.

[0067] The main difference between VGI and VGI+ is whether the QoS information is embedded. To simplify the description, only the service flow fragment search based on SSM-Tree is considered here.

[0068] Based on VGI+ service flow fragment query, for strict service flow fragment query, the result must fully meet the query requirements. Since the input and output parameters are distinguished for each service, this ensures that every successfully matched service can be called and the returned service flow fragment can run perfectly. This formula makes It is a super edge and can be combined with a large number of left edges, which means that all dependencies between Web services are consistent in the query environment and throughout the process;

[0069] Based on SSM-Tree, indexes are created on atomic services and service process fragments to achieve unified service query. The query of service process fragments is divided into the following two processes:

[0070] (2) First determine whether the service flow fragment is included in the index. If not, the query speed will become very fast because only a bitwise AND operation is needed to match the root node of the SSM-Tree. If the query condition can match the root node or even the core node, the calculation result of the OR operator in the SSM-Tree is that the service flow fragment does not exist. Each search and matching process may contain a service flow fragment that meets the query condition, and many traditional methods that consider control logic can be used to select appropriate fragments to achieve the purpose of reuse.

[0071] (2) The relationship signature is further used to locate the process containing the required service flow fragment, where all edges between the query condition vertices will be merged by a bitwise OR operation based on the signature until the matching program that does not satisfy the vertex signature and edge signature can finally reach the leaf node. The alternative process in the first step will be deleted. In particular, each vertex in a service flow fragment search Q may correspond to a single service or even a process corresponding to a multi-path operation. The dependency relationship between the different paths between the processes must be consistent with the query condition.

[0072] Based on the sorting process of VGI+, using TSM to sort the search results of service process fragments can have excellent performance in a large number of dynamic environments or fusions. The first step is the calculation of the dominance relationship, and the second step is the region division.

[0073] (1) SPF i The dominant set of the u-th QoS attribute is defined as P u SPF i}, which consists of more than SPF i The same QoS attribute of P u SPF j}={SPF j |Q u (SPF j ) u (SPF i ),j∈[1,n]}, n represents the total number of SPF service flow fragments; SPF i The inferior set of the u-th QoS attribute is defined as F u SPF i}, which consists of more than SPF i The same QoS attribute of F u SPF i}={SPF j |Q u (SPF j )>Q u (SPF i ),j∈[1,n]}, consider Figure 6 The two QoS attributes of SPF are shown in Figure 1. TSM constructs P-Tape and R-Tape. Figure 7 The corresponding advantage sets and disadvantage sets are shown. The corresponding dominance relationship can be obtained through the intersection of all advantage sets and disadvantage sets. The intersection of all advantage sets is called AP set, and the intersection of all disadvantage sets is called AI set. For a specific SPF i , all SPFs in the AP set can dominate it, and all SPFs in the AI ​​set are dominated by it, through Calculate the AP set by​ Calculate the AI ​​set;

[0074] (2) After determining the AP set and AI set of SPFi, TSM divides different areas according to the following rules:

[0075] if Then SPF-i will be assigned to area A, which is the dominant area.

[0076] if Then SPF-i will be assigned to area B, which is the blind spot area.

[0077] if Then SPF-i will be assigned to area C, which is the dominated area.

[0078] Figure 8 The TSM parallel system is shown and Figure 6 The SPFs of the current skyline (a service set S = {s1, s2, s3, ..., s m The skyline representation of} consists of services in S that are not dominated by any other service, that is, s k <s i ) may also be updated according to its own location, as follows:

[0079] If SPFi is in region C, then it is dominated by one or more current skyline services, so it does not belong to the skyline and the skyline remains unchanged.

[0080] If SPFi is in region B, then it is directly added to the current skyline.

[0081] If SPFi is in area A, it is directly added to the current skyline, and all SPFs in SPFi that are in the AI ​​set will be deleted from the skyline.

[0082] This embodiment also provides a massive service process fragment indexing and reuse device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the massive service process fragment indexing and reuse method is implemented.

[0083] This embodiment also provides a software product, including a computer program, which implements the method for indexing and reusing massive service process fragments when executed by a processor.

[0084] The present invention is an efficient service process fragment indexing and sorting mechanism, which can realize massive-scale service process fragment search and reuse under variable granularity conditions. The present invention uses the VGI+ mechanism to support the search and sorting of ultra-large-scale service process fragments. Under this mechanism, the functional information and non-functional information of the service process fragments are encoded and then stored in each node of the tree. Then, starting from the root of the tree, through each tree layer, including a large number of irrelevant service process fragments, until all leaf nodes are recursively searched and sorted. In the constructed sample data set, there are four million processes and one billion atomic services of CTG based on Web services. The present invention has high search efficiency and feasibility in this sample data set. With the popularization of online integrated development environments, the scale of services on the Internet is expanding, and the importance of service search and reuse will become more and more prominent.

[0085] It should be understood that the embodiments described above are part of the embodiments of the present invention, rather than all of the embodiments. In addition, the technical features in the various embodiments or single embodiments provided by the present invention can be combined with each other arbitrarily to form a feasible technical solution. Such combination is not restricted by the sequence of steps and / or the structural composition mode, but must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such combination of technical solutions does not exist and is not within the protection scope required by the present invention.

[0086] It should be understood that the above description of the preferred embodiment is relatively detailed and cannot be regarded as limiting the scope of patent protection of the present invention. Under the enlightenment of the present invention, ordinary technicians in this field can also make substitutions or modifications without departing from the scope of protection of the claims of the present invention, which all fall within the scope of protection of the present invention. The scope of protection requested for the present invention shall be based on the attached claims.

Claims

1. A method for indexing and reusing massive service process fragments, characterized in that: The following steps are involved: Step 1: Use digital signature technology to convert user query conditions and services into bit strings, and then use the signature to construct a directed graph WS-graph; Step 2: Build the index SS-Tree, include the bit string and the QoS attributes of the service in the SS-Tree, and thus form the SS-Tree+. Each inner node contains the aggregated QoS attributes, and the bit string of a node is calculated from the bit strings of all its child nodes. Step 3: Use each SS-Tree+ as a leaf node to construct a new SSM-Tree+ tree, and search the content to support both atomic services and service process fragments; Step 4: Based on the service flow fragment search of SSM-Tree+, the search results are trimmed to remove the branches that do not meet the requirements, and the remaining service flow fragments are comprehensively sorted to finally obtain several service flow fragments or atomic services with the best comprehensive QoS that meet the requirements.

2. The method for indexing and reusing massive service flow fragments according to claim 1, characterized in that: In step 1, the directed graph WS-graph is represented by G p =(W,E), where P represents the combination process of the entire service fragment, W = {w1,w2,...,w m } is the vertex set of the directed graph, E={<wi,wj>}={<wi.out→wj.in>} represents the edge set of the directed graph, i, j∈m; the vertex represents the Web service, and the edge represents the execution sequence between the connecting service and the connected service. <wi,w j > in, w i Indicates connection service, w j Indicates the connected service, w i = <w i .in,wi.out>,w i .in indicates service w i The input parameter, w i .out indicates service w i Output parameters of The vertex signature of a service is represented as vSig(w). The signatures of all its input and output parameters are bitwise ORed to obtain vSig(w).in and vSig(w).out respectively. Then the vertex signature is represented as vSig(w):=vSig(w).in*vSig(w).out, where * is a string concatenation operation. For connection service i All output parameter signatures are bitwise ORed to obtain For the connected service j All input parameter signatures are also bitwise ORed to obtain Then the signature of the edge between services is expressed as 3. The method for indexing and reusing massive service process fragments according to claim 1, characterized in that: In step 2, the index SS-Tree, for a given graph G*, each leaf node in the S-Tree corresponds to a vertex signature in G*; If and only if there is an edge between v1 and v2 in G*, an edge is introduced between the two leaf node entries d1 and d2 in the S-Tree. In the SS-Tree, use express The edge of A superedge between d1 and d2 is established if and only if there is an edge between the child node of d1 and the child node of d2. The signature of the superedge is obtained by performing a "bitwise OR" operation on the edge signatures between all child nodes of d1 and d2.

4. The method for indexing and reusing massive service flow fragments according to claim 1, characterized in that: In step 2, the SS-Tree+,adds a suffix vector to each vertex signature and embeds the QoS information into the SS-Tree; The QoS,represents several characteristics of Web services, and all QoS are normalized to the interval [0,1]. The value of each individual QoS attribute is calculated by the corresponding values ​​of the aggregated component services.

5. The method for indexing and reusing massive service flow fragments according to claim 1, characterized in that: In step 3, the atomic service and the composite service are distinguished by judging whether the leaf node of the SSM-Tree has a self-edge, where each composite service has a self-edge.

6. The method for indexing and reusing massive service process fragments according to any one of claims 1 to 5, characterized in that: In step 4, the service process fragment search based on SSM-Tree+ first determines whether the service process fragment is included in the index. If not, only a bitwise AND operation is needed to match the root node of the SSM-Tree; then, the search result returns a set of SS-Tree root nodes that meet the query conditions, first adds the root node of the SSM-Tree to the stack, and loops to take out the top node CurrentNode from the stack, and determines whether the input parameters and output parameters of all child nodes of CurrentNode meet the query conditions. If they meet, determine whether the child node is a leaf node and whether it contains self-edges. If both are satisfied, the child node is added to the Pset set. If only the self-edge is satisfied, the node is stacked and the loop continues until the stack is empty; wherein the Pset set is the root node set of the SS-Tree that meets the query conditions.

7. The method for indexing and reusing massive service flow fragments according to any one of claims 1 to 5, characterized in that: In step 4, the search results are pruned to remove branches that do not meet the requirements; the relationship signature is used to locate the process containing the required service flow fragment, where all edges between the query condition vertices will be merged by bitwise OR operation based on the signature, until the matching program that does not meet the vertex signature and edge signature can finally reach the leaf node, and the alternative process in the first step will be deleted; among them, because each vertex in a service flow fragment search Q corresponds to a single service or even a multi-path operation process, the dependency relationship between different paths between the processes must be consistent with the query condition.

8. The method for indexing and reusing massive service process fragments according to any one of claims 1 to 5, characterized in that: In step 4, comprehensive sorting is performed, and the specific implementation includes the following sub-steps: Step 4.1: The TSM sorting algorithm is to calculate the service process fragments SPFs that meet the functional requirements. It includes two steps: calculating the advantage relationship of the SPF fragments and obtaining the advantage set P u {SPFi} intersection AP set, inferior set F u The intersection AI set of {SPFi}; divide different areas to determine the current SPF set skyline and obtain the sorting. Defining SPF i The dominant set of the u-th QoS attribute is P u SPF i }, which consists of more than SPF i The same QoS attribute has a smaller value of the service composition P u SPF j }={SPF j |Q u (SPF j ) u (SPF i ),j∈[1,n]},Q u (SPF i ) indicates SPF i The u-th QoS attribute value, n represents the total number of SPF service flow fragments; define SPF i The inferior set of the u-th QoS attribute is F u SPF i }, which consists of services with a larger value of the same QoS attribute than SPF-i, i.e., F u SPF i }={SPF j |Q u (SPF j )>Q u (SPF i ),j∈[1,n]};​ The intersection of all advantage sets is recorded as AP set, and the intersection of all disadvantage sets is recorded as AI set. For a specific SPF i , all SPFs in the AP set can dominate it, and all SPFs in the AI ​​set are dominated by it, through Calculate the AP set by Calculate the AI ​​set; Determining SPF i After determining the AP set and AI set, TSM divides the different areas according to the following rules: if Then SPF i will be assigned to area A, which is the dominant area. Then SPF i Will be assigned to area B, which is the blind spot area. Then SPF i They will be assigned to Area C, which is the dominated area; The current skyline is updated based on its own location. If SPF i In region C, it is dominated by one or more current skyline services, so it does not belong to the skyline and the skyline remains unchanged; if SPF i In region B, it is directly added to the current skyline; if SPF i In area A, it is directly added to the current skyline, while SPF i All SPFs in the AI ​​set will be deleted from the skyline; the TSM ranking returned at the end is the skyline of the SPF; where a service set S = {s1, s2, s3, …, s m The skyline representation of} consists of services in S that are not dominated by any other service, that is, 9. A massive service process fragment indexing and reuse device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for indexing and reusing massive service process fragments as described in any one of claims 1 to 8 is implemented.

10. A software product comprising a computer program, characterized in that When the computer program is executed by a processor, the method for indexing and reusing massive service process fragments as described in any one of claims 1 to 8 is implemented.

Citation Information

Patent Citations

  • Service process searching and reusing method based on variable granularity indexing

    CN103064952A

  • Data search method and device, equipment and medium

    CN116383469A

  • Service oriented query and service query language framework

    US20160042068A1