CN1968322A - Web service finding and integration proxy system - Google Patents

Abstract

The invention discloses a Web service discovery and integration agent system, which is a Web service discovery and integration scheme applied in a workflow system. The system prototype of the present invention, WSDIBroker, is built on the UDDI registration center, using the public or private UDDI registration center as the physical storage for saving Web services and service provider information, and WSDIBroker provides functions and interfaces of service registration, service search and modification to the outside world . WSDIBroker is mainly composed of four components: S-WSML interpreter, S-WSML/UDDI converter, service search engine and class comparator. The main components of the service lookup engine are three filters: name and text description filter, QoS metrics filter and interface semantics filter. The system of the invention can combine the tasks in the workflow system with the Web service to better meet business requirements.

Description

A Web Service Discovery and Integration Proxy System

technical field

The invention belongs to the technical field of Web service application, and in particular relates to a Web service discovery and integration scheme applied in a workflow system.

Background technique

The basis of Web services is based on open Internet standards such as HTTP, XML, SOAP, and WSDL. This technology allows all systems on the network to interact, and they uniformly use SOAP as the communication protocol for object calls, and XML as the standard data format. WSDL and UDDI are service description protocols and service discovery protocols that are generally supported by the industry today, and HTTP is the most widely supported transport layer protocol. Because it is based on open standards, the systems built by organizations and enterprises can provide their own services to the outside no matter what technology they use, thus making mutual interaction possible.

A workflow is the computer-aided or automated realization of a business process. It transfers or executes documents, information or tasks between different actors according to a series of process rules. The workflow management system provides automatic execution of business processes by managing the sequence of work activities, allocating and mobilizing relevant personnel or IT resources for different activities.

From the office automation system in the late 1970s to the present, the workflow management system has been developed for quite a long time. Today, workflow systems are used in a variety of contexts, from document-centric collaborative processing in office environments to automated processing of application data flows in enterprise application integration. At present, many software developers provide workflow management system products. The products provided by different software suppliers emphasize different functions and characteristics, and users also use different products to meet different needs.

The infrastructure and solutions supporting Web services are rapidly developing and maturing. More and more business organizations are attracted by the advantages of Web services in application integration, and begin to introduce Web services into their business processes. A business process is a collection of several business activities, which are linked back and forth according to certain rules and cooperate with each other in order to realize a complete function. A business activity is a practical link that can complete a specific function or task, and it usually targets specific application logic in an information system.

As a key technology for modern enterprises to realize process management and process control, workflow technology provides a complete framework from model analysis, establishment, management, simulation to operation for the business process of enterprises. At the same time, the workflow management system provides the whole process support for this framework through a set of integrated and interoperable software tools. After decades of development, workflow technology has gradually matured and has been applied in many fields such as banking, telecommunications, health care, manufacturing, and so on.

With the development and maturity of Web service technology, it is more and more urgent to use workflow to model Web service applications. How to combine tasks in workflow with Web services becomes very important.

Contents of the invention

The purpose of the present invention is to provide a Web service discovery and integration agent system, which can combine the tasks in the workflow system with Web services to better meet business needs.

The technical scheme of the present invention is as follows: a Web service discovery and integration agent system, including a process manager and a UDDI registration center, the system also includes an S-WSML interpreter, an S-WSML/UDDI converter, a service search engine and a class comparison device;

The S-WSML interpreter is written based on the API provided by Jena, and its functions include the extraction of attributes in the S-WSML Profile;

The S-WSML/UDDI converter is responsible for interacting with the UDDI registry;

The service search engine is responsible for obtaining various attributes of the service style from the S-WSML interpreter, and then using these attributes as conditions to search and query, and display the results to the queryer;

The class comparator is also written based on the API provided by Jena. Its function is to judge whether class A and class B come from the same ontology, and to judge the containment relationship between class A and class B.

The UDDI registry can be public or private, and is used to store information about Web services and service providers.

The four most important functions of the S-WSML/UDDI converter are:

A. Create the entity and service information structure in UDDI according to the Java object containing the description information of the Web service, and then call the save_business() and save_service() interfaces of UDDI to store the information in the UDDI service registry:

B. According to the index of the entity and service to be deleted, call the delete_business() and delete_service() interfaces of UDDI to delete the corresponding entity and service of the UDDI registration center;

C. Call the get_registeredInfo() interface of UDDI to obtain the brief information of all entities and services registered by a specific user from the UDDI registration center;

D. Call the find_service() interface of UDDI to find all services with S-WSML TModel from the UDDI registry for use by the service search engine.

The main components of the service lookup engine are three filters: name and text description filter (filter 1), QoS metrics filter (filter 2) and interface semantics filter (filter 3).

The working process of the service search engine is to firstly obtain various attributes of the service style from the S-WSML interpreter, and then use these attributes as conditions to provide all the The service instance referenced by S-WSML TModel performs three-layer filtering. Filters 1, 2, and 3 respectively match the service name and text description, the QoS measurement, and the interface semantics for the service instance and service style, respectively, and obtain syntactic similarity degree, QoS similarity, if any one of them is less than or equal to the bottom line, the service instance will be discarded and not returned as a query result, and all service instances that can finally pass filter 3 will be judged according to the syntax similarity or QoS similarity After sorting by degree or semantic similarity, it is displayed to the queryer.

In the processing of the interface semantic filter, since the input and output parameters are directly related to the classes in the ontology, the matching of the parameters needs to compare whether the classes corresponding to the parameters are in the same ontology, and also compare whether their corresponding classes have a containment relationship and How many common attributes are there.

The characteristics of the system of the present invention are as follows:

1) A Web service discovery scheme to assist Web service integration in a workflow system is proposed.

In this solution, fully considering the characteristics of Web services that are different from other general applications and the requirements of workflow systems for Web services, search for Web services that meet business needs from the following three aspects:

●The name and textual description of the Web service;

● Metrics of Web services in terms of quality of service (QoS);

●Interface (input and output) semantics of Web services.

2) Define a Web service description language that meets the requirements of the scheme.

Among the three aspects mentioned above, the second aspect is the requirement for Web services in consideration of the quality management of the workflow system. Due to the autonomy of Web services, Web services must have the ability to describe QoS characteristics by themselves. The third consideration is the semantic problem faced by Web service integration, the self-description of Web service must provide the semantic information of the interface. These are the two requirements put forward by the Web service discovery mechanism of this patent for the description language of Web services.

Description of drawings

Fig. 1 is the basic structural block diagram of system prototype WSDIBroker of the present invention;

Fig. 2 is a structural block diagram of a service search engine;

Figure 3 is a flowchart of the registration service;

Figure 4 is a structural block diagram of the integration of WSDIBroker and the workflow management system.

Detailed ways

The present invention will be further described below in conjunction with drawings and embodiments.

As shown in Figure 1, WSDIBroker is the prototype system of Web service discovery and integration broker in the Web service discovery and integration scheme of this paper. WSDIBroker is built on the UDDI registration center, using the public or private UDDI registration center as the physical storage for saving Web services and service provider information. WSDIBroker provides functions and interfaces for service registration, service search and modification.

WSDIBroker is mainly composed of four components: S-WSML interpreter, S-WSML/UDDI converter, service search engine and class comparator. The main components of the service lookup engine are three filters: name and text description filter, QoS metrics filter and interface semantics filter.

The S-WSML/UDDI converter is responsible for the interaction with the UDDI registry. The four most important functions of the S-WSML/UDDI converter are:

1. Create the entity and service information structure in UDDI according to the Java object containing the description information of the Web service, and then call the save_business() and save_service() interfaces of UDDI to store the information in the UDDI service registration center;

2. According to the index of the entity and service to be deleted, call the delete_business() and delete_service() interfaces of UDDI to delete the corresponding entity and service of the UDDI registration center;

3. Call the get_registeredInfo() interface of UDDI to obtain the brief information of all entities and services registered by a specific user from the UDDI registration center;

4. Call the find_service() interface of UDDI to find all services with S-WSML TModel from the UDDI registry for use by the service search engine.

As shown in Figure 2, the service lookup engine obtains various attributes of the service style from the S-WSML interpreter, and then uses these attributes as conditions to search all the S-WSML/UDDI converters that are extracted from the registry and have S -The service instance referenced by WSML TModel performs three-layer filtering. Filters 1, 2, and 3 respectively match the service name and text description, QoS measurement, and interface semantics for the service instance and service style, respectively, to obtain the grammatical similarity , the value of QoS similarity, if any one of the similarities is less than or equal to the bottom line, the service instance will be discarded and not returned as the query result. Finally, all service instances that can pass filter 3 are sorted according to the level of syntax similarity, QoS similarity, or semantic similarity, and then displayed to the queryer. The similarity baseline is a parameter specified by the user when accessing the lookup service. It indicates the range of matching service instances that the user wants to find. The service lookup engine includes three filters, namely name and text description filter (filter 1), QoS measurement filter (filter 2) and interface semantic filter (filter 3). Service instances are matched and screened by these three filters in the order of 1, 2, and 3. The choice of filter order mainly considers the overall processing speed. Changing the pass order of the three filters will greatly change the time required for the search.

In the processing of the interface semantic filter, since the input and output parameters are directly related to the classes in the ontology, the matching of the parameters needs to compare whether the classes corresponding to the parameters are in the same ontology, and also compare whether the classes corresponding to them have a containment relationship and have How many common attributes. In the scheme of this patent, the ontology is required to be described by DAML+OIL.

Class comparators are written based on the API provided by Jena. Jean is an open source Java API package for processing RDF, with the functions of reading, writing, modifying, browsing and querying RDF models. The main functions of the class comparator include: 1. Judging whether class A and class B come from the same ontology; 2. Judging the containment relationship between class A and class B.

The S-WSML interpreter is written based on the API provided by Jena, and its functions include the extraction of attributes in the S-WSML Profile.

Combined with the structure of the Web service discovery and integration solution in the workflow system, the drawings illustrate two application examples.

Application example one:

As shown in Figure 3, before the user has access to the registration service, UDDI user registration must be performed. After logging in to the service, the user can see brief information about all the services and corresponding entities registered by the user. Users can perform two operations: publish new services, or browse, select and log out of service entries. The publish operation needs to provide the S-WSML interpreter of the Web service to be published to interpret the S-WSML file and generate the corresponding Java data object. The S-WSML/UDDI converter is responsible for creating the entity and service information structure in UDDI based on the Java object containing the description information of the Web service, and then using the save_business() and save_service() interfaces of UDDI to store the information in the UDDI service registry . When the user performs the logout service operation, first select the service item. Get the index of the entity and service to be deleted, and then hand it over to the S-WSML/UDDI converter for further processing. The S-WSML/UDDI converter calls the delete_business() and delete_service() interfaces of UDDI to delete the corresponding entities and services in the UDDI registry. Serve.

Application example two:

As shown in Figure 4, WSDIBroker is a new layer between the ISFlow system located in the UDDI registration center. All Web service discovery and registration operations are completed through WSDIBroker, and the UDDI registration center is only used as the physical storage of WSDIBroker. The figure below shows the structure of the ISFlow system after integrating WSDIBroker. After integrating WSDIBroker, the process Web service packaging deployment and registration module uses the registration service of WSDIBroker to register the packaged and deployed process Web service to WSDIBroker, and the visual process definition tool searches for Web services that meet the service style through the search service of WSDIBroker. The process web service packaging deployment and registration tool, the visual process definition tool and the system application call interface must provide support for WSDIBroker, including: a S-WSML service description file and an extended WSDL must be added to the process web service packaging deployment and registration module File editing and generation tools, visual process definition tools, and system application call interfaces must add calls to Web services described by S-WSML. Since S-WSML uses WSDL as the "base", a large number of existing WSDL interpretation tools can get reused.

Claims (6)

1. A Web service discovery and integration agent system, including a process manager and a UDDI registration center, characterized in that the system also includes an S-WSML interpreter, an S-WSML/UDDI converter, a service search engine and a class comparator; The S-WSML interpreter is written based on the API provided by Jena, and its functions include the extraction of attributes in the S-WSML Profile; The S-WSML/UDDI converter is responsible for interacting with the UDDI registry; The service search engine is responsible for obtaining various attributes of the service style from the S-WSML interpreter, and then using these attributes as conditions to search and query, and display the results to the queryer; The class comparator is also written based on the API provided by Jena. Its function is to judge whether class A and class B come from the same ontology, and to judge the containment relationship between class A and class B. 2. The Web service discovery and integration agent system according to claim 1, characterized in that the UDDI registration center can be public or private, and is used to store Web services and service provider information. 3. The Web service discovery and integration agent system according to claim 1 or 2, characterized in that the four most important functions of the S-WSML/UDDI converter are respectively: A. Create the entity and service information structure in UDDI according to the Java object containing the description information of the Web service, and then call the save_business() and save_service() interfaces of UDDI to store the information in the UDDI service registration center; B. According to the index of the entity and service to be deleted, call the delete_business() and delete_service() interfaces of UDDI to delete the corresponding entity and service of the UDDI registration center; C. Call the get_registeredInfo() interface of UDDI to obtain the brief information of all entities and services registered by a specific user from the UDDI registration center; D. Call the find_service() interface of UDDI to find all services with S-WSML TModel from the UDDI registry for use by the service search engine. 4. The Web service discovery and integration agent system according to claim 1 or 2, characterized in that the main components of the service search engine are three filters: name and text description filter (filter 1), QoS Metric filter (filter 2) and interface semantic filter (filter 3). 5. The Web service discovery and integration agent system according to claim 4, characterized in that the working process of the service search engine is to first obtain various attributes of the service style from the S-WSML interpreter, and then use these attributes as conditions , perform three-layer filtering on all service instances referenced by S-WSML TModel that are extracted from the registry and provided by the S-WSML/UDDI converter. Filters 1, 2, and 3 respectively perform service names on service instances and service styles Matching with the text description, QoS metric matching and interface semantics matching, respectively get the values of syntax similarity and QoS similarity, if any of them similarity is less than or equal to the bottom line, the service instance will be discarded and not used as the query result All service instances that can finally pass filter 3 are returned and displayed to the queryer after being sorted according to the level of syntax similarity, QoS similarity, or semantic similarity. 6. The Web service discovery and integration agent system according to claim 4, characterized in that in the processing of the interface semantic filter, since the input and output parameters are directly associated with the classes in the ontology, the matching of parameters requires comparison of parameters Whether the corresponding classes are in the same ontology, it is also necessary to compare whether their corresponding classes have a containment relationship and how many common attributes they have.

Images