Valid and invalid WSDLs 

Context in source publication

Context 1

... service. Moreover, there is just a brief summary of the functionalities for the service available in its non structural part. Consider the following two web services as example. If we want to use a web service claiming to return digital weather forecast data of National Weather, MLserver.php, a consumer will look for its wsdl file for its description to understand, how to use it? And there is a fair description of the web service that can be found on the wsdl file for the service i.e. XML.wsdl, the web service description contains almost 534 words that can help the consumer how to use it, and the consumer is in a position to know functionalities provided by the web service. But if we consider a similar example as Lijie et al . has referred in his work. claiming to provide the weather report of all major cities around the world, looking on the wsdl file of the service on the link We will just find this non structural text “Get all major cities by cities around the world This information is still insufficient for a consumer to tell that how he will use the service, either it is free or for commercial use and other similar information. In this paper we focused our work in finding the descriptions about the web services written in non structural form. We have created an Active-X control and web interface as well for the same problem. It allows the user to input the name of the service then an attempt to extract the information present on different forums, and web pages that refer to explaining the functions of the web services are retrieved. After retrieval of pages the required data is extracted from the received information. As search engine don‟t provide the search of the WSDL document in particular, we have proposed a scheme for searching WSDL over the internet. This is discussed in detailed in section IV of the paper. The rest of the paper organization is as follows. The Section II presents the related work to the problem, in this section the WSDL validity and information in WSDL is also discussed in detail. Section III contains motivation of the work. Section IV describes the proposed scheme for finding description of the PWS over the web, Section V contains implementation details, evaluation and results of the proposed work and in last section conclusion and future work is discussed. II. R ELATED W ORK Web services gained the interest of many researchers in recent past. To fulfil the consumer requirement, many researchers dedicated their efforts on web services. A need of web service description language to the programmers is highlighted in Charles Petrie in his work. This need of description language for a programmer is an essence to see that what a software component written a web service can do and how they will use it [9]. Yan li et al . in their work on the public web services on the internet studied number, complexity, quality, and function diversity of the services on the internet. Transformation of web of data to web of semantics and services is continuing [10]. A web service mining framework is proposed by George et al . service compositions to emerge automatically in bottom up fashion is allowed by this framework. Pat. P. W. Chan provided an algorithm for building reliable web services. The web services are described by WSDL (web service describing Language) and their interaction with other services are described by Web Service Choreography Interface (WSCI). The performance evaluation, verification of deadlock free web services by Petri-Net modelling are their contribution towards web services [11]. The challenges in discovering web services are address by Eyhab Al-Marsi et. al., a web service relevancy ranking function is presented in their study. This function can be use in finding best available web service in web discovery process [12]. III. M OTIVATION The researcher work for finding the descriptions of the WSDL and enriching the information about the web service from the web has raised many interesting issues [13]. Releated to our work we are addressing two important issues from them. These issues are the motivation for our work. A web service is described by WSDL file; this file contains structural or non structural description of a web service. These descriptions help the user to understand what a service look like and what a service can do and how to use this service. Lejie et al . shown in their study in enriching the descriptions of public web services from the information captured from different web pages that most of the web services do not contain detailed descriptions in their WSDL files, and in most of the public web services the information given in the WSDL files are too few that it will never help the user to understand what a service can do for him actually. The same results were shown by Eyhab et al . in their study. They have focused UDDI business registries that are used for publishing and discovering the public web services and search engines for finding the WSDL. Fig. 1 shows a graph of the results given by Lejie et al . and Eyhab et al . According to Lejie et al . less then 50 % are valid WSDL, whereas Eyhab et al . searched web services from UBR (UDDI Business Registries) that is used for publishing and discovering Web services into registries contains almost 75% invalid WSDS and 87% were validated from the collected WSDL through search engines. Most of the web services existing today are having invalid or no WSDL files so the developer of the user can find them but he is unable to know about their functionalities and purpose of existence. The information that is present in valid WSDL is very less. Firstly there are low numbers of valid WSDL available and secondly the information present in these WSDL is unable to explain the functionalities of the web service. The Fig. 2 shows a graph that can be use too understand this statement in light of the work done by different researchers in recent past. It is clearly demonstrated that the information available in almost 72-83 % of valid WSDL is less the 1KB. IV. P ROPOSED S CHEME In this section the proposed scheme is described for collecting the description about the public web service over the web. As information about a Public web service is an essence for the consumer to know how to use it, and what are the functionalities it will provide to the consumer. Moreover these information if not available in the wsdl then it will produce a reason for the consumer to search it on the web, and most of the search engine like google, yahoo etc don‟t provide specific wsdl searches. Keeping this in view, we have designed different queries for a entered key word in the interface and passed it to search engines, not only the search engines were considered for searching the description of the public web services, but we have also searched different Public web service s directories through different API‟s available in open source languages i.e. Php. The queries that were taken in consideration contains the exact name, url of the service and synonyms from dictionary if exist. To limit our search we just considered top 15 retrieved results from them. These results are further refined to get a better result. As shown in Fig. 3, there are four major steps involve in the proposed scheme for finding the description of the web service. The user query the system using an interface named as “querying interface”, this interface accepts the input parameter. The input parameter from the user is passed to underlying layer of search engines, forums and Public Web Services Directories (PWSD). A dictionary is also attached with the interface that will check for any synonym of the word if available, and pass this word to underlying interfaces. But name to most of public web services are given in such a way that they have no synonym. For the testing purpose we have included google.com, bing.com, yahoo.com, and ask.com in our search engine list on which the queries were passed. We have use different API to retrieved result from them and the required data is extracted from the result. Different websites and forums are also checked for the description about the web service. Here is an algorithm that is used to pass information to yahoo.com; the similar algorithm is used for ask.com, bing.com, and google.com The required information is extracted from the data received from the searches made on web, but this relevancy is assigned after removing any redundancy of the links from the different searches. As in many cases the search engines result was near or almost same. The extraction of the data is followed by following step The received data from these the pages are now assigned a relevancy number this relevancy is required so as to get. To calculate the relevancy of the segments following formula is ...