In the past few years organizations have been turning to cloud as a way to reduce costs. Typically by outsourcing their non-core competencies. This brought a huge development to cloud services, leading to a crescendo amount of new providers, increasing competition, therefore, also increasing the func- tionality and capabilities of these services. This new paradigm poses a big challenge to organizations willing to adopt such services. Which service to contract? Which are the best options in the market for my problem? What services can be combined to achieve my goals? These are some of the ques- tions a decision maker has to answer when searching for cloud solutions to his organization. Sometimes a single service is not capable of providing what the organization needs, in these cases a composition of several services is needed. This adds more complexity to the comparison and decision that has to be made.
However, this is no longer just a search for services that fulfill the functional requirements for a determined problem, (supports JAVA, allows SSL, etc…), and that can work together. It has become a matter of distinguish which of those services provides a better composed solution, having in mind the organization goals. Therefore, having a greater importance given to the non- functional requirements, (price, security, availability, etc…).
Describing a service and its capabilities, both functional and non-functional, is a challenge. Publishing this description in a way that can be easily discov- ered is another challenge. Note that we are not referring to web-services but to cloud services, and therefore, the already existing approaches like WSDL do not suce. The amount of service capabilities also makes it very dicult to manually compare and choose the best composite solution. It has become much more than the cheaper is the best solution. Some organizations start to pay more attention to other details, security or portability for example, even if that has a price cost. This is why we think a decision aid process can help the person in charge by comparing and presenting a recommendation on what are the best solutions, based on pre-determined constraints.
What we propose in this thesis is to provide tools, methods and method- ology to help the decision maker during this process of search and choice of these services for composing a composite solution.
This document is an intermediate report of the final thesis to be presented in June 2013. We here give an introduction to the topics addressed in our work. As well as the literature review and contrast of these same topics. An use case is also proposed as a case study to support the feasibility study of the work done in this thesis.
[button link=”http://labs.sapo.pt/wp-content/uploads/2013/06/Thesis_SMS_v1.0.pdf” style=”download” color=”silver”]Semantic Mashups of Linked-USDL Services[/button]
WSDL is currently the preferred language to describe Web services. Unfortunately, WSDL is an interface description language that was created to enable computers to parse the interfaces, operations and parameters of distributed Web services. The language was not created to be used by humans. Nonetheless, it is a common practice for end-users to try to understand what a service does by looking at a WSDL description.
Recently, a new language to describe services from a business perspective, called USDL (Unified Service Description Language), was proposed by a W3C Unied Service Description Language Incubator Group and is being supported by companies such as SAP AG and Siemens. This language was created having in mind that its use would benefit end-users in searching and understanding the business value of a Web service. The argument lies on the fact that end-users search for services that can bring them an added value, and do not typically look for a service for its IT interface. The analogy would be that people do not buy cars based on technical characteristics such as spark plugs, batteries, suspension type, or oil pumps, but they look for a car based on the value and benefits it brings (e.g. fuel consumption, price, ecological performance, seats, etc.). Therefore, when end-users rely on WSDL descriptions to search for services, i.e. a technical description language, they are not using the best approach for the task and they should turn their attention into languages, such as USDL, which describe the business nature of a service.
The innovation of USDL is that it proposes a master data model for services” to describe various types of services ranging from professional to electronic services. It aims at a holistic service description putting a special focus on business aspects such as ownership and provisioning, legal constraints, availability indicators, business goals, industry domain, business models, release stages in a service network, composition and bundling, pricing and legal aspects among others, in addition to technical aspects.
In this project, we propose to explore the lifting of SAPO services’ descriptions from WSDL to USDL (WSDL-to-USDL). This will enable service providers (such as SAPO) to archive two goals: (1) use business description mechanisms to enable end-users to find services with a greater precision and recall, and (2) enable end-users to have a comprehensive visualization of Web services goals, functions and business value.