SEMEASY (SEmantics makes Middleware EASY) is a project included in 8th Call of ITEA program, composed by a French and Spanish consortium. Its participants are Atos Origin, European Software Institute, Polytechnic University of Madrid, Thales Communications, Thales Services, BULL, Go Albert. The SEMEASY concept:
| “ | The general objective of the SEMEASY-ITEA project is to define and develop a common semantic infrastructure for SOA architectures based on middleware systems multi-layer, to develop the essential components and build tools which make easy the definition and development of these middleware components for the infrastructure users | ” |
Introduction edit
Currently, the Services Oriented Architecture (SOA) is promising to solve the well known issues of information systems agility. Unfortunately, the existing real world deployments show limitations.
- At the technical level, the middleware layers needed to support the deployment of SOA bring their own complexity, due in particular to the number of components to integrate and manage.
- At the business level, there is an inherent difficulty to find common vocabulary to describe the complexity of the business processes.
SEMEASY intends to build on the various middleware components and intends to better specify and develop those components, and provide ease of use to the business users by developing a common semantic infrastructure.
Challenges edit
The main challenges of SEMEASY are:
- How to provide business operational managers with tools that allow them to really drive a dynamic e-business and be able to quickly modify important business parameters and processes and a much faster way then is common today?
- How can these tools be used in a network of businesses and organizations that are interoperating/collaborating? What are the standards to enable this?
- How are we going to make these tools robust, usable and safe enough? How are we going to provide the right Quality of Service on these tools for this e-business environment?
- How to share knowledge efficiently and effectively between multiple companies? How to ensure the relevance of the information? How to deal with the diverse nature of information, both in form and content?
To take up these challenges, SEMEASY will innovate in the following areas:
- complexity management of technical modules:
- design, administration, configuration, security, maintenance, training;
- change management (versioning);
- interoperability of :
- technical components;
- semantics and relationships;
- business rules description;
- advanced cooperation services;
- security policies description;
- efficient knowledge management.
SEMEASY must hit the market and will first address 2 markets sectors (healthcare and e-Government) identified very early in the project process.
Consortium edit
- ATOS ORIGIN: Atos Origin (the merger between SchlumbergerSema and Atos ODS Origin) is the Spanish branch of a major international IT services company, Atos Origin plc. Atos Origin’s business is turning client vision into results through the application of consulting, systems integration and managed operations. The company's annual revenues are more than EUR 5 billion and it employs 47,000 people in 50 countries.
- BULL: BULL is one of Europe’s largest information technology companies specialising in the design, development, implementation and support of secure business critical IT infrastructures.
- THALES Communications/ THALES Services: THALES has leading European and world-wide positions in defence and aerospace electronics and is increasing its market share in the domain of Information Technologies and Services.
THALES is organised in three business areas: Defence (56%), Information Technologies and Services (26%) and Aerospace (18%).
- Universidad Politécnica de Madrid: Universidad Politécnica de Madrid (UPM) is one of the top research universities in Spain. It has over 50.000 students per year who major in various fields of engineering. The university has a large experience in European projects having participated in over 200 European projects in the last four years (FP5/FP6, Eureka/ITEA, etc.). The university already counts with several successful spin-off companies exploiting commercially results resulting from research projects.
- European Software Institute: Created in 1993, the European Software Institute (ESI) is a foundation launched by initiative of the European Commission and with the support of leading European companies and the Basque Government. The primary objective of ESI is to contribute to the development of the competitiveness of the European industry through the promotion, continuous improvement and knowledge in Information and Communication Technologies. To this end, ESI identifies, validates, packages and disseminates good software development and management practices.
- Go Albert: Go Albert was created in 1999 by a group of industrialists of the world of aviation and electronics. This origin naturally led the company to adopt processes of design and manufacturing from these industries giving to its products a level from reliability little common in the world of the software.
Technological State-of-Art edit
The current technological situation in the domain of Web Services, process automation, Web Services orchestration, Service Oriented Architecture, Smart Connectors, Registry Standards and Content Processing is presented here. A focus on the starting technological base of SEMEASY is summarized in Appendix 7.3.
In the domain of Web Services, the current state-of-the art is well advanced, in such a way that technologies and standards are available for describing, executing, securing, registering, deploying and managing them. A lot of efforts has also gone in making Web Services “technically” interoperable. Web Services have advanced to the process level, which allows for changing or orchestration of Web Services into a Complex (or Composite) Web Service, expressed in a standard execution language (BPEL).
In the domain of automation of process, the W3C and OASIS organizations propose languages able to describe Web resources in such a way software agents can be able to search and process automatically these resources. These new descriptions languages are ‘Ontologie Web Language’ (OWL) from the W3C and Web Service Modelling Ontology (WSMO) from OASIS. Various specific usages can potentially be derived from these orientations.
The Web Services orchestration servers of today allow business analysts and process experts to document process flows that cause immediate, actionable changes to how IT systems exchange data and events.
As Web services become more capable and popular, their strengths and weaknesses are more apparent. They could allow an unprecedented amount of sophisticated communication among business applications, but that communication would only be as reliable as the medium and the languages through which they communicate. The languages are easy to use but don't have the bullet proof reliability that message-oriented middleware provides. And the medium is the Internet ! One objective is to make event-driven development easier by adding event-broker software in the Enterprise Services Bus. ESBs will use standardized protocols to let event-driven applications communicate and make them as reliable as middleware-based applications.
In the domain of SOA (Service Oriented Architecture), projects to be successful, must accommodate a high level of variety and change involving a large number of systems, applications, data formats, standards, and connectivity types all existing in a perpetual whirl of change for both legacy systems, unstructured data sources (e.g. report files), and new applications. Driven by business and technical factors, this growing volatility makes the goal of establishing a workable and maintainable SOA strategy hard-to-achieve.
Smart Connectors have emerged as a promising alternative to traditional code generation approaches to SOA. Smart Connectors execute self-contained, reusable, application-based units of work with a heavy dose of integration glue. They can include business functions, transactions, or system service functions. But many of these existing systems are currently very hard to connect, because the necessary technological infrastructure is missing or is too expensive for many of the owners of these systems. Many of the systems used in these markets are requiring an event-driven approach, which is not offered by the current Web Services standard, architecture and tools. In addition, current mainstream connector technology does not cover the much more stringent security requirements.
In the domain of Registry standards, Web services registries, like other Web service components, need to be standards-based to foster interoperability across organizational boundaries. Most first-generation service registries were based on the UDDI standard, which focuses on registration of service descriptions (i.e., WSDL descriptions). This is an essential function, but SOA projects generate a broader array of service-related metadata and artifacts than just WSDL's. These include XML schemas, BPEL descriptions, XSLT transforms, and many others. Such artifacts also need to be centrally accessible to promote the benefits of reuse and control, and standard ways of storing and retrieving them, capabilities that aren't addressed by UDDI. Another standard in this area that accommodates these needs is ebXML Registry. It not only supports Web service registry functions, but also a tightly integrated repository and functions for the organization, storage, and control of any kind of service metadata or artifact. Both types of registries are currently only scarcely used by product vendors and there is still much room for further improvement on integration of these technologies in products, on visualization of registry entries and the general administration and design of registry artefacts.
In the domain of Data Services the current state-of-the-art includes basic Data Service modelling and generation (covered by the DSF as LASCOT result and tools like BEA AquaLogic).
In the domain of global security services based on XML standards (SAML, WS-Security, …) and on translation of security policies into technical security descriptions and actions (XACML, O-RBAC, …), the industry is today in the early stage of the process.
The objectives are identified by editors, the needs are expressed by aware customers but the practical way is to be clarified, evaluated and implemented. The semantic domain is conceptually quite new and its application to the security domain needs to be verified. On the other hand, editors, standardisation bodies and potential customers agreed that the high tech of new XML mechanisms such as XML security must be encapsulated and packaged into better ‘easy to use’ tools.
Main Results edit
The main result of the project will be a demonstrator applied to an existing complex business problem. Two scenarios of application (healthcare and e-government) have already been identified. This demonstrator will be based upon enhanced LASCOT components (BPEL, Common Information View (CIV), Security, Ontology) and new components.
Semantic Middleware edit
The Semantic Middleware contains a technical layer and a functional layer. The technical layer is the bridge between the lower level technical components and the functional layer, which includes the semantics. To enable a plug-in into the Semantic Middleware, each existing, enhanced or new component will provide a Semantic Interface that is able to translate the semantic instructions to lower level technical instructions.
Semantic Integration edit
The Semantic Integration is higher level pure semantic integration layer. This means that it doesn’t contain any technical components, but only semantic expressions in a standardised language (like OWL-S). The Semantic Integration artefacts are stored and version managed in the Semantic Repository. This repository includes links/associations between various semantic elements and ontologies and is responsible for the classification and unique identification of each semantic artefact.
GUI Integration edit
The semantic artefacts do have to be composed in a design environment especially built for humans and supporting a right ergonomy. For this purpose a GUI Integration is foreseen that allow humans to manage all the semantic artefacts in the Semantic Repository.
SOA Infrastructure edit
SEMEASY will incorporate an ESB solution based on open source. Within SEMEASY, the ESB components will be improved in term of security support, flexibility and easier management; i.e.: BPEL integration in an event driven architecture with web service security support and service registry; Smart Web Event Service Connectors infrastructure; UDDI extensions in order to provide service directory, message directory for the ESB platform.
Security Middleware edit
The security middleware will fully benefit from the innovative “semantic” approach proposed in the project. Added to the improvement of the XML security standards implementation, the extension of the security to the business process orchestration and the new way of description of security behaviour from a functional point of view, this new security middleware will reach a readiness to market that allows a rapid integration in commercial packages and operations.