Projects

Adaptive management of quality-assured service provision for sophisticated e-business processes in open dynamic environments

Stanisław Ambroszkiewicz (Institute of Computer Science)    Stanislaw.Ambroszkiewicz >at> ipipan.waw.pl

Description:The notion of e-business process is strongly related to the several crucial issues, such as the SOA paradigm, and the concept of e-business service. Classical version of the SOA paradigm has its origins in software engineering. From the point of view of e-business process, the concept of service need not be related to RPC (as it is in SOA) where a service is passive and is waiting for a client to be invoked. In other words, a service may be active and looking by itself for clients’ tasks that can be accomplished. This corresponds to the reverse auctions in business practice. A business service has well founded structure where its operations (corresponding to request-quote, order-contract, invoice-payment) are related to each other. These relations cannot be expressed explicitly in WSDL. Hence, the concept of service in SOA should be revised. Generally, from the e-business perspective the following interrelated problems seem to be important:
  1. Service architecture.
  2. Communication protocols in e-business processes.
According to the SOA paradigm, the activities in which service and client are engaged, are as follows: publication, discovery and service invocation. Usually, the invocation is preceded by a negotiation for a service level agreement (SLA) that is (by its nature) "output" based, i.e., the result of the service as required by the client is the subject of the agreement. The SLA negotiations require semantic interoperability, that is, understanding between client and service provider. Actually all activities related to how a service is used by a client (also publication and discovery) require understanding between the client and the service provider. These activities must be realized in the form of communication protocols as it is usually done in distributed systems. Since the system is supposed to be open and heterogeneous, the semantic interoperability cannot be hardcoded in the protocols. It must be incorporated in the contents of the protocol messages. The contents must be expressed in a generic declarative language. Client task as well as service description, and the output (final result) description of service performance must be expressed in this language. The important question is what this language is about? In other words, what is the grounding, i.e., the real (not abstract model-theoretic) semantics of the language. In our approach the grounding is a class of XML documents that services process (input documents and output documents). Once the documents are processed, they have precisely defined impact in the real world. Since this “impact” is hard (perhaps not possible) to describe formally, the proposed grounding is simple and sufficient to provide semantic interoperability. OWL was chosen as the language for describing XSD schemata of the documents processed by services. In our previous research projects the language was Entish [3, 4, 5], our own invention. To conclude, there are serious problems with Service architecture and communication protocols in e-business processes. The classic standards like Web services SOAP, WSDL, UDDI and BPEL, Semantic Web OWL-S, as well as ebXML and BPMN are not sufficient to solve the problems. There is a strong need to search for new standards necessary for creating open, heterogeneous and scalable systems for realizing complex e-business processes. These very new standards are the subject of the proposed research to be done by a PhD students.
Another e-business related topic concerns the application of the multi-agent technology to model and solve critical problems of the modern society: effective energy production and environment pollution reduction. In this case the starting point is the usage of classical mathematical modeling and optimization techniques which are to be enriched with a comprehensive account for, and handling of, the imprecision and uncertainty. PhD students carrying out research on the following and previous topics are expected to collaborate, both during their study in Poland and abroad, and a synergy effect is expected thanks to that.
You are here: Home Projects Adaptive management of quality-assured service provision for sophisticated e-business processes in open dynamic environments