23.03.2017 Views

wilamowski-b-m-irwin-j-d-industrial-communication-systems-2011

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

65-6 Industrial Communication Systems<br />

Even if the environment manager is still aware of who should not have access, ensuring that it happens<br />

is another issue. Again, ontologies and, more specifically, the use of a base and several sub-ontologies,<br />

provide a solution to this problem. By integrating an access concept (related to all types of information/<br />

tasks/components in the project) in the base and sub-ontologies, it can be specified that certain groups<br />

do not have access (read, write, none or all) to specific parts of the project.<br />

The second issue relates to rapid reconfigurability (aim2) and this has two dimensions, namely,<br />

(1) reconfigurability to bring about rapid production of a customized set of products demanded by the<br />

customer for delivery in a compressed timeframe and (2) introduction of new processes and/or components<br />

of a different level of granularity or made from new materials.<br />

Previously, in order to address dimension one, a considerable amount of work has been done on flexible<br />

manufacturing <strong>systems</strong> (FMS). However, these <strong>systems</strong> are inappropriate for the rapid production<br />

of customized products because they have only a fixed number of processes that are in place and these<br />

will probably be inadequate for the range of customization of products that might be required in product<br />

construction. In addition, FMSs frequently absorb considerably more resources than are needed to<br />

produce the specific product currently being produced. For these reasons, there has been a tendency in<br />

recent research to move to other means of achieving reconfigurability. In this chapter, we explore the<br />

two different approaches, namely, (1) use a combination of the ontologies and multi-agents [Hadzic 07]<br />

and (2) use the semantic Web paradigm through a combination of ontologies and Web services [Wang 04],<br />

[Dillon 07a,07b]. In approach (1), ontologies will be used as common knowledge base by agents working<br />

at different sites. The agents themselves will be goal oriented and will collectively or individually seek<br />

to achieve these, updating relevant instances of ontology classes when appropriate to allow monitoring<br />

of the status of various processes through this. Our previous work on Software Engineering Ontology<br />

[Wongthongtham 06] and Project Management Ontologies [Cheah 07] will help provide important<br />

inputs on the way this can be achieved. The approach adopted for the design of Semantic Web services<br />

will use further extensions of the RESTFUL and Web-based approaches and makes use of ontologies to<br />

help in the identification of Web services that will be specified through the use of the triple computing<br />

paradigm [Dillon 07a,07b]. Note this will not use the WSDL-based description but the use of a tuple<br />

space to locate the required Web service. Previous studies [Pilioura 04], [Banaei-Kashani 04] suggested<br />

that the existing Web services architecture based on remote procedure call (RPC) is not indeed “Web<br />

oriented” because RPC is more suitable for a closed local network and raised concern that “Web services<br />

do not have much in common with the Web” [Buschmann 96], and lead to potential weakness such as<br />

scalability, performance, flexibility, and implementability [Piers 03].<br />

Good manufacturing system architecture is crucial for multisite production. But how to define<br />

“good” or “bad?” In the keynote paper [Dillon 07a,07b] at IFIP NPC 2007 Conference, we reviewed<br />

four main architectural styles for SOA applications, namely, matchmaker, broker, peer-to-peer, Weboriented.<br />

The Representational State Transfer (REST) [Fielding 02] uses a resource identifier (URI) to<br />

provide an unambiguous and unique label for one particular Web resource. In the RESTful architectural<br />

style, all resources are accessed with a generic interface resulting in a dramatic decrease in the<br />

complexity of the semantics of the service interface during the service interaction W3C has recognized<br />

this REST style as an alternative to WSDL. Triple Space Computing [Bussler 05] is built on top of several<br />

technologies: Tuple Space [Gelernter 85], Publish/Subscribe paradigm [Eugster 03], Semantic Web,<br />

and RDF [Vinoski 02]. [Klyne 04] Tuple Space employs the “persistently publish and read” paradigm<br />

by leveraging the Tuple Space architecture and APIs. From an architecture perspective, Triple Space<br />

Computing is, in effect, based on the natural confluence of Asynchronous Broker and RESTful styles.<br />

The basic interactions between service provider and requester are rather straightforward: the service<br />

provider can “write” one or more triples in a concrete identified Triple Space. The service requester<br />

is able to “subscribe” triples that match with a template specified against its interests in a particular<br />

concrete Triple Space. Whenever there is an update in the spaces, the Triple Space will “notify” related<br />

service requesters indicating that there are triples available that match the template specified in its preceding<br />

subscription, as shown in Figure 65.1.<br />

© <strong>2011</strong> by Taylor and Francis Group, LLC

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!