Appendix A
Future Manufacturing Environment—One Person’s View
Given the trends in semiconductor device, software, system, and communication technology, it is reasonable to expect that future semiconductor manufacturing facilities should have sophisticated manufacturing equipment that can be integrated and can interoperate with factory and enterprise systems in a seamless and cooperative manner. My vision of this would include the following:
-
Standardized enterprise manufacturing object models;
-
Standardized object classes for creating frameworks of manufacturing object behaviors;
-
Self-revealing object class definitions provided with the equipment made available by semiconductor manufacturing equipment suppliers;
-
Modeling, simulation, and consistency management tools that can create and utilize the above object models;
-
Standardized object “wire” syntax abstractions and definitions;
-
Self-revealing object semantics based on standardized enterprise manufacturing models;
-
Standardized distributed knowledge servers, objects, and tools for consistent enterprise-wide management of models, names, semantics, object relationships, object messaging, transactions, syntax translations, protocols, business rules, and other common knowledge needed for factory and enterprise object interaction;
-
Standardized spread-spectrum wireless transceivers and protocols;
-
Standardized networks operating at speeds of gigabits per second and greater and featuring multiple dynamically sized channels; intelligent caching; self-adapting and fuzzy-logic-based dynamic modifications of bandwidth resource allocations; routing; flow control and switching based on packet traffic flow patterns; object interaction criteria; and information usage patterns;
-
Standardized “information-bus”-type communication protocols utilizing name- or object-based addressing schemes;
NOTE: As part of his July 15, 1993, presentation to the committee, John Birchak of Intel Corporation submitted this comprehensive description of his vision of manufacturing in the year 2010. This vision emphasizes object-oriented approaches, a view not shared by the entire committee.
-
Standardized interpretive definition languages for object structure definitions robust enough to define all types of real and abstract objects (e.g., structured and unstructured data, voice, video, graphics, and images);
-
Standardized high-level communication protocols optimized for object movement and interobject communication;
-
Utilization of neural networks and fuzzy logic for internal calibration of equipment and intelligent localized control;
-
Utilization of neural networks and fuzzy logic for the intelligent management of object-level information flow and interaction on communication networks; and
-
Above all, market availability of the above items in common operating systems, languages, tools, and so on.
The above items could be applied in the following scenario:
A new piece of manufacturing equipment has arrived in the factory and has been installed. Standard installation hook-ups include electrical, fluids, ventilation, communication transceiver, and others. When the machine is powered up, it senses the presence of the transceiver coupling and broadcasts its presence via standardized syntax utilizing a common semantic for unsolicited announcements.
A factory equipment server (of which there may be many) sees the announcement and responds by asking the equipment to send its abstract object class definition. Upon receipt of this class definition, the factory equipment server uses the definition to invoke the equipment object upon the equipment itself. Thereupon the equipment describes its operational definitions to the server using standardized semantics. The equipment server builds a mapping of the equipment’s functional capabilities and state behaviors into the model of its particular factory and then instructs the equipment to set itself to one of its defined states.
The equipment server then initiates a similar object dialogue with its responsible factory server, and so on. The equipment server then initiates a dialogue with the equipment to run its specific diagnostic and material test runs and calibrate itself based on in situ or external process material tests and verifications. When these are completed, the equipment sends its specific calibration profiles to the equipment server. The equipment server then places the equipment into its appropriate waiting state.
At some point the equipment server receives a revision of its factory model, which now incorporates the new equipment as part of its control and interaction environment. Upon receiving this, the equipment server creates an abstraction of an appropriate model for the new equipment and passes this model abstraction to the equipment. This model defines the equipment’s behavior as a function of its own operational model and the operational model of the factory. The equipment, in turn, uses this model from the equipment server to map its parametric and operational capabilities to those of the factory equipment server.
At this point the equipment’s capabilities are known to all the relevant factory and enterprise-level servers, and the equipment is now configured with its appropriate factory role. This role defines the equipment’s interactions with other factory objects (servers, other equipment, human operators, facilities services, and so on.) and grants it privileges to carry out its role. As factory models change and adapt to local or enterprise circumstances, appropriate object behavior models respond and are modified according to synchronized horizon times or real-time as appropriate.
Upon notification from other objects, the equipment can from this point forward continue object-level interaction with other objects. The equipment is essentially an object node in the factory’s communication network.
The essence of this scenario is that complex equipment and machines behave as coordinated objects according to the consistent models provided. The models represent a consistent set of interactions and relationships between factory information and physical entities. The consistency of the model structures is ensured through standardization, and the consistency of the model relationships and contents is ensured by rule-based systems under factory and enterprise control. Model behaviors incorporate equipment control within inherent equipment capabilities (e.g., alarm levels, localized control/feedback, and state/process sequencing), as well as external behaviors such as publication of information required for local or enterprise systems (e.g., process control tolerances for design engineers at remote sites and interaction with other objects or services).
The underlying communication services are tailored to distributed object and model communication. Objects are aware only of other appropriate objects and not of the underlying communication mechanisms. The actual routing of messages and information is a function of the underlying distributed communication services and is not a concern of objects. Standardized object communication classes are inherited by objects used in the factory and enterprise, and so they may share certain common communication capabilities as well as have specific capabilities as required. Selection of appropriate routing, broadcasting, distribution, multicasting, flow, management, and other protocols and algorithms is contained within the factory or enterprise communication classes.
The above scenario, while emphasizing the object nature of equipment, does not underestimate the importance of the underlying communication mechanisms, object modeling, knowledge management, tools, and resources. In addition, it does not underestimate the substantial amount of industrial, academic, and governmental cooperation necessary to achieve the research, development, and general availability of the hardware, software, and standards needed to bring this about. (Note: This author has been involved in all of these facets and has great respect for the magnitude of the work necessary to bring them into reality.)