Distributed Object System Egineering for Terminal Aerodrome Forecast Validation and Metrics Processing

Distributed object systems are a very complex intertwining of heterogene-oushardware, software, and operating systems coupled with communication networks of varying protocols and capacities. Distributed components offer improved performance through parallel processing, improved expansion and scalability opportunities through modularity, improved availability through replication, and improved resource sharing and interoperability through interconnection. This research provides a distributed system design methodology to validate terminal forecasts and gather metrics for the Air Force Weather Agency. Proven principles such as component reuse and architectural development are applied through the use of parameterized types and design patterns. A client/server measurement model is developed to show the impact of design decisions on computer resource utilization, system scalability, system performance, ease of implementation, and system evolution. An experimental Common Object Request Broker Architecture (CORBA) application is implemented to quantify the approach's effectiveness toward selecting an appropriate CORBA implementation and deploying the application in a distributed environment. While this research specifically uses CORBA for system development, the methodology presented is easily mapped onto any client/server architecture.