Presentation 8. Towards a dynamic design architecture

Summary form only given. This paper proposes an alternative architecture for systems of systems design. Systems integration has become a costly national problem. Witness the Joint Strike Fighter whose code base has driven massive cost over-runs. Consider the magnitude of revamping the National Airspace, to say nothing of inclusion of all manner of unmanned air vehicles into the mix. Recall the Army's failed Future Combat System, which was brought down by network complexity issues. Clearly, a new architectural paradigm must arise to manage the ever-morphing design of systems of systems. The current breed of static architecture simply cannot accommodate the effects of the hypernetworks that underlie modern system of system technologies. Hypernetworks, or networks of networks, are increasingly pervasive. In such environments, the link between cause and effect is at best stochastic, dependent upon a prolifera of quasi-related state-changing variables all interacting dynamically in real-time. Catastrophic failure always lurks at the edges between relative stability - chaos on one extreme and rigidity at the other. Nonetheless, modern system design still tends to favor the more comfortable side of rigidity where cause and effect are better correlated. The Department of Defense Architectural Framework (DODAF) is one such instance. Dependent upon relational database technology, DODAF cannot effectively manage the multitude of many-to-many relationships necessary to define the hypernetworked world of system of systems. Consider a modern aircraft where digital flight controls interact over a bus with digital engine controls, navigation system controls, environmental control systems and external control networks, all fed by continuous streams of data from myriads of sensors. This is merely a weak forerunner to the growing Internet of Anything (IoA). These phenomena defy the type of static cause and effect relationships predominant in existing reductionist design paradigms. They simply lack the robust and simultaneous many to many relationships necessary to describe hypernetworks. Spring-boarding on award winning work conducted between 2000-2006, this paper proposes a DODAF compliment means of better specifying systems design in such fashion to permit dynamic, time-domain driven interactions to come to the fore. The early premise was that the DODAF Operational View-5 (OV-5) “Activity Diagram” could actually decompose to the lowest permissible system view. Ironically, the limitation of this view is driven by proprietary software, where performance changing Computer Software Configurable Items (CSCI) remain hidden from view. Other DODAF views can be derived from the OV-5 relationships with adequate definition of properties, including appropriate state and temporal values. Assuming an underlying object oriented data base, a living ontology can be defined to permit synthesis of suitable mission driven OV-5 based scenarios reaching down to the interactions among component systems and subsystems. The follow-on effort builds upon this premise. It permits a user friendly way of building OV-5 activity modules complete with well defined processes, inputs, outputs, constraints and mechanisms (agents). The process repository is envisioned to lives in a Wordpress environment, open to all eligible contributors, persisting process modules composed of verb phrase controlled triples. Synthesis of the modules into workable scenarios, however, takes place in a graph database where graph traversal reveals the multiplicity of relationships necessary to analyze system dynamics. Here, graph metrics can be brought to bear to begin to optimize design solutions that serve to provide performance bounds within tolerances to assure an acceptable range of hypernetwork stability.