Hello All,
Have you ever thought that modeling systems requires a huge amount of effort for not much payoff? Have you ever wondered whether a model could capture the wisdom and insights of a system architect, to enable people modifying the system years later to make successful modifications without compromising the essential elements of the system? Have you ever wished that a model would include information which allowed you to understand (if someone else developed the model) or communicate (if you developed the model) some of the rationale and implications of the decisions made in doing the system design?
This talk will focus on enriched modeling techniques for architecture specification and analysis. Current architecture capture and description is done in many places including databases, documents, architecture specifications, DODAF approaches, etc. This diverse and non-seamless set of representations makes it difficult to unify system semantics and understanding about completeness and relevance for complex architectures which are being developed today.
The approach presented here for enhanced modeling includes the idea of "enhanced model types". Enhanced model types build upon the concept pf abstract data types. In addition to specifying type, operation, and value constraints, information is added about the context within which the referencved element relates to the other elements and model context surrounding it.
An enhanced model type contains all the traditional information about an architecture element including its connections, intended behavior, state parameters, etc. In addition the model type enhancement provides additional constraints of three types about each model element: range and value constraints, consistency constraints referencing other elements within the model, and derivation constraints which cause the computation of one or more model elements from others. Constraints are expressed within the model semantics of the model, rather than being checked within the semantics of programmed code segments (which have been usually invoked to do the checking in past modeling approaches).
These constraints must be considered as they apply to both passive and active architecture elements. For active architecture elements, causality semantics provides the capability to specify sequencing, concurrency, conditions of triggering, activation, termination, and other complex process related control. In other words, this enhanced model approach includes time-dependent aspects of the system and its interactions, as well as rigorous static constraint aspects of the architecture. This is a dramatic enhancement over current modeling approaches.
For those of you who participated in the June 2016 SFBAC INCOSE chapter meeting, this August presentation will offer specific ways in which to address many of the challenges that David Long pointed out in his presentation. Come to our August meeting, and hear this very thought-provoking presentation!
Presenter’s bio:
Mark Gerhardt – Chief Architect of nHansa, Inc.
Mark Gerhardt is currently Chief Architect at nHansa, Inc. He has been involved for over 35 years in the conception, construction, and deployment of large and complex mission critical and high-performance software-based systems. Previous positions include software engineering laboratory deputy director at Lockheed Mission Systems, Chief Software Scientist at ESL, Inc., and involvement in many radar, sonar, and EW products at Raytheon, Lockheed, Boeing, and TRW. He was previously also a Chief Scientist at TPSI Inc. and a Chief Architect at TimeSys Inc., both vendors of schedulability and performance analysis tools. He has also taught graduate level courses in Performance Critical Design and Model Driven Architecture at a local university in San Jose, CA.
During his career, Mark has designed and built numerous real-time systems for signal and radar processing, computer architectures, and fault tolerant systems. He also designed and implemented major embedded software applications including C3I and early-warning receivers. Mark's interests include software and system engineering methods, system and software architecture, object-based languages, and Ada. Mark is the Past Chair of ACM SIGAda and was a Distinguished Reviewer for Ada95. He has done extensive work involving Rate Monotonic Analysis and the use of schedulability tools during architecture development. He is also involved in the IEEE 1471/ISO 42010 work for the recommended practice on how to capture architecture.
Mark has been involved with the generation of the schedulability performance and timing (SPT) profile for UML as well as the current modeling and analysis of real-time embedded systems (MARTE) profile for UML developed within the Object Management Group. Mark has been involved with evolving UML profiles for performance since their inception from UML_RT to the present MARTE work. Follow-on work included development of a real-time decomposition method for mission critical architecture development.
Current research work is focused on building a system and accompanying process that support enhanced model types including self-contained rules about integrity, consistency and completeness for model elements and architectures. A research product is being constructed using graphical databases, rule-based systems, and an enhanced storage repository. Both the system under development and the system process of developing it are represented within the system as layered meta-models. Continued enforcement of rule integrity maintains consistency about the development process and the developed artifacts when using this system.
Mark received his Bachelor of Electrical Engineering degree Magna Cum Laude from the City College of New York and his Master of Science in Engineering (Computer Science) from Princeton University.
There will not be a July meeting of the INCOSE SFBAC due to the IS taking place in Scotland this month.
The next SFBAC meeting will be on Monday, August 8 in the the Rinconada / Embarcadero Room of the Palo Alto Library, with the social time starting at 6 pm, and the talk from 6:30 to 7:30 pm. Mark Gerhardt, Chief Architect of nHansa, Inc., will be presenting "Modeling Made More Meaningful." Read on for more information about his presentation.
Have you ever thought that modeling systems requires a huge amount of effort for not much payoff? Have you ever wondered whether a model could capture the wisdom and insights of a system architect, to enable people modifying the system years later to make successful modifications without compromising the essential elements of the system? Have you ever wished that a model would include information which allowed you to understand (if someone else developed the model) or communicate (if you developed the model) some of the rationale and implications of the decisions made in doing the system design?
This talk will focus on enriched modeling techniques for architecture specification and analysis. Current architecture capture and description is done in many places including databases, documents, architecture specifications, DODAF approaches, etc. This diverse and non-seamless set of representations makes it difficult to unify system semantics and understanding about completeness and relevance for complex architectures which are being developed today.
The approach presented here for enhanced modeling includes the idea of "enhanced model types". Enhanced model types build upon the concept pf abstract data types. In addition to specifying type, operation, and value constraints, information is added about the context within which the referencved element relates to the other elements and model context surrounding it.
An enhanced model type contains all the traditional information about an architecture element including its connections, intended behavior, state parameters, etc. In addition the model type enhancement provides additional constraints of three types about each model element: range and value constraints, consistency constraints referencing other elements within the model, and derivation constraints which cause the computation of one or more model elements from others. Constraints are expressed within the model semantics of the model, rather than being checked within the semantics of programmed code segments (which have been usually invoked to do the checking in past modeling approaches).
These constraints must be considered as they apply to both passive and active architecture elements. For active architecture elements, causality semantics provides the capability to specify sequencing, concurrency, conditions of triggering, activation, termination, and other complex process related control. In other words, this enhanced model approach includes time-dependent aspects of the system and its interactions, as well as rigorous static constraint aspects of the architecture. This is a dramatic enhancement over current modeling approaches.
For those of you who participated in the June 2016 SFBAC INCOSE chapter meeting, this August presentation will offer specific ways in which to address many of the challenges that David Long pointed out in his presentation. Come to our August meeting, and hear this very thought-provoking presentation!
Presenter’s bio:
Mark Gerhardt – Chief Architect of nHansa, Inc.
Mark Gerhardt is currently Chief Architect at nHansa, Inc. He has been involved for over 35 years in the conception, construction, and deployment of large and complex mission critical and high-performance software-based systems. Previous positions include software engineering laboratory deputy director at Lockheed Mission Systems, Chief Software Scientist at ESL, Inc., and involvement in many radar, sonar, and EW products at Raytheon, Lockheed, Boeing, and TRW. He was previously also a Chief Scientist at TPSI Inc. and a Chief Architect at TimeSys Inc., both vendors of schedulability and performance analysis tools. He has also taught graduate level courses in Performance Critical Design and Model Driven Architecture at a local university in San Jose, CA.
During his career, Mark has designed and built numerous real-time systems for signal and radar processing, computer architectures, and fault tolerant systems. He also designed and implemented major embedded software applications including C3I and early-warning receivers. Mark's interests include software and system engineering methods, system and software architecture, object-based languages, and Ada. Mark is the Past Chair of ACM SIGAda and was a Distinguished Reviewer for Ada95. He has done extensive work involving Rate Monotonic Analysis and the use of schedulability tools during architecture development. He is also involved in the IEEE 1471/ISO 42010 work for the recommended practice on how to capture architecture.
Mark has been involved with the generation of the schedulability performance and timing (SPT) profile for UML as well as the current modeling and analysis of real-time embedded systems (MARTE) profile for UML developed within the Object Management Group. Mark has been involved with evolving UML profiles for performance since their inception from UML_RT to the present MARTE work. Follow-on work included development of a real-time decomposition method for mission critical architecture development.
Current research work is focused on building a system and accompanying process that support enhanced model types including self-contained rules about integrity, consistency and completeness for model elements and architectures. A research product is being constructed using graphical databases, rule-based systems, and an enhanced storage repository. Both the system under development and the system process of developing it are represented within the system as layered meta-models. Continued enforcement of rule integrity maintains consistency about the development process and the developed artifacts when using this system.
Mark received his Bachelor of Electrical Engineering degree Magna Cum Laude from the City College of New York and his Master of Science in Engineering (Computer Science) from Princeton University.
Thank You,
Robin Reil
