1 / 30

Relationship-Oriented Architecture: A New Paradigm in System Engineering

Relationship-Oriented Architecture: A New Paradigm in System Engineering. Tobin Anthony, Ph.D. Christopher Costello Shawnee Heritage Government Solutions July 10, 2007. Agenda. Introduction The Problem Relationship-Oriented Architecture Example Case Studies. Introduction.

Download Presentation

Relationship-Oriented Architecture: A New Paradigm in System Engineering

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Relationship-Oriented Architecture: A New Paradigm in System Engineering Tobin Anthony, Ph.D. Christopher Costello Shawnee Heritage Government Solutions July 10, 2007

  2. Agenda • Introduction • The Problem • Relationship-Oriented Architecture • Example • Case Studies

  3. Introduction • Tobin Anthony, Ph.D. • 12 yrs with NASA GSFC • 10 yrs in aerospace and defense systems • Christopher Costello • 19 yrs in aerospace and defense systems • 12 yrs in small business engineering • Shawnee Heritage Government Sol’n • Majority-owned by Shawnee Indian tribe • Native-American 8(a) exemption pending • Principals in business since 2003

  4. ROA Concept • Simple Is Better Than Complex • A Picture Is Worth 1000 Words

  5. Diagnosis: Chaos The Problem: Inefficient management of design cycle leading to cost & schedule overrun The Cure: Develop a system architecture to guide system design But you end up with… Micromanagement A “Key Man” A Tool Expert Poor Communication

  6. Typical Program Office Program Manager System Engineer/ DPM SE 1 SE 2 SE 3 Subsystem 1 Subsystem 2 Subsystem 3 Subsystem 4 Subsystem 5 • Manages Effort • Interacts with Customer • Responsible for glue between subsystems • Manages requirement process Subsystem 6 Subsystem 7 Subsystem 8 How does the SE manage all these subsystems? How does the SE track the progress of the design? Only one engineer understands the entire system

  7. ROA Solves The tool expert by… Being tool independent The key man by… Simplifying the process Poor communication by… Providing illustrative products Micromanagement by… Defining architectural relationships

  8. What is ROA? • History • Developed by authors for use on UAV ground system program • Honed by years of frustration seeing programs “winging it” • Concept • Database, not tool-oriented • Simple means of illustrating system by emphasizing relationships between elements

  9. ROA Concept - Simplicity • Rigorous, but not rigor mortis • Define a core set of system products • Illustration of the system architecture • Minimal set to document and communicate the design • Engineer identifies systems elements and defines the relationship between them • Examples of system elements are: • Requirements • Functions • Components (HW or SW) • Test procedures • Development milestones Relationships Between Elements Determine How the System Is Used & Built to Meet Requirements

  10. ROA Concept – 1000 Words

  11. Glue the System Together Interfaces Requirements State Transitions Top Level Model Test Test Test Procedures Use Cases ? Schedule Any Element

  12. Perform Still Imaging Perform Frame Image Perform Image Transfer Perform Store Image Perform Image Capture ROA: A Simple Example System Requirement:Take a Picture

  13. Perform Frame Image Zoom Target Perform Site Target Perform Focus Target Target via View Finder Target via LCD ROA: A Simple Example System Requirement:Frame Image (derived)

  14. Perform Site Target Target via LCD Target via View Finder Retrieve Image from Focal Plane Perform Store Image Multiple Image Capture Write Image to Flash Perform Image Capture Single Image Capture ROA: Camera Development System Requirements: Capture, Store Image, Site Target (derived) Good News! No dependencies so nothing else effected X Bad News! Not enough room for a lot of flash memory! Not all the bells and whistles but it will sell! Marketing Really Wanted Perfectly Engineered!

  15. Perform Still Imaging Perform Site Target Target via View Finder Target via LCD Perform Image Transfer Retrieve Image from Focal Plane Perform Store Image Perform Image Capture Single Image Capture Multiple Image Capture Write Image to Flash ROA: TLM/Req/WBS Consistency WBS & IPT Structure Matches Architecture! To Multiple Levels!!! One Architecture Providing the Glue for All Management and Engineering Efforts!

  16. Benefits of ROA • Relationships between elements are defined • Enables evaluation of effects of changes to elements • System design can be linked with management activities • WBS • IPT structure development • EVMS • Elements are linked to requirements • Requirements without elements • Unverified requirements • Elements without requirements • Bloatware

  17. Benefits of ROA • Software Tool Independent • Can develop architecture in any software environment • Core • DOORS • Requisite Pro • System Architect • MS Office • Requires database back-end • Provides link between individual elements and management artifacts • Database can be simple spreadsheet

  18. Case Studies • Bringing UAV Ground System to CDR • Forensic Analysis of Safehold Mode Software

  19. Completing Design: Challenges • Behind schedule • Engineering signed up to unrealistic schedules • Management did not understand engineering needs • Management insight into progress limited • Mgmt to Engineering: “You’re Gold-Plating!” • Lack of trust through lack of communication • Could not get to CDR • Customer and management did not agree on level of detail • System engineering not ready in customer opinion – poor CPAR

  20. COMMAND PLAN CLIMB RATE COMMAND PLAN AIRSPEED COMMAND PLAN ALTITUDE COMMAND PLAN COURSE DISPLAY CLIMB RATE PERFORM VEHICLE FLIGHT MODE DISPLAY AIRSPEED DISPLAY ALTITUDE DISPLAY COURSE SELECT CLIMB RATE SELECT AIRSPEED SELECT ALTITUDE SELECT COURSE COMMAND FLIGHT MODE SELECT VECTOR FLIGHT MODE SELECT MISSION PLAN FLIGHT MODE COMMAND FLIGHT PARAMETERS PERFORM MISSION PLAN FLIGHT MODE PERFORM VECTOR FLIGHT MODE DISPLAY PLAN COURSE DISPLAY PLAN AIRSPEED DISPLAY PLAN ALTITUDE DISPLAY PLAN CLIMB RATE Why We Couldn’t Get to CDR • Airspeed, altitude, course are independent of mode • Built separately for each mode. • This model was given to different developers • Each model was implementeddifferently • Very difficult to integrate!!! • Not clear how this function fits into the entire system

  21. PERFORM VEHICLE FLIGHT MODES PERFORM PREFLIGHT VEHICLE PERFORM VEHICLE MISSION PLANNING PERFORM LAUNCH VEHICLE PERFORM VEHICLE COMMAND & CONTROL Completing Design: Hierarchical Functional Decomposition PERFORM START VEHICLE SESSION Group Common Functionality And Identify Dependencies

  22. PERFORM VEHICLE FLIGHT CONTROL PERFORM MISSION PLAN FLIGHT MODE SELECT MISSION PLAN FLIGHT MODE PERFORM VEHICLE FLIGHT MODE COMMAND FLIGHT MODE DISPLAY FLIGHT MODE PERFORM VECTOR FLIGHT MODE SELECT VECTOR FLIGHT MODE Completing Design: Consolidate Common Functionality

  23. DISPLAY FLIGHT PARAMETERS SELECT CLIMB RATE PERFORM VEHICLE FLIGHT CONTROL SELECT AIRSPEED SELECT ALTITUDE SELECT COURSE COMMAND FLIGHT PARAMETERS DISPLAY CLIMB RATE DISPLAY AIRSPEED DISPLAY ALTITUDE DISPLAY COURSE Completing Design: Detail-level Use Case Definition Each function can be linked to a hardware component Only building one vehicle control module for multiple flight modes

  24. Completing the Design: ROA Solution • Used the minimum set of artifacts • Requirements - what we have to do • Top-level model (glue) – how we are going to do it • Interface diagram – who we need to interface to • State diagram – when things are getting done • Sequence diagram – how the system behaves • ROA Design was implemented in MS Office framework • Used Word, PowerPoint, Excel, and Access • Files were under CM • Accessible by management, engineering and customer Everyone Working to the Same Goal, Focused on Objectives

  25. Completing Design: Results • Artifacts became management tool • Earned-value was related to each capability in TLM • Mgmt supported engineers when they had insight to progress • Customer gained confidence as they had insight to design • Successful CDR • Customer engineers defended our design to their management • Highest CPAR rating • Reduced build and test schedule because architecture was well-conceived and communicated! Next Phase of Program was Bid and Won Using Architecture Artifacts to Show Enhancements!

  26. Forensic Engineering: Challenges • Approached customer on ROA use • Customer liked methodology but wanted to see it applied to real problem • Safehold Software flagged as concern • Contractor identified software robustness as potential issue • Concerned about algorithm “sinkholes” • Recommended full-scale redevelopment • Software had no documented design • Algorithm document existed • No knowledge of functional dependencies • Large Red Team assigned to study problem

  27. Forensic Engineering: ROA Solution • ROA solution effort was two-pronged • Build top-level model (TLM) • Link functions to requirements • Done as a parallel effort to Red team • TLM identified functional interfaces as the software was built • ROA team scoured requirements and linked TLM elements to relevant requirements Developed an Illustration of the Software Design

  28. Forensic Engineering: Results • No need to redesign software • Some areas of improvement identified • Identified six requirements not being met by design • Red team came to same conclusion • Relied on software and hardware-in-the-loop testing • Thorough, time-consuming, but not very broad • Delivered collection of opinions – no artifacts ROA: 1.5 FTEs, 6 Weeks, Design Artifacts Red Team: ~15 FTEs, 4 Months, Opinion

  29. Future • Develop ROA cookbook • Provide a standards-based means of publishing methodology • Software tool • Develop MS Office-based toolset allowing ROA development • Develop interfaces to external systems • Software development • Earned Value Management System • Toolsets to accompany established requirements software

  30. Summary • ROA enables the system engineer to become the system designer • ROA provides a means of documenting system design • ROA models lead to good implementation

More Related