8/18/2019 COCOMO Farah-Stapleton November-2015-DRAFT9 MFS http://slidepdf.com/reader/full/cocomo-farah-stapleton-november-2015-draft9-mfs 1/26 Resource Analysis Based On System Architecture Behavior Monica Farah-Stapleton Professor Ray Madachy Professor Mikhail Auguston Professor Kristin Giammarco November 2015
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Information Technology (IT) systems are large, challenged by the rate of change ofcommercial IT, and represent a significant investment in time and resources – Introduction of new system/capability results in unintended or unexpected
system/environment behaviors
– Operational and financial impacts often assessed after the fact
– Resourcing decisions and precise architectural descriptions of system/environment often
minimally related
Approach
• Leverage precise behavioral modeling using Monterey Phoenix (MP) to assess
architectural design decisions and their impacts prior to, during, and after deployment
•Relate architectural modeling to resourcing through analysis of behaviors andUnadjusted Function Point Analysis (FPA) counts, leveraging complexity and size
metrics, e.g. Data Element Type (DET) and File Type Referenced (FTR)
• Align activities with Systems Engineering Research Center (SERC) Ilities Tradespace
• Uniform Framework: Describe behaviors and interactions of the system AND environment using the sameframework
• Leverage Small Scope Hypothesis: Exhaustive search through all possible scenarios (up to the scope limit),expecting that most flaws in models can be demonstrated on small counterexamples
• Separation of System Interaction from System Behavior: Specify behavior of each system’s components
separately from interactions between those components
• Unadjusted Function Point is a unit of measurement to express the amount of
functionality in a system, and can be used to estimate system cost. Of specificinterest are the input/output activities of the system.
• MP architecture model is based on behavior modeling, providing a bridge
between the requirements and high level design, and is supportive of cost
estimates early in the design phase.
•The concept of an event in MP is an abstraction for activity within the system.It is rendered as a pseudo-code, appropriate for capturing the functional
aspects of requirements, and supportive of refinement.
• UFP can be identified in the MP architecture model as an interaction
abstraction (i.e. COORDINATE or SHARE ALL constructs).
• The structure and the complexity of interactions in MP provide a source forassigning weights contributing to the UFP.
• Since an MP model is precise and formal, FP metrics can be identified by
automated tools such as http://csse.usc.edu/tools/MP_COCOMO
• Collaboration with AFIT for a joint Intelligence, Surveillance and Reconnaissance
(ISR) mission application involving heterogeneous teams of autonomous and
cooperative agents.
• NPS will provide cost modeling expertise, tools and Monterey Phoenix (MP) modeling
support. A focus will be on translations between models/tools in MBSE, specifically
mapping architectural elements into cost model inputs.
• Approach – Develop a baseline operational and system architecture to capture a set of military scenarios.
– Transition the baseline architecture to the MP environment.
– Utilize the executable architecture modeling framework of MP to perform automated assertion
checking and find counterexamples of behavior that violate the expected system's correctness.• Operational scenarios will be cycled through the MP modeling process, whereby alternate events are
captured for each actor in each scenario. This will produce a superset of scenario variants from thebehavior models, suitable for input to tradespace analysis and cost models.
– Design and demonstrate an ISR UAV tradespace.
– Develop cost model interfaces for components of the architecture in order to evaluate cost
• Continue extending the scope and tradespace interoperability of cost models and tools
from previous phases.
• Cost modeling will engage domain experts for Delphi estimates, evolve baseline
definitions of cost driver parameters and rating scales for use in data collection, gather
empirical data and determine areas needing further research to account for differences
between estimated and actual costs.
– Prototype cost models and tools will be extended accordingly for piloting and refinement.• For tool interoperability we will integrate cost models in different ways with MBSE
architectural modeling approaches and as web services. We will also automate systems
and software risk advisors that operate in conjunction with the cost models.
• NPS will provide domain expertise for SysML cost model integration with Georgia
Tech and USC to add software cost model formulas and the risk assessment
capabilities. – This is also allied with Task 2 where we will assess Monterey Phoenix (MP) for automatically
providing cost information from architectural models. MP will extract software sizing cost
model inputs to compute costs, and we will assess mapping MP architectural elements into