Software Dependability Modeling
Lead Personnel: TBD
Objectives: Develop a "value-based"
model for understanding and evaluating software dependability
alternatives and the tradeoffs among them, together with methods
and tools for evaluating and improving dependability at both design-
and run-time. Such model makes explicit the idea that a given
system dependability must be evaluated, not in absolute terms,
but in terms of the value that it provides the stakeholders.
Approach: Define dependability as a composite
property that integrates such attributes as availability, reliability,
safety, security, survivability and maintainability. Develop an
Information Dependability Attribute Value Estimation (iDAVE) model
for reasoning about the return on investment (ROI) of software
dependability investments. It integrates cost estimating relationships
(CER's) from the Constructive Cost Model COCOMO® II; dependability
estimating relationships (DER's) from the Constructive Quality
Model COQUALMO; and value estimating relationships (VER's)
supplied by the system's stakeholders. The other related models
such as quality attribute requirement conflict calculation (QARCC),
and software defect dynamics modeling will be added into the model
as appropriate. The assumption made by this model is:
- The iDAVE model's initial dependability value estimating
relationships (VER's) assume that a baseline business case
analysis has been performed for various components of value
(profit, customer satisfaction, on-time performance) as a function
of the time-phased information processing capabilities at nominal
dependability attribute levels.
Use the baseline dependability level
as the start point and increase the dependability investment level
from the baseline dependability level. Then analyze the
higher-payoff and lower-payoff regions of a project's dependability
investment. Develop a tool helps user to do such analysis.
Please click iDAVE to download the initial
spreadsheet version of the iDAVE Tool.
- D. Port, L. Huang, "Software Dependability Risks and
the Insurance Process," The 5th
International Workshop on Economics-Driven Software Engineering
Research, May 2003.
- L. Huang, B. Boehm, "Reasoning about the Value of Software
Dependability: the iDAVE Model," to appear on Supplementary
Proceedings 14th. IEEE International
Symposium on Software Reliability Engineering (ISSRE),
Further calibrate iDAVE to JPL experience, and to experimentally
use the model to evaluate the relative ROI of alternative combinations
of dependability technologies to planetary mission outcomes. Develop
and calibrate domain-oriented DER's and VER's for additional
dependability attributes such as security, safety, accuracy, and
Develop MDS-Based Extensions of the USC Information Dependability
Attribute Value Estimation (iDAVE) Model.