GuideArch: Guiding the exploration of architectural solution space under uncertainty

Abstract

Abstract-A system's early architectural decisions impact its properties (e.g., scalability, dependability) as well as stakeholder concerns (e.g., cost, time to delivery). Choices made early on are both difficult and costly to change, and thus it is paramount that the engineer gets them "right". This leads to a paradox, as in early design, the engineer is often forced to make these decisions under uncertainty, i.e., not knowing the precise impact of those decisions on the various concerns. How could the engineer make the "right" choices in such circumstances? This is precisely the question we have tackled in this paper. We present GuideArch, a framework aimed at quantitative exploration of the architectural solution space under uncertainty. It provides techniques founded on fuzzy math that help the engineer with making informed decisions.

Uncertainty, risk, and information value in software requirements and architecture

“…They also defined models for computing the impact of options on the goals and documented uncertainty about model parameters using three point estimates, a method commonly used by engineers and project managers that consists in eliciting a pessimistic, most likely, and optimistic value for each model parameter. They then applied a fuzzy-logic based approach, called GuideArch, to support design decisions under uncertainty [21].…”

“…Previous work dealing with uncertainty in early requirements and architecture decisions [21,42,49,62] suffers important limitations: they use unreliable methods for eliciting uncertainties (some confuse group consensus with certainty); they tend to evaluate alternatives against vague, unfalsifiable criteria; they provide no information about the risks that accompany uncertainty; and they provide no support for assessing to what extent obtaining additional information before making a decision could reduce these risks.…”

“…3. We introduce the concept of Pareto-optimal strip, a generalisation of a Pareto-optimal front, designed to resist modelling and measurement errors present in multi-objective decisoin problem under uncertainty (Section 3.5); We have developed a tool supporting our approach and have applied it to data from a real system from the literature [21]. Our tool and all models discussed in this paper are available at www.cs.…”

Comparison of exact and approximate multi-objective optimization for software product lines

“…ERS [10] is an Emergency Response System presented with seven quality attributes and objectives: battery usage, response time, reliability, ramp-up time, cost, deployment time and development time. The model was built based on expert judgment and used to explore uncertainty in the early architectural design, it included lower, upper bounds and middle values for the contribution of each feature to the quality attributes.…”

Uncertainty handling in goal-driven self-optimization – Limiting the negative effect on adaptation

“…This study used a probabilistic technique to tackle goal satisfaction uncertainty at design time, while we focus on runtime handling of contribution uncertainty, preference uncertainty and effect uncertainty. Esfahani et al (2013) presented GuideArch to quantitatively guide the exploration of architectural solution space, including ranking the architectures, finding the optimal, and identifying the critical decisions, under the uncertain impact of architectural alternatives on properties of interest. This study employs fuzzy math to represent and reason about uncertainty where ours uses a probability analysis to try to eliminate the contribution uncertainty.…”