Uniting academic achievements on performance analysis with industrial needs

Abstract

In our mission to advance innovation by industrial adoption of academic results, we perform many projects with high-tech industries. Favoring formal methods, we observe a gap between industrial needs in performance modeling and the analysis capabilities of formal methods for this goal. After clarifying this gap, we highlight some relevant deficiencies for state-of-the-art quantitative analysis techniques (focusing on model checking and simulation). As an ingredient to bridging the gap, we propose to unite domain-specific industrial contexts with academic performance approaches through Domain Specific Languages (DSLs). We illustrate our vision with examples from different high-tech industries and discuss lessons learned from the migration process of adopting it.