A Systems Perspective on Technical Debt

Abstract

This paper explores applications of the Technical Debt metaphor, which originated in the agile software community, at a systems level. We present a summary of current understanding of the technical debt concept and propose a broader application of the approach at a system level. The technical debt metaphor provides a method to identify cases when design or implementation decisions are being made for expediency and assess their long term impact on productivity and risk. In addition, technical debt management offers a logical approach to tracking expedient decisions and determining when it makes sense to carry the debt and when it s better to pay it off by updating designed and retrofitting units. In this paper, we describe various kinds of technical debt that may be present in a system baseline and consider how the presence of debt impacts further evolution of the system. Types of debt include architectural debt, configuration debt, testing debt and deprecation debt among others. We also present use case studies that illustrate how technical debt assessment and management can have a positive impact on system realization. Cases include moving a baseline from R&D to operations, reusing a baseline in a new context, and planning a technical refresh. We discuss methods of managing technical debt in a baseline including identification, quantification, tracking and payoff.