Architectural Risk Analysis Research Articles

The Top 10 Risks of Machine Learning Security

Our recent architectural risk analysis of machine learning systems identified 78 particular risks associated with nine specific components found in most machine learning systems. In this article, we describe and discuss the 10 most important security risks of those 78.

Software Security

Summary form only given. Software security has come a long way in the last few years, but we've really only just begun. I will present a detailed approach to getting past theory and putting software security into practice. The three pillars of software security are applied risk management, software security best practices (which I call touchpoints), and knowledge. By describing a manageably small set of touchpoints based around the software artifacts that you already produce, I avoid religious warfare over process and get on with the business of software security. That means you can adopt the touchpoints without radically changing the way you work. The touchpoints I will describe include: code review using static analysis tools; architectural risk analysis; penetration testing; security testing; abuse case development; and security requirements. Like the yin and the yang, software security requires a careful balance-attack and defense, exploiting and designing, breaking and building-bound into a coherent package. Create your own Security Development Lifecycle by enhancing your existing software development lifecycle with the touchpoints

Architectural Risk Analysis of Software Systems Based on Security Patterns

The importance of software security has been profound, since most attacks to software systems are based on vulnerabilities caused by poorly designed and developed software. Furthermore, the enforcement of security in software systems at the design phase can reduce the high cost and effort associated with the introduction of security during implementation. For this purpose, security patterns that offer security at the architectural level have been proposed in analogy to the well-known design patterns. The main goal of this paper is to perform risk analysis of software systems based on the security patterns that they contain. The first step is to determine to what extent specific security patterns shield from known attacks. This information is fed to a mathematical model based on the fuzzy-set theory and fuzzy fault trees in order to compute the risk for each category of attacks. The whole process has been automated using a methodology that extracts the risk of a software system by reading the class diagram of the system under study.

Static analysis for security

All software projects are guaranteed to have one artifact in common $source code. Together with architectural risk analysis, code review for security ranks very high on the list of software security best practices. We look at how to automate source-code security analysis with static analysis tools.