# Software Updates Strategies: a Quantitative Evaluation against Advanced Persistent Threats

arXiv: https://arxiv.org/abs/2205.07759 · HTML: https://mlsanity.com/paper/2205.07759/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 16 May 2022.

## Built on by (most cited first, 3 of 3)

- [Will AI Make Cyber Swords or Shields: A few mathematical models of technological progress](https://mlsanity.com/paper/2207.13825/) (2022)
- [Large Language Models Are Unreliable for Cyber Threat Intelligence](https://mlsanity.com/paper/2503.23175/) (2025)
- [CVE-TTP KG: Knowledge Graph Linking Software Vulnerabilities to Attack Behaviors](https://mlsanity.com/paper/2606.31557/) (2026)

## Abstract (arXiv)

Software updates reduce the opportunity for exploitation. However, since updates can also introduce breaking changes, enterprises face the problem of balancing the need to secure software with updates with the need to support operations. We propose a methodology to quantitatively investigate the effectiveness of software updates strategies against attacks of Advanced Persistent Threats (APTs). We consider strategies where the vendor updates are the only limiting factors to cases in which enterprises delay updates from 1 to 7 months based on SANS data. Our manually curated dataset of APT attacks covers 86 APTs and 350 campaigns from 2008 to 2020. It includes information about attack vectors, exploited vulnerabilities (e.g. 0-days vs public vulnerabilities), and affected software and versions. Contrary to common belief, most APT campaigns employed publicly known vulnerabilities. If an enterprise could theoretically update as soon as an update is released, it would face lower odds of being compromised than those waiting one (4.9x) or three (9.1x) months. However, if attacked, it could still be compromised from 14% to 33% of the times. As in practice enterprises must do regression testing before applying an update, our major finding is that one could perform 12% of all possible updates restricting oneself only to versions fixing publicly known vulnerabilities without significant changes to the odds of being compromised compared to a company that updates for all versions.