<div>In 2019, cyber security is considered one of the most significant threats to the global economy and national security. Top U.S. agencies have acknowledged this fact, and provided direction regarding strategic priorities and future initiatives within the domain. However, there is still a lack of basic understanding of factors that impact complexity, scope, and effectiveness of cyber defense efforts. Computer security incident response is the short-term process of detecting, identifying, mitigating, and resolving a potential security threat to a network. These activities are typically conducted in computer security incident response teams (CSIRTs) comprised of human analysts that are organized into hierarchical tiers and work closely with many different computational tools and programs. Despite the fact that CSIRTs often provide the first line of defense to a network, there is currently a substantial global skills shortage of analysts to fill open positions. Research and development efforts from educational and technological perspectives have been independently ineffective at addressing this shortage due to time lags in meeting demand and associated costs. This dissertation explored how to combine the two approaches by considering how human-centered research can inform development of computational solutions toward augmenting human analyst capabilities. The larger goal of combining these approaches is to effectively complement human expertise with technological capability to alleviate pressures from the skills shortage.</div><div><br></div><div>Insights and design recommendations for hybrid systems to advance the current state of security automation were developed through three studies. The first study was an ethnographic field study which focused on collecting and analyzing contextual data from three diverse CSIRTs from different sectors; the scope extended beyond individual incident response tasks to include aspects of organization and information sharing within teams. Analysis revealed larger design implications regarding collaboration and coordination in different team environments, as well as considerations about usefulness and adoption of automation. The second study was a cognitive task analysis with CSIR experts with diverse backgrounds; the interviews focused on expertise requirements for information sharing tasks in CSIRTs. Outputs utilized a dimensional expertise construct to identify and prioritize potential expertise areas for augmentation with automated tools and features. Study 3 included a market analysis of current automation platforms based on the expertise areas identified in Study 2, and used Systems Engineering methodologies to develop concepts and functional architectures for future system (and feature) development.</div><div><br></div><div>Findings of all three studies support future directions for hybrid automation development in CSIR by identifying social and organizational factors beyond traditional tool design in security that supports human-systems integration. Additionally, this dissertation delivered functional considerations for automated technology that can augment human capabilities in incident response; these functions support better information sharing between humans and between humans and technological systems. By pursuing human-systems integration in CSIR, research can help alleviate the skills shortage by identifying where automation can dynamically assist with information sharing and expertise development. Future research can expand upon the expertise framework developed for CSIR and extend the application of proposed augmenting functions in other domains.</div>
| Identifer | oai:union.ndltd.org:purdue.edu/oai:figshare.com:article/10014803 |
| Date | 30 October 2019 |
| Creators | Megan M Nyre-Yu (7525319) |
| Source Sets | Purdue University |
| Detected Language | English |
| Type | Text, Thesis |
| Rights | CC BY 4.0 |
| Relation | https://figshare.com/articles/Determining_System_Requirements_for_Human-Machine_Integration_in_Cyber_Security_Incident_Response/10014803 |
Page generated in 0.0019 seconds