Risky Business #716 -- This ain't your grandma's cloud
Aug 8, 2023
auto_awesome
Brian Dye, Corelight's CEO, discusses cybersecurity news such as Tenable's criticism of Microsoft over Azure bugs, ransomware attacks on hospitals and schools, and Russia hacking Japan's defense networks. The podcast delves into cloud security challenges, tech controversies, cybersecurity laws, and network security event models, optimizing security operations with XDR solutions.
Criticism towards Microsoft for Azure bug fix delays and quality issues.
Ransomware attacks on hospitals and schools highlight cybersecurity vulnerabilities.
China hacking Japan's defense cyber networks emphasizes the need for improved cybersecurity measures.
Deep dives
Facial Recognition Technology: A Double-Edged Sword
The use of facial recognition technology by a unit of the Department of Homeland Security, aided by Clearview AI, has led to successful arrests of individuals involved in heinous activities, leveraging previously seized material. This showcases a positive application of the technology. However, a contrasting story highlights a pregnant woman in Detroit wrongfully arrested for carjacking due to a mistaken identity from a facial recognition system. This incident underscores the need for policy development to prevent misuse of the technology, especially considering documented cases of false charges primarily targeting black individuals.
Challenges in Hardware Security: Inception and Downfall Attacks
Recent attacks named Inception targeting AMD Zen CPUs and Downfall affecting Intel-based systems reveal challenges in ensuring hardware isolation and security. Speculative execution vulnerabilities like these demonstrate that hardware security mechanisms, such as SGX, have not yet delivered the expected resilience. The complexity of optimizing CPU performance while maintaining robust security proves to be a significant hurdle in achieving reliable hardware isolation.
Navigating the Contrasts of Technology Applications
The juxtaposition of successful and problematic uses of technology, such as facial recognition, illuminates the need for nuanced approaches. While applauding beneficial applications for apprehending criminals, like those involved in child exploitation, the erroneous arrest of individuals based on flawed facial recognition highlights the importance of balancing technology advancement with protective policies to mitigate misuse and discrimination in law enforcement practices.
Evolution of Security Analytics Architectures
Security analytics architectures have evolved over the years, with three main models emerging: a fully centralized data lake approach, the SOC triad model, and the XDR model. The data lake approach offers flexibility and effectiveness, especially with a large security engineering team, but can be costly. The SOC triad model serves as a starting point for organizations and provides meaningful detections, albeit with limitations in IR capabilities compared to the data lake approach. The XDR model blends the best of both worlds, offering cost effectiveness, ease of deployment, and some flexibility, although there may be constraints imposed by vendor-specific functionalities.
Considerations in Choosing Security Analytics Architectures
Organizations should evaluate their security analytics architectures based on factors such as the size of their security engineering team, the need for threat hunting capabilities, and the level of in-house detection engineering expertise. Large security engineering teams can make the fully centralized data lake approach more feasible and effective. The SOC triad model is suitable for organizations with budget constraints and basic security needs, while the XDR model offers a balance between cost effectiveness and flexibility, with ongoing advancements enhancing its capabilities.