#Japan #Tokyo #NVIDIA #AI Security #Check Point

Introduction

Check Point Software Technologies, a global leader in cybersecurity, has unveiled its latest development—the AI Factory Security Blueprint. This reference architecture provides a comprehensive framework for securing private AI infrastructures throughout every layer, from hardware to applications. Notably, the blueprint harnesses Check Point's industry-leading firewall technology and AI security measures, relying on NVIDIA’s BlueField data processing capabilities to foster a ‘Security by Design’ environment for AI factories and data centers.

Importance of AI Infrastructure Protection

According to Nataly Kremer, Check Point’s Chief Product Officer, AI infrastructure ranks as one of the most critical and vulnerable assets for organizations. The AI Factory Security Blueprint aims to bolster businesses by safeguarding their investments in AI infrastructure. Kremer emphasizes that this initiative supports built-in security from the design phase, rather than retrofitting security measures after deployment.

AI data centers are increasingly becoming essential yet risky components of enterprise infrastructure. Companies are creating private AI environments to safeguard intellectual property, adhere to data sovereignty requirements, and cut costs associated with public cloud services. This approach demands significant investments in assets such as GPU clusters, training pipelines, and proprietary models. However, the rapid growth of AI technology often outpaces the establishment of necessary security architectures.

Emerging Threats and Security Needs

AI computing environments differ from traditional data centers, integrating high-performance GPU clusters, distributed training pipelines, and extensive data lakes. This evolution has created new attack surfaces that previous security tools are ill-equipped to handle. Threat vectors include everything from poisoning attacks on training data and model theft to lateral movements within Kubernetes namespaces and supply chain attacks exploiting open-source dependencies.

Four-Layer Protection Defined in the Blueprint

The AI Factory Security Blueprint establishes a four-layer protection strategy:

1. Boundary Layer: Check Point’s Maestro Hyperscale Firewall offers Zero Trust network access, virtual security group segmentation, and scalable policy application. This system addresses north-south traffic from external users, internet sites, and corporate networks

2. Application and LLM Layer: The AI Agent Security protects inference APIs and LLM endpoints from prompt injection, data theft, malicious queries, and API misuse. Notably, this goes beyond what traditional web application firewalls (WAF) can cover, integrating seamlessly across cloud, virtual, and appliance formats.

3. AI Infrastructure Layer: In collaboration with NVIDIA, Check Point integrates firewall and threat defense capabilities directly into NVIDIA’s BlueField Data Processing Units (DPUs) via the DOCA software platform. This ensures that high-performance AI prompt defenses and traffic inspection occur without consuming CPU/GPU cycles.

4. Workload and Container Layer: Check Point collaborates with third-party micro-segmentation solutions to implement micro-segmentation and control east-west traffic within Kubernetes clusters, preventing lateral movements in inference namespaces.

Compliance with Security Frameworks

This blueprint adheres to the CISA principles which advocate for 'secure by design' AI systems. This principle promotes embedding security in the fabric, hardware, and orchestration layers from the outset, rather than tacking it on afterward. Check Point's architecture insists on upholding Zero Trust practices in every interaction, ensuring all users, API calls, and service requests are authenticated and continuously validated.

Moreover, the blueprint complies with various governance frameworks including the NIST AI Risk Management Framework and Gartner’s AI Trust, Risk, and Security Management framework, which helps organizations align with regulatory requirements across AI legislation like the EU AI Act and the GDPR.

Conclusion

In a rapidly changing technological landscape, the AI Factory Security Blueprint by Check Point serves as a pivotal resource for organizations seeking to protect their AI infrastructures. By embedding security deep within the design, companies can better withstand the evolving threats facing their vital data and intellectual property. For further details on the AI Factory Security Blueprint, please visit Check Point's official page.

Check Point Software Technologies protects over 100,000 organizations globally, advocating for a secure AI transformation in the complex digital environment. Through a prevention-focused strategy and integrated architecture, Check Point significantly enhances the security posture of hybrid networks, multi-cloud environments, and AI systems.