AI Needs Edge-to-Cloud Access

In a recent interview, Chris Mellor engaged with Jim Liddle, Nasuni’s Chief Innovation Officer for Data Intelligence and AI. The conversation delved into crucial elements of AI infrastructure including edge AI, data resilience, and the implications of malware in AI environments, highlighting Nasuni’s strategic approach in these areas.

The Landscape of AI and Data Storage

Nasuni has positioned itself as a pioneer in AI, even as the industry increasingly gravitates towards companies like Nvidia for AI training and inferencing. Liddle pointed out that while many enterprises are embracing AI, less than 5 percent are genuinely leveraging AI for training due to the considerable costs involved. For many companies, he asserted, the focus should be on deriving the best value from their existing domain information rather than engaging in expensive training exercises.

Importance of Edge-to-Cloud Integration

Emphasizing a hybrid model, Liddle stated that Nasuni operates on an edge-to-cloud architecture. This paradigm facilitates seamless data movement from various locations to the cloud, which he argued is critical since a significant portion of employee activity occurs at the edge. With myriad data points spread across various sites, Nasuni’s architecture enables efficient data access without the burden of complex migrations.

A Unified Global Namespace

Central to Nasuni’s strategy is the notion of a global namespace, which ensures that all enterprise data remains accessible regardless of its physical location. Liddle explained that having a single source of truth for data enhances contextual accuracy, vital for AI applications that require comprehensive and relevant data inputs.

Navigating AI Data Retrieval and Inference

While discussing data retrieval, Liddle highlighted that organizations often need to build on existing AI models without the need to retrain them from scratch. Instead, he noted, the emphasis should be on retrieval-augmented generation (RAG) or a nuanced approach that utilizes existing models in conjunction with proprietary data.

The Role of AI Data Resilience

Liddle raised an important consideration regarding the resilience of AI data infrastructures in the face of growing cybersecurity threats, particularly ransomware. He noted that AI systems, while beneficial, can also become targets. As businesses increasingly rely on automated systems and AI agents, ensuring data resilience becomes paramount. Nasuni employs sophisticated snapshot technology to enhance data recovery and mitigate risks associated with ransomware attacks.

Tracking AI Agent Behavior

Furthermore, the dialogue touched upon the need to monitor AI agents’ behavior. As enterprises integrate AI more into their workflows, being aware of agents’ operations and the potential for compromised behavior has become crucial. Liddle indicated that while deploying agents provides significant benefits, protecting these systems against vulnerabilities is essential in maintaining enterprise integrity.

Conclusion

Overall, the insights shared by Liddle showcase Nasuni’s commitment to supporting the evolving landscape of AI through robust storage solutions that prioritize edge-to-cloud access, data resilience, and a unified architecture for information retrieval. As the need for sophisticated data handling grows alongside AI applications, the importance of these features cannot be overstated.