AI's Role in Network Efficiency and Observability

Recent discussions among industry leaders indicate that Artificial Intelligence (AI) is significantly impacting network operations and the expectations surrounding infrastructure performance. Experts from Carahsoft, AvidThink, and Aviz Networks explored the role of AI-powered observability, open networking through SONiC, and automation in creating more efficient network solutions.

Panelists

• Scott F Robohn – Consulting CTO, Carahsoft
• Roy Chua – Principal, AvidThink
• Random Access Memory (RAM) Mohan Hariprasad – Technical Product Marketing Manager, Aviz Networks

RAM Mohan Hariprasad began by addressing the current rapid evolution of networking. He noted the critical need for networks to scale in response to advancements in Graphics Processing Units (GPUs), Data Processing Units (DPUs), and the increasing demands of Generative AI (GenAI).

As new compute hardware continues to emerge, it is essential for networking capabilities to adapt at both hardware and operational levels.

Roy Chua highlighted that the changes extend beyond technology, emphasizing the importance of operational flexibility and intelligence within the infrastructure.

Observability and AI Integration

Hariprasad explained that traditional observability frameworks are becoming insufficient. Aviz is moving away from a “click-ops” model by utilizing large language models (LLMs) for enhanced real-time observability.

This new approach allows users to engage with AI agents to receive contextualized insights directly, reducing the need for navigation through multiple dashboards.

Such AI-driven systems are designed to improve the speed at which insights can be obtained, delivering a cohesive view across diverse vendor environments.

Vendor Flexibility and Cost Management

The discussion underscored a shared desire among enterprises for flexibility in their infrastructure choices without incurring additional costs.

Companies seek to adapt to new hardware and increased speeds while maintaining operational efficiency.

With a standardized operational layer offered by Aviz ONES, organizations can select upgrade paths according to their requirements, avoiding vendor lock-in and unexpected expenditures.

Standardization through SONiC

Robohn portrayed the evolution towards open networking as a continuation of the Linux narrative, asserting that SONiC fulfills a similar role for switches as Linux does for servers.

Hariprasad added that SONiC provides a robust standard, enhanced by Aviz through observability and automation, bridging innovation with production readiness.

Chua noted that the operational success of SONiC within hyperscalers instills confidence in its reliability for enterprise use.

AI-Native Networking Future

The panel concluded with insights into the future of AI-native networking. This involves developing scalable, latency-aware infrastructures to accommodate AI and Machine Learning (ML) workloads.

• Networks designed for AI must adapt to manage, troubleshoot, and optimize operations through AI.
• The next step involves standardizing data across various hardware and operating systems to facilitate effective network management through AI.

With these initiatives, Aviz aims to establish a unified AI framework that provides centralized intelligence in complex environments.

As Hariprasad concluded, the focus is not solely on speed, but also on creating smart infrastructures that learn and adapt over time.

In summary, the panelists discussed the current trends and forward-looking strategies in AI and network observability, aiming to highlight how organizations can effectively navigate the evolving landscape.