Axians: AI and Automation Reshape Modern Data Centers

IT services company Axians has released an analysis examining the role of artificial intelligence (AI) and automation in the evolution of modern data centers. The publication highlights that increasing data volumes and new AI workloads place unprecedented demands on traditional data center infrastructures, requiring a rethink.

AI training phases consume significant resources, including throughput, energy, and cooling, differing from classical applications that generated predictable loads. Furthermore, data centers must transition from being mere infrastructures to strategic innovation platforms, ensuring data availability, security, and performance.

Axians emphasizes the advantages of Ethernet networks over previously used proprietary technologies. Modern Ethernet solutions offer comparable reliability and speed more cost-effectively and flexibly. This openness is crucial for meeting the demands of rapidly changing AI models.

The growing complexity of data centers increases the need for automation. Axians points to the Juniper Apstra solution, which simplifies the planning, deployment, and management of AI-compatible networks throughout their lifecycle. The solution enables virtual modeling, automated deployment, and continuous network status monitoring through Intent-Based Analytics.

The analysis also addresses the importance of data sovereignty, particularly for government bodies and regulated industries. Maintaining proprietary large language models (LLMs) within a data center offers full control over data and models, maximizes security, and reduces reliance on external cloud providers. Axians advises companies to first analyze their current network traffic patterns before implementing automated networks.