Cisco updated its SD-WAN integration with Megaport for on-demand AI data center access.

By integrating its Catalyst SD-WAN platform with Megaport’s AI Exchange and Software-Defined Cloud Interconnect (SDCI), Cisco aims to help enterprises connect with AI data centers for low-latency, high-performance networking that can handle AI workloads.

Using Cisco’s Cloud OnRamp automation and multicloud optimization features, users can connect to AI workloads hosted by major public clouds and neoclouds, including GPU-as-a-service (GPUaaS) offerings for AI training and inference.

Underpinning the solution is an on-demand SD-WAN fabric designed for optimized routing and low latency while maintaining uniform security policies, enabling dynamic scaling and efficient workload distribution.

The network giant claimed users can expect faster deployment of AI connectivity, along with improved visibility into traffic patterns, performance, alongside security postures.

Users will also be able to scale bandwidth on demand and maintain sovereign control of their network paths, with service-level agreement (SLA)-backed connectivity between multiple cloud and AI ecosystems.

Cisco also positioned the integration as an opportunity for AI vendors and neocloud providers to create new AI-workload based revenue streams. By using the SD-WAN and Megaport framework, it claimed providers can simplify their edge routing architectures, and deliver high-performance connectivity to customers.

The offering builds on Megaport’s existing Cisco SD-WAN integration with its network-as-a-service (NaaS) offering.

It also offers an alternative to recent releases such as Google Cloud’s Cross-Site Interconnect (CSI), a managed data layer connectivity offering designed to simplify global WAN architectures, and an AI-driven WAN solution released by Huawei last month.