Hybrid cloud and artificial intelligence platform leader Nutanix and Netris, a pioneer in Network Automation, Abstraction, and Multitenancy , have announced a strategic engineering collaboration to construct a joint multitenant infrastructure solution purpose-built for neocloud operators and enterprise data centers.
The integrated stack combines Nutanix Cloud Platform’s compute orchestration and virtualization layer with the Netris NAAM platform. Specifically engineered to satisfy the requirements of NVIDIA Cloud Partner technical specifications, the design enables cloud providers and enterprise IT to quickly deploy secure, GPU-accelerated IaaS services in bare-metal and virtualized deployments.
Overcoming Multitenancy and Infrastructure Problems When Scaling AI Clouds
With the growth of enterprise interest in generative AI and high-performance computing, neocloud vendors and enterprise teams are now fast-moving into creating extensive GPU clusters. Yet the transformation of processing resources into commercially viable GPU-as-a-Service is challenging due to a number of technical issues, including tenant isolation and network orchestration.
The creation of enterprise-ready AI cloud requires the provision of soft and hard network isolation throughout all infrastructure layers, from virtual machines and Data Processing Units (DPUs) to physical switches. The manual setup of network configurations becomes a bottleneck for fast customer onboarding and revenue generation.
Also Read: Rafay Systems Partners with Lenovo AI Innovators to Drive Infrastructure Monetization Across Global Markets
By unifying compute provisioning with physical network orchestration, the Nutanix and Netris architecture removes these obstacles and enables tenants’ onboarding, strict data isolation among competing workloads, and the creation of cloud-like self-service management experience.
Some of the core capabilities and functionalities provided via the joint solution are:
Automatic Physical Fabric Configuration: Provides automatic configuration of the switching fabric and network services when allocating compute resources.
Bare Metal and Virtualized Network Consistency: Converts parameters of the Nutanix tenants into consistent network policies for both VMs and bare-metal GPU nodes.
Cloud Networking Constructs Access: Provides access to cloud networking constructs, including VPCs, Elastic IP addresses, NAT Gateways, Layer 4 Load Balancers via user-friendly web console, REST API, and Terraform providers.
Optimized AI Factory Network Fabrics: Provides optimized network fabric for AI applications, optimized for high throughput, North-South, East-West, and scale-up network traffic flows.
Hard and Soft Isolation Across the Stack: Extends the hardware-based tenant isolation from hypervisors to the physical network switches level for secure multi-tenant environment with shared GPU cluster.
“Neoclouds and NVIDIA Cloud Partners are at the forefront of the AI transformation, but building the cloud operating stack required to monetize GPU clusters at scale is remarkably complex,” said Ketan Shah, Vice President, Product Management, Nutanix. “By pairing Nutanix’s Agentic AI solution with the Netris NAAM platform, we will provide AI cloud operators with an out-of-the-box solution to accelerate their journey from metals to margins. We chose to partner with Netris, a leader in network automation, abstraction, and multitenancy, to give operators true self-service networking across both bare-metal and virtualized environments. This will enable our partners to deliver secure, multitenant IaaS environments quickly while maintaining the simplicity and enterprise reliability that Nutanix is known for.”
“AI infrastructure is scaling incredibly fast, and the network has to scale with it,” said Alex Saroyan, CEO and Co-Founder of Netris. “An AI cloud needs two layers working as one – Nutanix’s compute orchestration and virtualization, backed by deep operational leadership and experience, and Netris network automation, abstraction, and multitenancy. What makes this collaboration matter is that those layers stay continuously in sync, giving neoclouds and enterprises a single automated stack for secure, multitenant GPU services.”
“AI infrastructure providers are evolving beyond GPU capacity to deliver trusted inference environments, workload orchestration, agent lifecycle management, and managed AI services. The differentiator will be a cloud-like operating model that combines performance with control, isolation, and security at every layer of the stack.” Steve McDowell, Principal Analyst, NAND Research