“NeuralMesh enabled us to deploy composable storage that delivers secure multi-tenancy across our AI cloud infrastructure.”

Philip Lin
CEO, YTL AI Cloud

Tenant Density Without the
Trade-Offs.

FAQ

Your Questions, Answered

NeuralMesh runs multiple fully isolated tenants on a single cluster — each with their own private network, performance controls, security policies, and administrator. It also supports composable clusters, delivering dedicated hardware per tenant. Both models run on the same platform from one control plane.

Tenants are first-class entities inside the cluster. Each gets its own network space — private VLANs and IP ranges — plus its own capacity quota, QoS ceilings, security policies, and administrator role. Storage capacity and compute are shared dynamically. Isolation is enforced at the network and data plane, not just software policy.

Infrastructure cost drops as tenant count grows — shared storage capacity and compute convert every dollar of hardware to productive capacity. Your anchor tenants get dedicated hardware. Smaller tenants get logical isolation on shared infrastructure. You serve both profitably from one platform. New tenants come online in minutes with no hardware reconfiguration and no maintenance windows.

No — both models are fully supported and complementary. Composable clusters suit large tenants requiring physical isolation. Native multitenancy is optimized for high tenant density on shared infrastructure. Most service providers use both. Running both from one control plane is a capability most AI storage systems can’t offer.

Three things set it apart. Network spaces enforce isolation at the data plane — cross-tenant requests are rejected before they reach data, with overlapping IP ranges supported. Elastic resource sharing means compute and capacity flex dynamically with no idle reserved capacity. The dual administrator model keeps cluster and tenant admins fully separated by design.

Resources

Dive Deeper

Build a Better AI Cloud

See how NeuralMesh helps AI cloud providers serve more tenants, at lower cost, without adding operational overhead.