Virtuozzo, a leader in infrastructure software for AI and the cloud, announced Virtuozzo Infrastructure System (V/IS) version 8.0. The release updates NVIDIA support for GPU workloads, introduces lite deployments starting with one server and local storage, and brings its built-in VMware migration tool into general availability.
“Organizations need the flexibility to pursue new AI opportunities while improving the economics of their overall infrastructure,” said Kurt Daniel, CEO of Virtuozzo. “With updated NVIDIA support and a smaller deployment footprint, Virtuozzo helps them bring GPU-powered services to more customers and locations. Our focus is on making cloud and AI easier to deliver, more efficient to operate, and profitable to grow.”
V/IS version 8.0 supports NVIDIA R595 and vGPU v20, enabling dedicated and shared GPU resources on the same host. Providers can run AI alongside traditional workloads while helping customers control data location and access.
Also Read: HighByte and Databricks Partner to Make Industrial Data More Accessible for AI Applications
This release introduces a lite deployment option for environments needing only one or two servers. Compute, management, and storage run together using local disks instead of shared storage, reducing infrastructure investment for smaller customers, branch offices, and factory locations.
Providers and channel partners can use the new option to offer:
- Dedicated customer nodes for workloads requiring separate hardware
- Managed infrastructure on customer premises for applications and data that need to remain local
- Deployments for smaller businesses sized for current needs, with an expansion path as demand grows
Each independent deployment includes its own admin panel, monitoring, and alerting. Configurable system reservations help reduce the resources allocated to system services, leaving more capacity for customer workloads.
Lite deployments use standard V/IS licensing, with per-core and per-TB logical storage pricing. A one-server configuration can deliver approximately 53% lower total cost of ownership than VMware.
Nodes with local storage can be added under common cluster management as customers need more capacity. When customers require automatic VM failover, providers can introduce a storage network and shared V/Storage across at least three nodes to form a standard high-availability cluster.
Additional improvements in the release include generally available migration from VMware vSphere 8 without VDDK, a new onboarding wizard, a RHEL 10 operating system, faster object storage, and greater control over API access.
SOURCE: Globenewswire