The industry-leading fabric for Cloud Networking and much more
The Pluribus Unified Cloud Fabric – the next generation of the Adaptive Cloud Fabric – brings the benefits of hyperscale agility, simplicity and availability to data centers, unified cloud networks and many other networking applications.
The Unified Cloud Fabric unites networks across multiple dimensions: switches and servers, overlay and underlay networks, private and public clouds. It incorporates comprehensive automation, distributed security, and pervasive visibility, runs on diverse open network hardware platforms, and works with any workload or virtualization environment.
Unified Cloud Fabric technology is also powerful and flexible enough to go beyond the cloud and solve diverse networking problems, from layer 1 test lab automation to service provider metro networking.
Integrated with NVIDIA DPUs
The Unified Cloud Fabric architecture delivers a holistic compute virtualization fabric across any workload environment (including ESXi, Hyper-V, Xen, KVM, bare metal, and Kubernetes). Providing a zero-trust administration model between compute and network, and radically simplifying the networking stack running on the server OS with better overall performance and lower TCO.
The Pluribus Netvisor ONE network operating system (OS) integrates with the NVIDIA BlueField data processing unit (DPU) hardware architecture to:
- Unify and simplify network segmentation with a common SDN overlay for multiple compute environments.
- Radically simplify the deployment of micro-segmentation policies (across different hypervisor environments) by eliminating changes to the physical network and traffic hair-pinning.
- Deliver pervasive visibility of traffic flows to/from any VM, without any blind-spots and without the overhead of network tapping.
- Leverage the DPU hardware architecture to enable a clean, zero-trust demarcation between the server and the NOS delivering the network/security services accelerated by the DPU.
- Boost networking and security performance while offloading server CPUs with NVIDIA BlueField DPU chip acceleration.
Features and Benefits
Cloud Operating Model
The Unified Cloud Fabric brings an agile, simplified public cloud operating model to private clouds and on-premises data centers, enabling network operations teams to increase operational efficiency and deliver new services in seconds.One-touch Fabric-wide Provisioning
A single command programs the entire fabric (for example add a fabric-wide VLAN or distributed VRF or shut down a port on all nodes), and every command is atomic with unlimited rollback to quickly revert to a good state, dramatically reducing manual provisioning time and the potential for errors.Rich L2/L3 Overlay Services
A wide variety of highly scalable layer 2 and layer 3 VPN services provides a robust toolkit for network architects and engineers to solve virtually any networking problem, and every fabric service benefits from one-touch provisioning and strong multi-tenant segmentation and security.Network Segmentation / Microsegmentation
Network segmentation, or deep slicing, provides isolation for specific traffic types like IoT as well as supporting multi-tenant services across both switches and servers. When deployed in server-based DPUs, the fabric provides even finer-grained microsegmentation with distributed firewalls at the individual application level, enabling zero-trust security.Pervasive Visibility
Integrated FlowTracker telemetry captures all key parameters of every TCP, UDP, ICMP, DNS, DHCP flow between all endpoints connected to the fabric for unparalleled visibility and analytics. The KubeTracker™ fabric service enables visibility into East-West traffic between pods/containers inside a Kubernetes cluster, empowering NetOps teams with cloud-native application visibility. Learn moreScalable, Flexible, Distributed Architecture
Unified Cloud Fabrics scale simply to thousands of network nodes, stretch seamlessly across highly distributed data center and edge locations, and support flexible and arbitrary topologies including leaf-spine, hub-and-spoke, ring or mesh.Easy to Deploy, Easy to Operate
With a comprehensive set of features and capabilities all pre-integrated, pre-tested and inherently automated, the Unified Cloud Fabric delivers networking solutions that just work, right “out of the box.”Cloud Operating Model
The Unified Cloud Fabric brings an agile, simplified public cloud operating model to private clouds and on-premises data centers, enabling network operations teams to increase operational efficiency and deliver new services in seconds.
One-touch Fabric-wide Provisioning
A single command programs the entire fabric (for example add a fabric-wide VLAN or distributed VRF or shut down a port on all nodes), and every command is atomic with unlimited rollback to quickly revert to a good state, dramatically reducing manual provisioning time and the potential for errors.
Rich L2/L3 Overlay Services
A wide variety of highly scalable layer 2 and layer 3 VPN services provides a robust toolkit for network architects and engineers to solve virtually any networking problem, and every fabric service benefits from one-touch provisioning and strong multi-tenant segmentation and security.
Network Segmentation / Microsegmentation
Network segmentation, or deep slicing, provides isolation for specific traffic types like IoT as well as supporting multi-tenant services across both switches and servers. When deployed in server-based DPUs, the fabric provides even finer-grained microsegmentation with distributed firewalls at the individual application level, enabling zero-trust security.
Pervasive Visibility
Integrated FlowTracker telemetry captures all key parameters of every TCP, UDP, ICMP, DNS, DHCP flow between all endpoints connected to the fabric for unparalleled visibility and analytics. The KubeTracker™ fabric service enables visibility into East-West traffic between pods/containers inside a Kubernetes cluster, empowering NetOps teams with cloud-native application visibility. Learn more
Scalable, Flexible, Distributed Architecture
Unified Cloud Fabrics scale simply to thousands of network nodes, stretch seamlessly across highly distributed data center and edge locations, and support flexible and arbitrary topologies including leaf-spine, hub-and-spoke, ring or mesh.
Easy to Deploy, Easy to Operate
With a comprehensive set of features and capabilities all pre-integrated, pre-tested and inherently automated, the Unified Cloud Fabric delivers networking solutions that just work, right “out of the box.
How it works
The Unified Cloud Fabric is a distributed application that runs in Pluribus-enabled network devices.
Built into Netvisor ONE
The Unified Cloud Fabric is built into the deployment-proven Linux-based Netvisor® ONE network operating system (OS), so it runs as a distributed network application on every network edge device, including switches and DPUs. Robust, fault-tolerant distributed database technology ensures that every device has a complete and consistent view of the entire network fabric, and that the fabric is highly resilient and available.
Controllerless SDN Automation
The Unified Cloud Fabric creates a distributed software-defined networking (SDN) control plane, eliminating the need for external controllers and lowering cost and complexity compared to controller-based SDN solutions. This SDN-enabled, protocol-free control plane automates and integrates both the physical data center network underlay and virtualized network overlay into a unified fabric.
