Pluribus Adaptive Cloud fabric offers significant capabilities around automation and provisioning in VMWare environments. ONVL’s vPort database, which is one of the key components enabling its incredible awareness & tracking of endpoints across the fabric, is further enhanced through the integration of VMWare vCenter object metadata. These VMWare attributes (port-group, vnic, vmkernel, vmnic etc.) become searchable keywords within the vPort DB. At a high level, the Pluribus Fabric integration with VMWare provides:
- vSphere-aware distributed end-point directory (vPort database)
- vSphere-aware Insight Analytics, which complement vRealize Network Insight.
- Multicast auto-provisioning for vSAN transport + Fabric network provisioning based on vSAN cluster configuration
- NSX provisioning of L2 VTEP gateway on physical switches (ovsdb service container). Switch Hardware VTEP orchestration can be executed via NSX controller.
Detection & Provisioning
With Pluribus <> VMWare integration, One-Touch Automation around provisioning (for respective features) across Compute, Network & Storage, can be achieved via vCenter.
Layer 2: LAGs, vLAGs & VLANs
With regards to Auto-provisioning of LAGs/vLAGs , it will perform auto host link detection & LAG formation. **
With regards to Auto-provisioning of VLANs (from vCenter) on host facing ports:
- If the fabric is Layer 2, it will auto-provision VLANs on uplink ports
- If the fabric is Layer 3, it will auto-provision VNI and tunnels
- With VXLAN Extension, Populate L2 Networks created in vCenter, across (any VXLAN tunnel in) the Multi-DC fabric
As an example, Once a new workload + Port Group + VLAN is provisioned, it is automatically added to
- ToR switch to host links/LAG
- Fabric wide (inter-switch links)
- With VXLAN, across the tunnel into multiple DCs
VMs & Hosts
The integration allows the fabric to identify the switch port used for each VM, track movement of VMs from one host (ESXi) to another, as well as identify the VLAN requirements of each VM. It will also track VM configuration changes such as additions, deletions, or modifications of VLANs.
The fabric will also track the additions or deletions of VMs and hosts, and dynamically provisions VLANs on the servers facing physical ports. (taking into account, the hots, switch ports connected to the VM, vNIC that connects the VM to a virtual switch, VLAN information of port groups, and port groups required for the VM).
** Note that Pluribus switches can achieve Automatic LAG formation between themselves. I discovered this during a home test lab before I read about it, so the first instance was a bit of a surprise for me. Here, NetVisor OS will achieve Auto-LAG between the switches & the ESXi hosts as well. NetVisor uses LACP and LLDP to bundle the ports.
vCenter Connection Service (vCCS) is used to connect to vCenter. it facilitates both operational simplification, as well as visibility of entities, resident in the overlay being supported by the Pluribus fabric. The entities include:
- Server, hypervisor, physical NICs
- VM: Guest OS, vNIC properties such as MAC address, IP address, and VLAN.
- VM Kernel ports and the associated services like vMotion and vSAN.
Some of the information vCCS obtains from vCenter, includes:
- The host where the VMs exist & ONVL switch ports connected to the VM.
- vNICs that connects the VM to a virtual switch.
- VLAN assignment of port groups, and port groups required for the VM.
It currently supports only one vCenter connection, & requires ESXi 5.x and 6.x versions of vCenter.
Visibility: vSphere Network Traffic
Insight Analytics can have 3 sources of data. These include the switches running ONVL (NetVisor’s Embedded Telemetry), Network Flow (Netflow/sFlow) from any flow enabled devices, and lastly, vProbes: A virtual appliance deployed inside a virtual server. This monitors all traffic on the host hypervisor, and enables visibility into :
- Intra-server and inter-server + virtual server to physical server traffic
- East-West traffic across ESXi hosts
- VM-to-VM applications/services
- Vmkernel services including VSAN, vmotion and IP storage traffic
It connects to Hypervisor vSwitch, and observes all traffic within assigned port groups. It automatically discovers all application flows within virtual server. Traffic can be classified by portgroup, vSwitch, Host, VM type, or location.
Pluribus also has a vCenter plugin for Adaptive Cloud fabric. vRO offers GUI driven integration with Role based Access Control.
Hasan Mansur 