All Edge reference architecture components implement a high availability design by default. In addition, VNF's can take advantage of platform capabilities to extend their availability in compute, storage, and networking.

vSphere High Availability

Redundancy with vSphere uses VM level replicas in conjunction with the VNF high availability architecture. vSphere HA can instantiate a new VM in the event of host failure thus providing redundancy for a VNF and VNFC pair. vSphere HA can be fully automated without the need for manual intervention for failure and recovery.

Figure 1. vSphere High Availability

VMware NSX-T Data Center Availability

NSX-T Data Center, by default, provides high availability to VNFCs in the overlay network. NIC Teaming and schemes, such as Equal-Cost Multipath (ECMP), Graceful Restart, and Link Aggregation Group (LAG), provide redundant connectivity. The NSX-T architecture also separates management, control, and data plane traffic to further optimize service availability.

VMware vSAN

vSAN is fully integrated in vSphere and provides policy-driven fault handling with platform awareness such as chassis and rack affinity to store object replicas. The virtual storage is integrated with vRealize Operations so that in case of failure, the failed VM can be cloned and spun up automatically. Storage vMotion can be used to perform live migration of virtual machine disk files (VMDK) within and across storage arrays when these are deployed in the environment. This ensures continuous service availability and complete transaction integrity at the same time.