6.3.2.2 Design Option 2a – Combined Edge/Compute Cluster

When the leaf and spine network architecture is used, VLANs that back vCloud Director external networks are trunked only to the Edge/Compute cluster. The vCloud Director placement engine deploys edge VMs to a cluster based on VLAN connectivity. vCloud Director automatically places all edge gateways into the Edge/Compute cluster because it is the only cluster where the external connectivity (VLANs) exists. However, vCloud Director will also opportunistically place regular tenant VMs into this cluster (hence its name: Edge/Compute).

This design option has all the scale advantages of leaf and spine architecture. However, the drawback is the possibility of tenant workloads taking up limited space of the Edge/Compute cluster. There are two options to remediate this:

• vCloud Director edge gateways are always deployed by the vCloud system administrator. Prior to edge gateway deployment, the administrator can verify that there is enough capacity in the Edge/Compute cluster. If not, some tenant workloads can be migrated to another cluster. This must be done from within vCloud Director (Resource Pool / Migrate to option). Live migration is possible only if the Edge/Compute cluster shares the same VXLAN prepared VMware vSphere Distributed Switch™ (VDS) with the other clusters. This setup requires at least four network uplinks on the Edge/Compute cluster hosts (two uplinks for the edge VDS with external VLANs and two uplinks for the VXLAN VDS that spans all clusters).

• Artificially limit the size of the Edge/Compute cluster so the placement engine does not choose it for regular tenant workloads. This can be achieved by leveraging the resource pool, which is created manually by the system administrator in the Edge/Compute cluster and attached to the provider VDC instead of the whole cluster. An artificial limit is set by the system administrator and is increased only when a new edge gateway needs to be deployed.