Google Cloud Virtual Edge Deployment Overview

More customers are moving workload to Public Cloud infrastructure and expect to extend SD-WAN from remote sites to public cloud to guarantee SLA. There are multiple options, leveraging distributed VeloCloud Gateways to establish IPSec towards public cloud private network or deploy virtual edge directly on Google Cloud Platform (GCP).

For small branch deployment that demand throughput less than 1G, single virtual edge can be deployed in the Private GCP network. For larger data center deployments that demand multi-gig throughput, hub clustering can be deployed.

Note: In the VeloCloud Hub clustering design, a Layer 3 Instance is leveraged on the LAN side to run BGP between hubs in the cluster and the Layer 3 Instance for route distribution in LAN. Since the GCP router does not support dynamic routing protocol, a third-party virtual router is required in the GCP infrastructure.

This document illustrates a basic topology and a high-level workflow to deploy a virtual Edge (vVCE) on GCP. For more information, see Virtual Edge Deployment on GCP.

Prerequisites

GCP Machine Types

Bandwidth throughput and the number of network interfaces need to be considered when sizing the VeloCloud SD-WAN Virtual Edge. The minimum number of network interfaces required is three (GE1, GE2, GE3).

Table 1. Bandwidth Throughput
Throughput 30 Mbps 50 Mbps 100 Mbps 200 Mbps 400 Mbps 1 Gbps
vCPU 2 2 2 2 4 4
Memory 4 GB 4 GB 4 GB 8 GB 8 GB 8 GB
Table 2. Machine Type
Machine Type vCPUs Memory (GB Max NICs
n2-highcpu-4 4 4 4
n2-highcpu-8 8 8 8