Craig Bruenderman

Overview Greetings, this series will cover the design/build of VeloCloud Clusters in a multi-region AWS deployment. This exercise dips into a bunch of different technology skill areas, and I couldn't find anything comprehensive written, so here we are. This will be an ambitious multi-part post, so strap in. I'm decent with routing/switching & Velocloud, and moderately experienced with AWS, so please comment if I need to correct anything. How to eat the elephant Part 1 - Review the design options Review the mechanics of various Velocloud HA options Some general review of the relevant parts of AWS and scope Review AWS Transit Gateway Review selected AWS and cluster layout Part 2 - Initial deploy and configuration of Velocloud devices Part 3 - AWS Deployment Deploy and configure various AWS network constructs TGWs and routing Part 4 - Further configuration of Velocloud NSD via Edge configuration Establish BGP peering with AWS TGWs Validation through VCO, CLI, and AWS Dashboard

Deploying VCEs in AWS Part 1 gave a design overview of what I want to build. In Part 2, I'll bootstrap two Velocloud Hub Clusters. Each Cluster will exist in a different AWS Region and consist of two VCE Cluster Members, one in each Availability Zone. INHO, the Velocloud AWS vVCE deployment documentation is not great. It only covers CloudFormation, which not everyone uses. I've done this with Python + Terraform as well, but that's pretty advanced and it would be nice to have a guide for those who just want to click their way through the Velocloud Orchestrator (VCO) and AWS Dashboard. I'll do my best to document that... I'm going to start out with some pre-work in the Velocloud Orchestrator, and then move on to actually spawning the VCEs in AWS EC2. Network Interfaces We need to understand a few things about interfaces as they pertain to AWS, the VCE's underlying Linux OS, and the settings in VCO. Our understanding of network interface cards gets a bit fuzzy i