VeloCloud Q&A: How to size an Edge

My next Q&A entry tries to answer an important question: what edge size you need for each of your locations? An Edge can run on bare metal, or virtual, on top of ESXi/KVM, or inside your major cloud provider (Amazon, Microsoft, Google, Alibaba).

The main elements to consider are the Maximum Throughput (as the sum of upload and download capacity across all the local WAN links), and the Maximum Tunnel Scale (the number of tunnels running simultaneously from the local Edge to the rest). So while at the branch you might go for a lower series device (5xx or 6xx series), you might consider something bigger at HQ or DC (2xxx or 3xxx series, or virtual clusters). Keep the branch hardware small by implementing a hub and spoke topology, and possibly utilising the dynamic branch-to-branch tunnels if you have traffic going back and forth between (if the latency via the hub impacts real time applications).

For a virtual implementation, the underlying hardware will dictate the performance of the Edge, the image is the same. Remember, the server and ESXi licenses are not included in a VeloCloud BoM. And same goes for public cloud.

See more in the video bellow:

And find the datasheet here.