Why am I seeing 3TB of data transfer on my relay?

[noahpark233 (Noah Park)]

I setup a relay server / discovery server in AWS on a single EC2 instance. I’m running the server in private mode (i.e. pools="" ). I thought that if I connected site A and site B to this discovery and relay server the 2 locations should connect and transfers happen primarily direct. However, it looks like I incurred a hefty $300 expense for data transfer for 3TB of data that went through the relay.

What do I need to check to make sure I’m setup correctly (Since I’m guessing I’m not) to make sure the transfers happen directly?

[AudriusButkevicius]

One of the devices should be available on the internet, as in not behind a NAT/firewall or opening ports to the internet using UPnP or NAT-PNP

[canton7 (Antony Male)]

Also note that AWS is rarely the cheapest option for a small number of always-on servers: their value comes from if you need to spin machines up and down quickly.

4TB will only cost you $40/mo on Linode, $20/mo on DigitalOceant, $11/mo on SolVPS, similar amounts on others (I’m not affiliated with any of these).

[noahpark233 (Noah Park)]

Thanks. This is really helpful. I only used AWS because its what I’ve used before. I’ll check out the others you mentioned today.

[capi (Martin)]

Since above some providers were mentioned for relays, I’d also point out to Hetzner (in Germany and Finland): their smallest “cloud-server” for €3/month includes 20TB of traffic. Might be interesting for someone wanting to run their own relay.