Backconnect proxy
A single gateway address that transparently routes each request through a different address in the provider's pool.
A backconnect proxy gives you one endpoint that resolves to many addresses behind it. You configure one host and port; the provider maps each connection onto a different member of the pool.
#Why it is built this way
Managing thousands of individual addresses in client code is impractical: they change, they fail, and they need health checks. A backconnect gateway moves all of that to the provider. Your configuration stays a single line even as the pool underneath changes continuously.
#Practical consequences
- Your firewall rules and connection settings reference one host, not thousands.
- You cannot choose a specific exit address directly — you express intent through parameters such as country or session identifier.
- The gateway’s own location is not the exit location.
Nearly all modern residential and mobile services are backconnect. Per-address lists are now mostly a datacenter and ISP arrangement.
#What the single endpoint hides
You configure one hostname and port. Behind it the provider selects an exit node, applies your geo-targeting, enforces your plan limits and records the traffic. Several consequences follow that are easy to miss.
| You observe | What is actually happening |
|---|---|
| One address in your configuration | Many addresses in use, chosen per request or per session |
| A connection failure | Could be the gateway, the selected exit, or the destination |
| Consistent latency to the gateway | Says nothing about latency from the exit to the target |
| An unchanged endpoint | The pool behind it changes continuously |
#Why it is built this way
Consumer devices cannot accept inbound connections reliably. They sit behind NAT, they move between networks and they disappear without notice. So the devices connect outward to the provider and hold the connection open, and the provider routes your request down an existing one. The name describes that reversal: the node connects back, rather than being connected to.
This is also why an exit can vanish mid-request. The node’s owner closed a laptop, and no protocol exists to warn you first.