Onion Routing
Also known as: onion-routed, layered encryption routing, Tor routing
Onion routing is a technique for anonymous communication in which a message is wrapped in successive layers of encryption, like an onion, and relayed through a chain of volunteer-run nodes. Each node peels off one layer to learn only the next hop, so no single relay ever knows both who sent the message and where it is heading. Invented at the U.S. Naval Research Laboratory in 1996, it is the basis of Tor.
Onion routing solves a problem that encryption alone does not: even if your message content is unreadable, the routing information on the outside of every packet still shows your IP address talking directly to a destination. That metadata can be as damaging as the content. The technique was first described in the 1996 paper "Hiding Routing Information" by Goldschlag, Reed, and Syverson at the U.S. Naval Research Laboratory, originally to protect intelligence communications.
The mechanism is layered encryption. Before sending, your client picks a path of several relays and encrypts the message once for each relay, in reverse order. The result is an onion: the outermost layer is encrypted to the first relay, the next layer to the second, and so on. The first relay decrypts only its own layer, which reveals nothing but the address of the second relay; it forwards the still-encrypted remainder. Each hop peels exactly one layer. Only the final relay sees the innermost contents, and even then it does not know who originated them.
The security property is structural: the first hop knows your IP but not the destination, the last hop knows the destination but not your IP, and no relay knows both. An adversary would have to control or observe multiple specific relays at once to link the two ends. This is distinct from a mixnet, which additionally batches and delays traffic — onion routing forwards each packet immediately, prioritizing low latency.
How it works
Onion routing the concept vs. Tor the network — an important distinction. Onion routing is the general technique of nested layered encryption across relays. Tor is one specific, widely deployed implementation of it: a global network of thousands of volunteer relays with directory servers, perfect forward secrecy, congestion control, and hidden (.onion) services, formalized in the 2004 paper "Tor: The Second-Generation Onion Router." Tor is to onion routing what a particular car is to the internal combustion engine. Other systems use onion routing too; Tor is simply the dominant one.
A standard Tor circuit has three hops:
1. Guard — the entry relay. Knows your real IP but not your destination. Deliberately kept stable for weeks to resist Sybil-style guard-rotation attacks. 2. Middle — sees only the guard's IP and the exit's IP. Knows neither end. Rotated frequently. 3. Exit — peels the final layer and contacts the destination. Knows the destination but never your IP.
For connections between two Tor users (hidden services), a rendezvous point is used and no exit node is involved at all, so neither party's IP is exposed to anyone.
How RVNT uses Onion Routing
RVNT embeds arti, the Tor Project's official Rust client, directly in the app — no separate Tor Browser or system service. Standard mode routes identity-server lookups, DHT queries, and peer messaging over 3-hop circuits; maximum-privacy mode extends these to 5-hop circuits and increases cover traffic. When both peers run Tor, RVNT uses hidden-service rendezvous so neither IP is ever exposed. Local-network and offline-mesh peers skip Tor, since a nearby peer already knows you are close. See the Tor integration docs.
Frequently asked questions
Is onion routing the same thing as Tor?
Not quite. Onion routing is the general technique of wrapping a message in layers of encryption and relaying it through nodes that each peel one layer. Tor is the best-known implementation of that technique — a specific global network of volunteer relays with directory servers, forward secrecy, and hidden services. Tor uses onion routing, but the concept predates and is broader than Tor.
Why is it called onion routing?
Because the message is wrapped in successive layers of encryption, one per relay in the path, like the layers of an onion. Each relay peels off exactly one layer to learn only the address of the next hop, never the full route. The layered-encryption metaphor gives the .onion address suffix its name too.
Can onion routing be broken?
No single relay can deanonymize you, since none knows both the source and destination. But a powerful adversary who simultaneously observes traffic entering the first relay and leaving the last can sometimes correlate timing and volume to link the two ends. This end-to-end correlation attack is onion routing's main theoretical weakness, which is why RVNT adds mixnet-style timing defenses.
Every definition here describes something RVNT actually ships — a post-quantum, end-to-end-encrypted, peer-to-peer messenger with no phone number and no servers.