The v3-only era: life after legacy onion addresses disappeared
Four years after Tor pulled the plug on v2 onion services, we look back at why the old addresses died, what v3 fixed, and how the ecosystem settled.
notes from running a 24/7 tor client: outages, weird latency spikes and other things that make badges jump.
Four years after Tor pulled the plug on v2 onion services, we look back at why the old addresses died, what v3 fixed, and how the ecosystem settled.
From SHA-1 collisions to the October 2021 cutoff: how the Tor Project deprecated v2 onion services, and why thousands of sites initially stayed behind on legacy addresses.
Tor Browser 13 was more than a visual refresh. Its rebase onto Firefox ESR 115 changed how the anonymity browser absorbs security fixes on desktop and Android.
Snowflake turns thousands of volunteer browsers into disposable WebRTC proxies that stand in front of Tor. How the design works, what limits it, and why censors struggle to keep up.
How obfs4 bridges hide Tor traffic inside random-looking bytes, where to request bridge lines from BridgeDB, moat or email, and why active probing still lets censors find them.
WebTunnel disguises Tor bridge connections as everyday HTTPS traffic. We explain how the pluggable transport works and what performance it costs.
Every .onion connection is a small choreography of descriptors, introduction points and a rendezvous relay. Here is how the six-circuit handshake works, in plain language.
OnionBalance spreads onion services across backend instances, but it is not a DDoS shield. How frontend-backend balancing really works, where it failed in 2023-2024.
Onion services can be healthy yet unreachable. We explain how hidden service directories store descriptors, why republishing rules matter, and what overload does to reachability.
Tor rebuilds circuits every ten minutes but keeps the same guard relay for months. Research shows why switching guards early makes deanonymization more likely.
Circuit builds, relay congestion and quiet guard rotations explain most onion latency swings. Probes over weeks show why a hidden service feels instant at noon and crawls by midnight.
Tor Metrics charts look authoritative, but user numbers are modeled estimates rather than counts. What the curves really measure, their blind spots, and how not to over-read them.