This repository grew out of operating a non-exit Tor middle relay and using it as a reason to dig deeper into Linux services, TCP behavior, routing, latency, DNS dependencies, and how a distributed network behaves outside a classroom diagram.
The relay itself is only one part of the repo. I also kept the troubleshooting notes, research, concepts, and network-analysis work that came out of running it.
- Linux service administration
- Tor relay configuration
- TCP/IP and port reachability
- Logging and troubleshooting
- Latency and jitter
- DNS dependencies
- Routing behavior and internet-path analysis
- Privacy-aware infrastructure
| Area | Purpose |
|---|---|
cachemeoutside_relay/ and my_relay/ |
relay-specific configuration and notes |
Troubleshooting/ |
failures, observations, and fixes |
analysis/ |
network and performance analysis |
concepts/ |
networking concepts connected to the relay |
research/ |
supporting research and deeper dives |
Cloudflare_labs/ |
related networking experiments |
This was one of the projects that moved networking from something I studied into something I could observe.
Running a real network service forced me to think about questions such as:
- Is the service running, or is it actually reachable?
- Is a slow connection a host problem, a path problem, or normal internet variation?
- What can logs prove?
- What changes when traffic crosses systems I do not control?
- How much can I infer from the edge of a distributed network?
Those questions now show up again in my newer work on BGP failover and network observability.
- Mini Internet — controlled BGP failover and convergence experiments
- Network Flight Recorder — local-first observability, incident evidence, diagnosis, and guarded recovery
- Cloud Network Architecture Journal — routing and architecture notes
This repository is a documented project/lab record; it should not be read as a claim that the relay is continuously online today.