- 611comments
- 817comments
- 290comments
- 360comments
- 21comments
- 132comments
- 155comments
- 91comments
- 72comments
- 84comments
- 223comments
- 83comments
- 12comments
- 6comments
- 42comments
- 194comments
- 82comments
- 17comments
- 27comments
- 106comments
- 45comments
- 75comments
- 47comments
- 58comments
- —discuss
- 461comments
- 14comments
- 146comments
- 9comments
- 27comments
Nice to see this here; happy to discuss our paper with folks.
The TLDR is that there's a classic idea (dating back at least to Chaum's early work if not earlier) that we can get privacy in networks by splitting knowledge across multiple parties and/or multiple architectural elements.
We explore that idea and give examples of where we and others have applied it recently (in some of our services at Invisv and things our colleagues elsewhere have built).
PGPP (for privacy on a cellular network) and Multi-Hop privacy relays such as the one provided by PGPP, or Apple's iCloud relay, are excellent services. Hopefully this kind of protection starts getting built into more and more consumer products. https://invisv.com/pgpp/
A quick summary of The Decoupling Principle: Separate WHO you are from WHAT you do
An example is Apple's iCloud relay. When you browse a URL, it is encrypted and sent to the first server, owned by Apple. This knows WHO you are, but not what you are browsing. The Apple server then sends your request to the second server, operated by a third party such as Fastly. This server gets to see your URL, but it only has Apple's IP address, so it knows WHAT is being browsed, but not WHO you are.
https://media.wired.com/photos/61e0c9a0cdd4f18de04f594a/mast...