From b68a18fa85b8577629416f1947e2bede4dbbdddf Mon Sep 17 00:00:00 2001 From: Ajay Ramachandran Date: Tue, 15 Nov 2022 15:16:34 -0500 Subject: [PATCH] Updated K Anonymity (markdown) --- K-Anonymity.md | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) diff --git a/K-Anonymity.md b/K-Anonymity.md index 8ffb205..cc4f5f1 100644 --- a/K-Anonymity.md +++ b/K-Anonymity.md @@ -213,4 +213,14 @@ The client then gets a list of possible videos that have that prefix. This list The client then gets back these 11 videos (in this case) and checks to see if the video it is looking for is in the list. If it is not, it treats it as not found. Since the list may, or may not contain the video, this makes it even more secure than a 1 / 11 chance. -This system allows the client to be lightweight and always get the latest segments on every video while still ensuring the server cannot predict which videos they are watching. \ No newline at end of file +This system allows the client to be lightweight and always get the latest segments on every video while still ensuring the server cannot predict which videos they are watching. + +*** + +Note: The extension also sends an anonymous request after segment is skipped to give you a "time saved" metric on your submissions. This can be disabled, and it is recommended to disable this setting if you want to fully prevent leaks. So, if you watch a video, and it does have segments, and you do skip a segment, then the server does get access the that segment ID, which is directly linked to the video. + +Why is K-anonymity still used then? Even with skip counting enabled, K-anonymity still helps with the following cases: +- Unlisted videos with no segments (Since those are private to everyone who doesn't have access to the URL, leaking the video ID is not a good idea) +- Videos with no segments you skip + +K-anonymity was originally developed as an optional feature but was enabled by default for the above reasons. \ No newline at end of file