←back to thread

160 points Metalnem | 2 comments | | HN request time: 0.439s | source
1. agwa ◴[] No.44494778[source]
Sunlight and static-ct-api are a breath of fresh air in the CT log space. Traditional CT log implementations were built on databases (because that's the easiest way to implement the old API) and were over-complicated due to a misplaced desire for high write availability. This made operating a CT log difficult and expensive (some operators were spending upwards of $100k/year). Consequentially, there have been a rash of CT log failures and few organizations willing to run logs. I'm extremely excited by how Sunlight and static-ct-api are changing this.
replies(1): >>44495370 #
2. eddythompson80 ◴[] No.44495370[source]
I wonder if this is the solution something like SponsorBlock is looking for[1][2]. They have a similar-ish problem. How to replicate crowdsourced data that trickles in slowly, but ideally you want replicated quickly.

WAL replication, rsync, bittorrent, etc all things that don't quite work as needed.

[1] https://github.com/mchangrh/sb-mirror/blob/main/docs/breakdo...

[2] https://github.com/ajayyy/SponsorBlock/issues/1570