buffer-read-write-up is a sporadic newsletter filled with great content on
decentralization and metadata resistant systems. Got an idea or suggestion?
Open an issue!
Ever wondered who's downloading what in BitTorrent? Scott Wolchok and J. Alex Halderman, the authors of "Crawling the DHT for fun and profit" show how easy and cheap it is to crawl the BitTorrent Vuze DHT to get insights of which IP:port pairs are storing and requesting files in the network. They show how the crawler they built can a new trackers index with millions of entries in a question of hours and how it can track more than 8M ips downloading over 1.5M files in 16 days.
For more info, check the original paper or the short summary.
The authors propose a two-tier architecture for performing safe garbage collection in CRDTs and a migration protocol for upgrading replicas to the highest, updated tier. The replicas which are part of the tier-1 (core) maintain and perform operations on the state which has been through the latest garbage collection in the network. The second tier of replicas hasn't performed one or more garbage collection operations. Thus, tier-2 replicas can only perform // TODO: finish!
TODO: description
TODO: description
- Awesome IPFS just got even more awesome
TODO: description