Shoutout to take responsibility
Response to last day’s tweets.
The system can handle nodes going down, as shown here: https://twitter.com/coz_official/status/970494691471798273
I call it an edge case because it requires a node being out of sync, going offline, then back online during consensus, and send ”change view” (because out of sync) when the consensus is already at a high iteration number (6 or more). Every time a view is changed, the view time is increased. This mean that there is an increasing delay in each iteration to be able to receive all consensus messages. What we experienced yesterday is a combination of 1. Delay is too long, 2. Consensus messages being forwarded with priority but not received with priority (queued behind other messages). These will be fixed in upcoming patch.
Even though delay occures, the consistency is and never was threatened. NEO’s consensus algorithm prioritize finality, so the worst thing that can happen is a block delay. Canesin and members of City of Zion + Red4Sec addressed these claims on both Twitter (https://twitter.com/canesin/status/970392581279535105?s=19) and Reddit (https://www.reddit.com/r/NEO/comments/81xzz5/dev_on_twitter_rips_into_neo_looking_for/dv68dda/).
Our development community is open and anyone sending me and/or CoZ Council a message will get a reply. My email is open to everyone (firstname.lastname@example.org), and I also answer on everything on our Discord (@lerider). We are very reachable and I feel it is a pity that some people make claims without even looking into the problem. It is one thing that ”crypto-fans” are making these claims; that is fully understandable. However, when so called ”industry leaders” are spreading claims like this, then we should take a timeout and consider if we really want to have these persons as ”industry leaders” and if they really are as knowledgable as they claim to be.