Lot of explanations are in the article and on docs.jami.net, but quickly
1/2: push notifications & the OpenDHT proxy, so if dhtproxy.jami.net goes down or is blocked, push will not work, accounts with push will not connect, but it’s optional, so disabling push & enabling “run in background” and account will uses the DHT without the proxy
- If bootstrap.jami.net goes down or is blocked and no other DHT node is cached, the account will not be able to the DHT, but specifying any another DHT node (can be another Jami on the LAN, another dht node in the world) and you will be able to communicate to users on that DHT
- If ns.jami.net usernames on ns.jami.net will not work, but any other nameserver will. usernames are optional and hashes can always be used. So instead of adding firstname.lastname@example.org you will use hashes
- if one TURN goes down, if you have ipv4 only + NAT and no upnp, you will not be able to do connections with other people without upnp. But other will not care. And changing to another TURN will work.
So with or without these services, Alice and Bob will do exactly the same, contacting each part via the DHT, do a p2p link, discuss/transfer files.
Read about DHT, there is plenty of doc online about this