There are many website like https://bitnodes.earn.com/ but I’m failing to understand if they collect nodes from compulsory submission on the web interface for statistics purposes only or they use a way to scan the whole network ?
A second part of the question regarding the type of the node is whether it’s required to run a full node to create blocks as a miner.
Sites like bitnodes use a crawler to discover many nodes on the Bitcoin network. They start a node which connects to some nodes initially, and then requests for more nodes to connect to from those nodes. They connect to the nodes they received and repeat the process until they have queried every listening node that they can.
Note that this does not get every node. It can only scan listening nodes and only the nodes which have been up long enough for another node to return its IP address in a query for nodes to connect to.
