To become a node in the nkn network, an ID must be generated which takes time and a 10 nkn fee must be paid. This is a strong disincentive for hackers to attack the network - it would cost them a lot to do this. But there is no such barrier for those who have already done those things and then deciding to stop mining at a later date - those nodes can still be used.
Once a miner has decided mining isn’t worth their time any more, there’s nothing in the way to stop them from using these nodes to attack the network.
Former miners of Bitcoin still need to pay huge electricity costs to power their antminers if they want to sustain an attack, making it not worth it. To have nodes in Proof of Stake coins you need to put your money at stake - and it is seized by the network if the nodes you staked attack the network. But for NKN there’s no direct disincentive for a former miner to not attack the network. There is only an opportunity cost associated with it - if they are attacking the network then it means they couldn’t have been using their hardware for other purposes.
How can we call NKN secure if this is the case? A former miner can perpetually attack the network with very little cost. I’m not an expert on what an attacker can do if they have over 50% of the nodes in a network, but I believe they can stop new blocks from being produced. Is that correct?
Requiring a number of tokens to be staked by a node to mine, and forfeiting them if that node attacks the network, is one solution to this problem. However this puts a cap on the number of nodes that may be in the network, since the max number of nodes would be total_number_of_nkn_tokens / tokens_staked_per_node. If there was a way to making the staking requirement dynamic so that there was no limit, I would probably be in favour of that approach.