Bitcoin maxis have a very ambitious and grand vision and I don't think they should compromise it. It sets a high bar that drives innovation. I do think that over the short term any nation scale usage of bitcoin for everyday payments will look very different from that vision and that is ok.
For e.g., mining operations can be distributed within the country itself to ensure decentralization within the country. Then this internal blockchain network interfaces with the slower and more expensive bitcoin blockchain to do final settlements?
Could this work?
This is a good question and we should always be asking "Why Bitcoin?" and think about the trade-offs.
For small amounts of bitcoin a trusted third party, especially an institutional trusted third party, makes a lot of sense. Convenience, availability and low transaction fees dominate the user's needs.
For large amounts of bitcoin, you probably don't want to use a trusted third party and you can afford to run your own infrastructure. For instance a national bank should not keep all its bitcoin on coinbase.
Bitcoin allows you to choose between these two trust models. It is very hard to remove a trusted party from a system which fundamentally relies on a trusted party. However it is often easy to inject a trusted party into a system to make it more useable or scalable. Thus, for maximum flexibility you want your base layer to be as trustless as possible and then add trust as needed to solve different use cases.
1. You can easily support billions of bitcoin transactions a second since the transactions only change centralized database entries.
2. Users can send and receive funds using a website. You could even issue tokens that user could use to authorize fund transfers at Point of Sale (see Credit Cards).
If you want to ensure that this trusted party isn't hacked, you can have five trusted parties instead of one and use multisig so that 4 of 5 of them must consent and sign to do an on-chain transaction. You can something similar for the internal ledger maintained by the trusted parties to ensure that one of them doesn't maliciously alter user balances.
The typical rejoiner to this is: Why do you need Bitcoin to do this? Why not do it with fiat? They aren't wrong per se, you could do this with fiat and it would work pretty well. However building a system like this with Bitcoin does have some advantages:
1. User's can withdraw funds on-chain,
2. there is a clearly defined on-chain layer for settling large balances between institutions,
3. bitcoin isn't created by fiat from a government bank. This removes the risk of inflation in longterm international contracts, etc...