How does an ISP work?

Question

Computer networks

Good afternoon, I am new to network technologies, so do not judge strictly, it became interesting how the Internet provider works and why different providers act differently, if this question has already been asked, then I apologize.
Two providers, one works on GPON, ONT can be in the attic in the house, twisted pair to subscribers or at the subscriber's house, if the tariff is gigabit, white dynamic ip addresses via dhcp, there is no binding to the mac address. Each subscriber has its own client-VLAN, and ONT has its own vlan, i.e. it works via QinQ.
And the second provider via ethernet, subscribers from one or more switches can be in the same vlan, on a territorial basis, gray ip addresses are also via dhcp, with binding to a mac address and from iptv via multicast.
1. Are there any guests / international standards on how a provider should build a network, or at least how smart guys do it and how not to do it? Should I make a separate vlan for each subscriber or combine groups of subscribers into broadcast domains? What are the pros and cons for each of the methods for the provider and for the subscriber?
2. What determines the choice of how to build a network at the physical and logical levels?
3. Is it related to the type of cable? Or is it possible to do for optics, so that subscribers are in the same broadcast domain (for example, from one area), and those that are on twisted pair, on the contrary, for each separate vlan?
4. Does the first provider have dynamic ip addresses, i.e. is there some kind of pool that was allocated to it for subscribers and the client receives the one that is free now? does the second provider also issue an ip from the address pool or do they all come out under the same one? Is this due to the fact that the second simply ran out of ip addresses due to a larger number of users and the first provider will also transfer everyone behind NAT when there are a lot of clients? How do large providers use the ip range, dynamic ip or for nat (it is clear that they buy out small providers and everywhere in different ways, is there a standard that they strive for or should strive for, or if everything works, then it will do?)
5. About iptv. I have a router with dns from Google at home, a cable from it to Smart TV and the provider has its own application for viewing. The problem was that I did not have a list of channels and it was decided that I set dns on the router from the provider, as explained in the TP, this was due to the fact that since I had a gray IP it corresponded to the NAS address and since this address did not match the address of any of the subscribers, then the list was empty. But how can a NAS address be issued to a subscriber? it's just network storage, nonsense or not? checked my IP on the site and asked a friend connected to the same provider from my house to check. we had different ip. so technical support is lying? (this is the second provider, yes, everything would work with a static ip, but I'm wondering what's wrong with gray ip)
Thanks for reading and for the answers :)

Answer 1

Everything is easier.
An ISP buys internet from uplinkers and sells it to subscribers.
On this the main essence of how it works ended.
Everything else is purely individual. Common things - the usual world practices for ensuring quality and security in the form in which a particular provider could implement them.
Most things are justified only economically. For example, if a provider can stretch an optical fiber into a house, it will not pull it to each subscriber - and it’s expensive and you can’t just plug an optical fiber into a computer, so it’s easier to put a switch in the house and distribute the last mile over ethernet.
On the other hand, if this is a provider that initially works according to docsis, on coaxial, the technology itself requires that each user have their own final modem, which can already be connected to the computer either via USB (for a long time) or via ethernet - modern options, but coaxial will enter the apartment.
If the provider works outside the city, distributes the Internet via satellite - it has its own characteristics.
All options are not particularly related to how many IP addresses the provider has, but it is clear that small providers will intuitively have fewer of them. In addition, white IP can be sold as a standalone service.
Your question does not pull on a question but on a small dissertation.
But what it all depends on is money.
Standards - there are different individual moments, but there is no standard that will describe in detail the conditions of the provider.
Minimum requirement - do you have an Internet? connect a neighbor in any way but for money, register as a legal entity, pay taxes - and you are already a provider. Yes, and get a license (depending on the country, though).

Answer 2

1) No. The task of the provider is to make money by providing the service. He earns.
How to build a network is his business, he decides it based on the situation, available equipment, experience, finances, etc.
2) From the current situation and plans for the future.
3) No.
4) IPv4 addresses have already ended and they are not being given out to anyone. As a result, the provider has as many addresses as he managed to grab before they ran out. Some have more, some have less. Now addresses can be obtained either by buying a range from someone, or most often when a large provider takes over a smaller one.
The existing addresses have to be saved - at least dynamic addresses should be allocated. If it's not enough anyway, transfer users behind NAT.
Almost everyone provides a white static address at the request of the subscriber - of course, not for free.
5) Of course it is possible. Although the situation is not entirely clear.
You have a gray network behind your router, your ISP has a gray network - addresses may overlap.
Although I believe that it didn’t work for a simple reason - the list of channels hung on a server with a gray address in the provider’s network, if you resolve names with third-party DNS servers, they don’t know anything about the provider’s gray network, and they don’t give out the address, or give out incorrectly.

Answer 3

1. No. Provider - a person providing services, in this case, tyrnet access services. I connected to the backbone, installed billing, fulfilled the requirements of SORM, received a license - that's it, the provider. As he wants / can / knows how - he builds a network, the user will vote with a ruble if the network is crooked.
2. From the thickness of the provider, from the amount of money, from what plans he has for development. The technical side here may not play a role at all.
3. No. This is by no means due to technical issues, but mainly to how much money is invested in development
4. The provider's address policy is determined by how many IPv4 blocks it has. Since they don’t give them out right now, you have to be effective managers :) Whoever has PPPoE - he grinds his teeth, but gives everyone a white IP, who doesn’t have it - he can trade them, renting them out, who has very little - gives gray ones, releases through NAT. There is no single practice or any method to tell the provider what he is doing wrong, if he does not give a white IP - perhaps he simply does not have them :D. This can lead to such perversions, such as, for example, in Krasnoyarsk, at one time, all providers agreed and divided the 10.0.0.0/8 network inside Krasnoyarsk :) And they began to issue IPs from this network to clients so that clients from other Krasnoyarsk providers could connect to these IPs . Then I had to butt heads with the support for a long time, poke them in the nose with RFC1918...
5. The provider's DNS may respond differently to requests from different subnets and give different zones with different information - do you know about this? Google's DNS, for example, may not know anything about the provider's IPTV server, and this may be done quite intentionally.