In I2P there are no single domain registrars, each

In I2P there are no single domain registrars, each user can independently register himself a domain or subdomain. Each user has several files, in which the correspondence of domains to addresses in the network I2P (not IP-addresses) is prescribed. This is the so-called address book. Also, each user has a list of subscriptions-the addresses of the servers from which the addresses are periodically updated. Subscription is just a text file, which lists the addresses of the servers from which the links to the server are downloaded by their names. From these servers, addresses fall into the so-called router address book.Everyone can register their own domain in the I2P zone undergoes some kind of bureaucratic procedure. There will inevitably be a problem with the fact that several people can register the same domain, and both of them will be correct from the point of view of the protocol. How to solve this problem at the level of the I2P protocol, unfortunately it was not possible to find out, but there are already several servers on the network – domain name catalogs. Most likely in the future, the struggle will unfold for registration in them. It is rumored that this can be one of the ways to earn money on this network, when the domain is created freely, but it is registered in the catalogs for money (it is not excluded that it is a subscription fee). But in this case it is interesting how the problem of competition of such directories will be solved, because it may turn out that in one catalog the domain is registered for one server, and in another – for another. At the moment, there is expanse for potential phishing sites. But in this way, domain names are resistant to censorship.I would like to make one more clarification, so you do not expect too much from the I2P. It is believed that this is a distributed network. However, addresses are primarily allocated. Sites all work on one particular server. This server can be closed, but it is very difficult to find (learn its IP). It is also very difficult to identify the user who worked at one time or another with the server. It should be clarified that the server also does not know, no user IP-addresses, only its identifier in the I2P network, which periodically changes. But for reliability, the user should not show this ID to anyone (he himself can learn it).At the moment, work through I2P is very slow and unreliable. Sometimes, sites start to open quite quickly. Apparently it depends on the paved tunnel. Actually, this is the problem of all distributed networks, if there are not many participants in it. I want to believe that this project will not remain a toy for a very narrow circle of users, and it will at least become a little more famous, and at the same time reliable.