Hello all, after all the discussions I cannot understand what the problem would be with a scheme like the following: a) Member LIRs are divided into 16 categories, named /5, /6, /7, .. / 20, /21 b) Your category is the smallest route that could contain the total IPv4 allocated space of your LIR c) Your fee in euro is T/2^C, where C is your category and T the total RIPENCC costs. Seems simple and fair, it is still a category based model and thus it does not change the tax status of RIPE, and in example: - Up to 2048 IP addresses you are class /21 and pay about 100 euro - Up to 4096 IP addresses you are class /20 and pay about 200 euro - Up to 8192 IP addresses you are class /19 and pay about 400 euro - Up to 16384 IP addresses you are class /18 and pay about 800 euro - Up to 32768 IP addresses you are class /17 and pay about 1600 euro - Up to 65535 IP addresses you are class /16 and pay about 3200 euro ... and so on. Some corrections improvements might be considered to take in other exhaustable resources (IPv6 address space, ASn) and even to some extent non-exhaustable ones (number of tickets per year, number of records or operations in the database). Has a schema like this ever been considered or voted upon ? Regards, A.