Classless interdomain routing (CIDR) circumvents the old Classes A, B, and C network boundaries by combining arbitrary networks and masks to form a prefix that is used for routing.
The BGP routing protocol was updated to version 4, primarily to allow prefixes (the combination of the network and the mask), rather than announcing classful networks to peers. Therefore, large providers could announce both supernets and subnets to their peers, significantly reducing the prefix count in core Internet routing tables. Indeed, based on their knowledge of address allocation history, many large providers began to actively filter incoming route announcements, ignoring those that did not meet their minimum size policy.
NSI also managed the allocation of domain names within the United States. Not surprisingly, the biggest area of contention was the .com domain because various companies disputed the allocation of particular domain names based on their registered names. Several court battles ensued, some inevitably involving NSI in one role or another. For the most part, NSI merely abided by the rulings of the courts.
NSI maintained the databases of the so-called root nameservers, which contain the zone files for all top-level domains. Although NSI operated root nameservers, there were many others around the world. Some operated on a voluntary or government-funded basis, and some operated to provide commercial service. NSI makes the root zone files available to operators of the root nameservers through methods other than DNS, such as FTP. All nameservers connected to the Internet are primed with the addresses of the root nameservers as a point of origin for the demand-based creation of a DNS cache.
The NSF funding of NSI served as an interim measure that enabled the Internet community to establish policies and procedures for governing the functions, and eventually to become self-sufficient. As a result of heavy consultation with IANA, the IETF, RIPE, the APNIC, and the NSF (among others), the American Registry for Internet Numbers (ARIN) became operational on December 22, 1997. IANA then transferred the role of assigning Internet addresses from NSI to ARIN.
ARIN, a not-for-profit body, operates under a self-funding procedure achieved through registration charges for new IP address allocations. The charge is levied annually and is proportional to the size of the address allocation. The funding model is based on experience reported by the APNIC and RIPE, and recognizes that the per-address price of allocations decreases with the size of the allocation. For example, an allocation of 8192 hosts (32 Class Cs) costs 30 U.S. cents per address per year. These charges are not applied to preARIN allocations.
Was this article helpful?