IPv4 was vanishing quite rapidly. Contemporary studies indicated that it may be depleted within the next ten years – around 2005! These findings challenged the Internet community to start looking for a solution. Two possible approaches were at hand:

1. Minimal: Keep the protocol intact, just increase the address length. This was the easier way promising less pain in the deployment phase.

2. Maximal: Develop an entirely new version of the protocol. Taking this approach would enable incorporating new features and enhancements in IP.

Because there was no urgent need for a quick solution, the development of a new protocol was chosen.Its original name IP Next Generation (IPng) was soon replaced by IP version 6 which is now the definitive name. The main architects of this new protocol were Steven Deering and Robert Hinden.

The first set of RFCs specifying the IPv6 were released at the end of 1995, namely, RFC 1883: Internet Protocol, Version 6 (IPv6) Specification [RFC1883] and its relatives. Once the definition was available, implementations were eagerly awaited. But they did not come.

The second half of the nineties was a period of significant Internet boom. Companies on the market had to solve a tricky business problem: while an investment in IPv6 can bring some benefits in the future, an investment in the blossoming IPv4 Internet earns money now. For a vast majority of them it was essentially a no-brainer: they decided to prefer the rapid and easy return of investments and developed IPv4-based products.

Another factor complicating IPv6 deployment was the change of rules in the IPv4 domain. Methods to conserve the address space were developed and put into operation. The most important of these was Classless Inter-Domain Routing (CIDR). The old address classes were removed and address assignment rules hardened. As a consequence, newly connected sites obtained significantly less addresses than in previous years.

The use of CIDR may well have delayed the need for IPv6 in the eyes of many people, but not in all. Somewhat perversely, the use of CIDR accelerated the perception of a lack of address space in the eyes of those who were relatively new to the Internet. Since the Internet was booming, requests for new address blocks were increasing, yet the size of assigned address blocks were being reduced. Thus, new or expanding sites had to develop methods to spare this scarce resource. One of these approaches is network address translation (NAT) which allows a network to use an arbitrary number of non-public addresses, which are translated to public ones when the packets leave the site (and vice versa). Thus, NATs provided a mechanism for hosts to share public addresses. Furthermore, mechanisms such as PPP and DHCP provide a means for hosts to lease addresses for some period of time.