IPv6: the basis for the next generation internet

oday's lntcrnet operates ovcr one common network U layer datagram protocol, Internet Protocol version 4 or IPv4. Virtually all Internet communication services have been using the same basic IPv4 packet format for over 25 years, proving that IPv4 was extremely well designed aud in a sense is an unprecedented success in an otherwise rapidly changing world of computer networks. However, for morc than 10 years researchers have been discussing the need for an improved version of IP, originally called .next-generation IP (IPng), now called IP version 6 (IPv6). T h e fact that IPv4 has been so tremendously successful and widely deployed makes it very difficult for any successor protocol to enter the scene. I t is obvious that marginal improvements over IPv4 would not justify the s t rong impact and therefore huge cost that the introduction of a new network layer protocol would imply. Hcnce, in the early '90s a new dcsign addressing most of the recognized weaknesses of IPv4 was started within the Internet Engineering Task Force (IETF). T h e result was IPv6, which has been around i n its basic features for some 10 years .already. The single most significant advantage IPv6 offers is increased address space. Ultimately, this will lead to network simplification, first through less need to maintain rout ing s t a t e within t h e ne twork and second th rough reduced need for address translation; hence, it will improve the scalability of the Internet. Due to early unbalanced IP address allocation policies, t he need for more address space is not yet so pressing in the western world. However, already today some geographic regions, especially Southeast Asia, must make extensive use of multiple levels of network address translators (NATs) to provide lnternct access for those who need it. This problem will dramatically worsen in two phases. The first phase is the introduction of third-generation (3G) mobile communications. If every mobile terminal requires a permanent IPv4 address, we will quickly exhaust the remaining 20-30 percent of'IPv4 addrcsses. Yes, i l is true that 2G and 3G network providers make use of private and/or temporary addresses through the use of NATs and protocols like DHCP, and that NATs to some extent enhance the privacy of mobile users; on the other hand, it also greatly incrcases network complexity and hinders easy reachabilty for mobile terminals. This is not a critical problem for Web surf ing, but is a huge ba r r i e r to t h e widespread introduction of peer-to-peer applications. The sccond phase will be the introduction of truly ubiquitous networking. When every appliance or sensor needs an 1P address, the demand for address space will grow dramatically. At that time the seemingly huge 128-bit address space of IPv6 may he just adequate. Since the introduction of a new network layer protocol with new packet and header formats is a complex and costly process, IPv6 contains many other enhancements toward better mobility support, integrated security and multicast, a new routing mode called anycast, as well as flow labels to ease quality of service management. Once the IP layer needs to he changed, we may as well include all features dcemed useful for the foreseeable future. The ncxt change may be another 25 years out. A significant obstacle to the success of, IPv6 is application transitioning. Although supporting IPv6 in new applications is relatively straightforward, realizing a dual v4iv6 capability for every old application is not. The volume of legacy applications and tools, as well as their life span, is just too big. Hence, the introduction of IPv6 as a ncw network layer will he a gradual process, lasting many years. On the other hand, there is no single killer application demonstrating the superiority of IPv6 from a service point of view. Advantages may become obvious only when looking at the largc scale (i.e., millions of nodes); they may never bccome convincing in small trials with a few hundred computers. Network management tools exploiting of new IPv6 capabilities in large networks arc one family of IPv6 applications despcrately, needed. Demonstrating the benefit of IPv6 for peer-to-peer applications like voice over IP or distributed gaming could stimulate user demand. However, as o f today little is happening in this domain, leading to a situation where IPv6 deployment is delayed duc t o lack of end user demand. Dcspite the strong support European and Asian govcrnments have given IPv6, little is happening in this area. Nevertheless, it is a fact that IPv6 provides important and useful changes that are critical to the long-standing success of IP packet technologies.