Internetevolution.PDF

Internet evolution Dah Ming Chiu, Chinese University of Hong Kong Abstract Can Internet evolve, and where is it evolving towards? This article gives a plausible model for Internet evolution, and tries to explain its implications for different stakeholders. To understand the Internet, it is not enough to only describe what it is, but also guess at how it is evolving. There seems to be less agreement about the latter than the former. This short article tries to provide one viewpoint for discussion. How does and will Internet evolve? Despite the tremendous success of the Internet, almost all agree that as new applications and services are developed, it needs to evolve to address new demands and requirements. One way Internet evolves is through planned transition. An example of planned transition is from IPv4 to IPv6. The original Internet protocol (IPv4) was designed to use 32 bit address, which proved to be inadequate given Internet’s big success. After a non-trivial amount of effort, IPv6 was standardized, changing IP address from 32 bits to 128 bits, plus some other minor tweaks to the protocol. The ensuing transition from IPv4 to IPv6, however, took a long time (and is still on- going). The slow process of transition necessitated a whole suite of interim solutions to make sure IPv4 and IPv6 can co-exist and interoperate, which is more complicated than the new protocol itself! Some causes for the slow transition may be due to the introduction of a network address translation (NAT) mechanism that allowed non-public addresses to be used, relieving the most pressing issue with 32 bit address; while another reason may be the lack of incentive for existing network users to upgrade. This led to discussions of Internet being ossified (or unable to evolve), and how to get out of that. A notable paper called for Internet diversification [1], proposing to let different architectures (resource management policies) to be operated and co-exist over the s