热度 24
2015-7-16 21:17
1984 次阅读|
0 个评论
The American Registry for Internet Numbers (ARIN) has just sent out an alert that its share of the unique URL addresses for the Internet Protocol, Version 4 (IPv4) is running out. Since only a few of four billion or so unique URL addresses possible with IPv4 are left, ARIN has announced that there’s not enough in its stock to satisfy an entire order. So it has activated a stricter procedure by which to assign the few remaining IPv4 addresses. Specifically, in its recently announced policy for unmet requests , when an organization qualifies for a block size that no longer remains in the ARIN IPv4 inventory, they'll be given the option to either accept a smaller block that is available to fully satisfy their request, or to be placed on the waiting list for unmet requests. “The number of days remaining before depletion are dwindling," writes ARIN's chief information officer Richard Jimmerson in a blog post just before July 4. "It is very likely that we are already processing a request that we will be unable to fulfill." An article in the Washington Post on the ARIN announcement seems to imply that major changes are imminent, but that we are not to worry because IPv6 is available to save us and give us all the URL addresses we need. Only the second part of that is true. Except for many mobile smartphone and tablet users whose wireless service providers are just now making the shift, the transition to IPv6 will take years and even decades. Many existing Internet Service Providers with IPv4 will be around for a long time. Indeed, there is already a growing market for buying used and highly desirable IPv4 addresses . IPv6 with its 340 trillion trillion trillion possible unique combinations has been sitting in the wings as a replacement for IPv4 since about 2000. But few if any large organizations with the bankroll to establish a presence on the Internet have felt it was economically viable to invest in it until recently. It was not until three to four years ago that major Internet and Web entities under the umbrella of the Internet Society finally started making the transition. Many of those most active in making the shift are in mobile smartphone markets, including ATT, Google, T-Mobile, and Verizon. Most of them are still only half-way through their deployment and testing on IPv6. And where available, it is being offered only as an optional alternative to IPv4. Currently, only about seven percent of users who come to Google do so over IPv6 connections. (Source: Google) Even with the number of its remaining unique addresses drying up, IPv4 still accounts for 93 percent of worldwide Internet traffic. According to a continuously updated chart maintained by Google, as of July 3, 2015, the percentage of users worldwide who access Google via ISPs with IPv6 is currently about seven percent. Google's per country adoption charts indicate that in the United States just slightly over 20 percent of the users who come to Google do so via IPv6 connections. The reason there is not a rush on IPv6 is simple: economics. Most local, regional, and in some cases national Internet Service Providers are not able or are unwilling to pay the expense of transitioning from their existing base of IPv4-based routers, switches, and servers, except on a slow and years-long incremental basis. In the United States, only the largest of ISPs have committed to the transition, including Charter Communications, Comcast, Global Crossing, Hurricane Electric, Liberty Global, and Time-Warner Cable. IPv6 will slowly eat away at the continued IPv4 dominance, due mostly to new network providers and companies with no investment in existing IPv4. But most second and third tier regional, statewide, and local ISPs who have major investments in IPv4 are not rushing to make the shift. Engineers and technicians at the several regional ISPs I have dealt with directly over the last decade or so point out that there is no compelling end-use application that cannot be done with existing IPv4. Since about 2000, through the use of such traditional techniques as the use of several levels of subdomains, dynamic rather than static network address translation (NAT) , destination and stateful NAT, NAT loopbacks, port address translation, and Internet connection sharing, they have been able to keep up with demands, not only for more bandwidth, but more features and flexibility. Where such enhancements and workarounds put additional load on their network hardware, IPv4 Internet Service providers are shifting to the use of switches and routers based on more powerful and lower cost multicore processor designs. Also aiding their efforts at improving IPv4 is the shift to systems based on software defined networking (SDN) . Such routers and switches depend not on dedicated hardware for separate network functions, but on software-based network function virtualization (NFV) allowing lightning -fast reconfiguration of a variety of network elements. While such enhancements of existing IPv4-based systems involve additional investment in hardware and software, this can be done at a lower cost, and on a more piecemeal, as needed, basis rather than by replacing their existing IPv4 systems with IPv6. So, despite the fewer unique URL address out of the four billion or so made possible by IPv4, the end of the Internet has we have known it will not occur soon. IPv4 will continue to be the norm for many decades to come. In the U.S. that transition will occur more quickly only if one of three things happen: First, a dramatic use case for IPv6 emerges that triggers a demand from users of the Internet, causing IPv4-based ISP companies to make the shift; second, government action is taken either in the form of significant incentives or through a direct statutory requirement; and third, the economics of maintaining IPv4 becomes unviable. Nothing I see now or in the near future makes any of these likely any time soon.
标签: IPv4
