A story published Monday says IANA will be out of addresses by the end of this week. The message comes from the IDG News Service via Computerworld.
This isn’t exactly breaking news, although it’s hard to believe the long-predicted address exhaustion is finally upon us. As of this writing (Tuesday, Jan 25), the exhaustion counter to the right says we’ve got 7 more days and about 29 million unique addresses to go. Alas, the story is a lot more complicated than that. For one thing, the five RIRs, which get their address blocks from IANA, have some stockpiled. Still, tensions are mounting…
As luck would have it, I just sent in a version of the story as my monthly “comment” for Telemanagement, due to be published in a few days. See below…
Is There An IPv6 Crisis?
In case you hadn’t heard, IPv6 is coming. But it’s not coming quietly. To paraphrase that old adage about psychotherapy: IPv6 will set us free, but first it’s going to make us miserable.
IPv4 has been in continuous use on the public Internet for the last three decades. Today we’re scrambling to move to IPv6. News of the transition seems to evoke three general reactions.
One group knows we’re very quickly running out of IPv4 address space, which allows for a mere 4.3 billion addresses, and they’re plenty worried. At the forefront of this group are the organizations that keep the Internet running, including the Internet Society (ISOC) and the American Registry for Internet Numbers (ARIN), one of the world’s five regional Internet registries (RIRs). They’re proselytizing for IPv6 through workshops, presentations and social media. And just this month, ISOC announced that June 8, 2011 will be World IPv6 Day. Some of the biggest online players – like Google, Facebook and Yahoo! – have agreed to spend 24 hours testing their content on the public Internet using IPv6 technology.
Another group comprises the many ISPs, governments, Web publishers and other businesses whose operations will soon face disruptions resulting from the v4-v6 transition, or lack of it. In Canada, it appears that most organizations – including the federal government – are standing pat. For private-sector firms, the transition is likely to be costly and messy. It also places them in something resembling the prisoner’s dilemma. They don’t want to make any commitments until they see what their competitors are doing, and neither do their competitors.
Then there’s everybody else.
In my discussions with mere mortals, including students, it’s hard to generate anxiety about running out of v4 addresses or excitement over its big shiny replacement. The numbers involved are awesomely big – I mean, like, awe-inspiring. So big they defy comprehension.
Consider the 39-digit string at the top of this article. It represents the theoretical number of unique addresses created by IPv6, which can be written more compactly, and professionally, as 2^128 or about 3.4*10^38. That’s around 340 trillion, trillion, trillion addresses. Let’s translate these figures into slightly more familiar terms.
Planet Earth is 4.5 billion years old. If we’d been assigning IPv6 addresses at the rate of one billion per second since Earth’s formation, we would by now have used up less than one trillionth of the IPv6 address space. In principle, we’ll eventually have 5*10^28 addresses for every one of the nearly 7 billion people on the planet. We’re not likely to be worrying about another round of address exhaustion any time soon.
That said, we need to keep a few other considerations in mind. First, IPv6 wasn’t created simply for its sheer scope. The IETF, which has been working on v6 for 15 years or so, has incorporated many other improvements: in address allocation, traffic routing, network renumbering, mobility support and a comprehensive security protocol suite known as IPsec, to name a few.
Moreover, IP addresses are not intended for individuals, but for the hundreds of different kinds of devices people use now and will use in the future – a list that’s growing rapidly and includes once unlikely suspects like domestic appliances.
At the recent Consumer Electronics Show, by way of illustration, Samsung and LG were engaged in a battle of connected fridges. The WiFi-enabled Samsung allows you to stream music from Pandora and organize your life in the Google calendar; the company rep was especially enthused about how handy this fridge is for sending tweets. The competing LG has an interactive display that allows you to inventory your food supply, then check it via handheld from the grocery store. You can interact with this appliance using drag-and-drop, or share your culinary thoughts by talking to it.
These Jetson-like scenarios are now bumping up against a cruel reality: no IP address, no calling your fridge from Loblaws. IPv4 exhaustion actually began years ago, although the IETF developed a number of awkward if effective technical workarounds like network address translation (NAT) to postpone the fated day.
And we’re still not done with the workarounds, because IPv6 has one big structural problem: it’s not interoperable with IPv4.
Can we get there from here?
Once I started looking for some hard numbers about v4 exhaustion, I quickly discovered that the numbers are, well, hard to pin down. By late December, the consensus was we were hovering around 2% of the available v4 pool and dropping rapidly. What does that represent in unallocated addresses?
As Iljitsch van Beijnum explained in Ars Technica, the individual address counts are complicated. The master list is maintained by IANA, the Internet Assigned Numbers Authority, which works under ICANN. The address totals are made up of what’s allocated to each of the five RIRs, which all have different inventory profiles. The counts include addresses that are reserved, and others that have special uses or are hidden away in legacy stockpiles.
To provide a visual, though not necessarily accurate, sense of this technological drama, a handful of organizations have made IPv4 “exhaustion counters” available for download. The one you see on this page comes courtesy of Hurricane Electric, a backbone provider that operates the world’s largest IPv6 network. On January 22, it showed a little over 41 million unallocated v4 addresses. A day later, the counter had moved down to 39 million and change.
As part of its awareness program, ISOC also provides an IPv6 readiness test. I took the plunge, more out of curiosity than hope for an upbeat result. The good news? My home connection (from reseller National Capital Freenet) scored 7 out of 10 for my “IPv4 stability and readiness, when publishers offer both IPv4 and IPv6.” But the one that matters is “your IPv6 stability and readiness, when publishers are forced to go IPv6 only.” Sigh, zero out of ten on that one.
ISOC assured me that for World IPv6 Day, “No problems are anticipated for you with this browser [Chrome], at this location.” They go on to add this sympathetic message:
“Don’t feel bad – most people are in this position right now. Most internet service providers are not quite yet ready to provide IPv6 internet to residential customers. The good news is, nothing will suddenly break.”
ISOC’s gentle touch may be a nice way to treat worried end-users. But the suggestion that most ISPs aren’t “quite ready” for IPv6 is highly optimistic.
In Canada, the only ISP that’s conducting a public beta on IPv6 is reseller TekSavvy (call for beta testers in DSLReports.com forums posted here February 2010). It’s hardly surprising that this initiative comes from CEO Rocky Gaudrault, who is well known and admired in Net neutrality circles for his pro-consumer advocacy. Among the incumbents, I picked Bell for evidence of IPv6 activity at the retail level. The company may be experimenting behind closed doors, but a search of both the Bell and BCE websites for IPv6 turns up no hits.
This finding is consistent with what other analysts are saying about the deployment of v6 in the backbone market. Network engineer Andree Toonk, writing in early January on his BGPmon blog (The State of IPv6 in Canada), notes that Canada’s wholesale carriers may have already ceded most of the market for IPv6 transit to the (non-Canadian) Tier 1 carriers. This lost opportunity may turn out to be a crisis all on its own.
This story continues…