Jeff Kabachinski

Hmm—I wonder if this is a looming network catastrophe or another Y2K ho-hum event? The issue: We are quickly running out of Internet address space. I am not surprised, in that we turn on more than 200 million new Internet services every year and add new users and mobile devices at a mind-boggling rate. For example, Asia accounts for nearly two thirds of the address consumption. Then consider the populations of China and India that are technically astute with the proliferation of cell phones and social media. Just a few of the reasons why we have used up nearly all of the 4.4 billion addresses that came with IPv4.

IPv6 is the next iteration of the protocol that runs the Internet. Will the upcoming migration to IPv6 disrupt clinical computer networks and the delivery of health care? This is not an issue of if to convert, but when and how. The current version—IPv4—has been in place for 20 years. While IPv4 has taught us a lot about what works and what needs improvement, it is pretty amazing that a protocol has endured for that long in the fast-paced IT environment! It is also amazing that a fairly simple set of protocols is still viable after all this time. I think IT time is like dog years—every year in real time is 7 years in IT time. That means IPv4 has been in operation for 140 years! If it were not for the limiting address space, who knows how long it would be before a major upgrade would be needed? Again, the depletion of IPv4 addresses is the main reason for the need for change. In November 2010, the Internet Assigned Numbers Authority (IANA) allocated another two groups of eight (Class A) IP address blocks, leaving a mere seven remaining in the pool. However, in reality there are really only two that can be allocated in the “normal” way. Because of global policy, when they get down to five they must allocate those to the five Regional Internet Registry (RIR). The RIRs have also hastened consumption by preallocating address space in anticipation. However, IPv4 addresses will be around for a while, but like any other commodity, from whom and at what price? Latest estimates are that we will exhaust the IPv4 address pool as early as February 2011.

IPv6 Address Format

IPv4 addresses are very familiar. They’re 32-bit (four-byte) addresses—written in decimal format—with each byte delimited by a period (decimal point or “dot”). Consider 120.23.68.244 as an example. On the other hand, IPv6 addresses are 128 bits or 16 bytes in length written in hexadecimal with each four-digit hex integer delimited by a colon. For example: 67cf:9:2100:f2ef:241:2ffe:cb83:64ad. Now we’re getting someplace! This example address is called the preferred format where leading zeros in each hex integer can be left out to shorten it a bit. In the example address, note that the 9 integer represent 0009 and 241 equals 0241. In the collapsed format if the integer equals zero it can be left out entirely. Therefore, the IPv6 address of 260B:0000:0000:0000:005F:1C00:800D:528B becomes 260B::5F:1C00:800D:52.

This expanded address space will allow for 2128 numbers of addresses. That’s 18 quintillion.1 I saw one analogy that compared this number to grains of sand. It said that if each IPv6 address were a single gain of sand then the entire IPv6 address space would create ~300 million planets each the size of the earth! OK—yeah—that really put it in perspective. Too big to think about. In real life the actual usable address space is probably 250 or 260, or only 150 million planets in our analogy. Bottom line is that IPv4 was originally planned to be able to provide addresses for 15 years; with IPv6 the claim is 50.

The trouble with IPv4 is larger than just the address limitation. IPv6 offers several new methods to also deal with security and off-loading network services to the protocol itself. For example, Internet routers will not need to keep track of so many details any longer when the routing is embedded into the globally routable IPv6 address. With things like cloud computing becoming popular, the traffic will increase (an understatement if I ever made one!). TCP port parallelism (simultaneously streaming a single service’s data to many ports on one device) is another bandwidth hog when each packet needs to be switched through the network. Consider every Internet packet that must go through a router.

The obvious thing is that we are facing a major IP upgrade that will affect network operations. How much of an upgrade is to be determined. While IPv6 is a substantially different language from IPv4, we will want all the improvements that IPv6 brings. The main sticking point in this upgrade is that IPv6 is not backward compatible (you’d think Microsoft was in charge here). So there will be a need to run both protocols simultaneously, called dual stacking, maybe for years. Dual stacking is not easy or straightforward. There will be other options to convert but due to protocol differences probably not automated translation to IPv6. Also, any IPv4 device (like a phone) that is currently happily talking to another IPv4 device may be best left alone—at least for now. Be sure to check the sidebar below for tips on what you can do right now to prepare.

Reference

  1. 18 zeros = quintillion (in US). You would think that quintillion (quint = 5) is 5 times a million or 30 zeros (a million has 6 zeros so 5x is 30)—it is in Britain! But in the US quintillion comes after quadrillion (15 zeros). In GB, it makes more sense. For example: In Britain, quadrillion is 4 times a million (24 zeros = 6 zeros in a million X 4). In the US, Britain’s quadrillion is a septillion, see?

Jeff Kabachinski, MS-T, BS-ETE, MCNE, has more than 20 years of experience as an organizational development and training professional. Visit his Web site at kabachinski.vpweb.com. For more information, contact .

What To Do Now

It is difficult to give attention to something that may not directly affect you for a year or two, especially with electronic health record and meaningful use activity already adding to the fray. At this point, maybe all that is needed to stay ahead of the game is to set up an IPv6 watch group or task force in your organization. It is always better to have several sets of eyes and brains working together, and also to divide and conquer. A cross-functional group that meets monthly to discuss implications would be a great joint effort from IT and biomedical/clinical engineering to engage in. Things to keep an eye on:

  • Vender preparedness. What devices will need to be IPv6 compatible? Do they need replacement, or can they be upgraded? What is the cost in materials and labor to make the replacements or upgrades?
  • Staff preparedness. What do you need to know about it? Who needs what training and by when?
  • Organization preparedness. What is on the horizon that we might need to address? In other words, what Twitter-like bandwidth-consuming app is sneaking up on us? Will remote-based personal health care substantially add to the data glut? (Or “torrents of raging data,” I like to say.)

Use the IPv6 watch group concept to stay alert to IPv6 activity. For example, many broadband Internet service providers and network hardware and software vendors have been sluggish in upgrading. Your network performance, connectivity, and security could be affected, so it is worthwhile to keep an eye on this one!

—JK