By Art Reisman
Editor’s note: Art Reisman is the CTO of APconnections. APconnections designs and manufactures the popular NetEqualizer bandwidth shaper. APconnections removed all Deep Packet Inspection technology from their NetEqualizer product over two years ago.
First off, let me admit my track record is not that stellar when it comes to predicting the timing of eminent technology changes.
In 1943, Thomas Watson, the chairman of IBM forecast a world market for “maybe only five computers.” Years before IBM launched the personal computer in 1981, Xerox had already successfully designed and used PCs internally… but decided to concentrate on the production of photocopiers. Even Ken Olson, founder of Digital Equipment Corporation, said in 1977, “There is no reason anyone would want a computer in their home” (read about other predictions that missed the mark).
As a young computer scientist 1984-ish, I would often get questions from friends on whether they needed a personal computer. I was on the same bandwagon as Ken Olsen, telling anybody that asked — my dentist, my in-laws, random strangers in the park — that it was absurd to think the average person would ever need a PC.
I did learn from my mistake and now simply understand that I really just suck at predicting consumer trends.
However, while the adoption of the personal computer was a private consumer-driven phenomenon, IPv6, on the other hand, is not a consumer issue. And, my track record as an innovator of technology for business is much better. My years of guiding engineering decisions in Bell Labs, and now running my own technology company, provide a good base for understanding the headwinds facing IPv6.
Since the transition to IPv6 is not a consumer adoption issue, it has many more parallels to the Y2K scare than the iPod. But, even then there are major differences.
Y2K had a time bomb of deadline. You could choose to ignore it, but most IT managers could not afford to be wrong, so they were played by their vendors with expensive upgrades.
My prediction is that we will not transition to IPV6 this century, and if we attempt such a change, there will be utter chaos and mayhem to the point that we will have to revert back to IPv4.
Here’s my argument:
- There is no formal central control for certification of Internet equipment. Yes, manufactures are self-proclaiming readiness, but even if they all do a relatively good and professional job of testing — even with a 99 percent accuracy — on switchover day, the day everybody starts using IPV6 address space, the cumulative errors from traffic getting lost, delayed, or bounced from the one percent of equipment with problems will bring the Internet to its knees. I don’t think the world will sit around for a few weeks or even months without the Internet while millions of pieces of routing equipment from thousands of manufacturers are retrofitted with upgrades.
- There’s no precedence. The only close precedent for changing the Internet address space would be the last time when AT&T added an extra digits to the dialing plan. At the time they controlled everything from end to end. They also had only one mission , and that was to complete a circuit from A to B. Internet routers, other than in the main backbone, do all kinds of auxiliary functions today such as firewalls, Web filtering, and optimization, hence further distancing themselves from any previous precedence.
- We have a viable workaround. Although a bit cumbersome, organizations and ISPs have been making due with a limited public address space using NetWork Address Translation for more than 10 years already. NAT can expand one Internet address into thousands. Yes, public IP addresses for every man woman child for earth and every other planet in the Milky Way is possible with IPV6, but for the forseeable future, NAT combined with the 4 billion addresses available in IPv4 should do the trick, especially given the insurmountable difficulty with a switchover.
- Phased Switchover nonsense ? The pundits of moving to IPv6 are touting a phased switchover. I am not sure what this accomplishes . If one set of users starts using a larger address range, for example, the Indian Government, they will still need to keep their original address range in order to communicate with the rest of the world. To realize the benefits of IPV6, the world as whole, will need 100 percent participation. Phased switchover by a segment of users, will only benefit vendors selling equipment.
Despite these predictions, the NetEqualier is ready for IPv6. We have already done some preliminary validation on IPv6 implementation in our NetEqualizer. In fact, we have even run on networks with IPv6 traffic without issues. While we have some work to do to make our product fully functional, we’ve already sufficiently tested enough to have confidence that if and when the IPv6 switch over happens, we will not cause any issues.