IPv4 vs IPv6 -- Why does this have to be so hard?
 

I was reading a news article on IPv4 depletion -- an issue far too few understand, in terms of dire consequences to the daily lives of anybody surround by networked and Internet-connected gadgets. What I fail to understand is why this turned into such a mess. I've seen computer software that does some pretty amazing automated translation tasks, so I don't exactly understand why IPv4 to IPv6 had to be so hard.

What is IPv6? What's IPv4?

A modern IP address comes in this form: 123.456.789.101
Or 1.1.1.1, or 22.22.22.22, or any combination of 1-digit, 2-digit, and 3-digit addresses. (1.22.333.4, for example)

IPv6 comes in hexadecimal, instead of numeric. And it's 16 characters long.
For example: A0B1:C2D3:E4F5:A6B7:0000:0000:0000:0000

I'm reminded of the Bank of America merger with NationsBank in 1999. All Bank of America 10-digit account numbers had "00" added as a prefix, to match the 12-digit length of NationsBank accounts. So an account of 1234567890 became 001234567890. Why can't this be done for IPv4 to IPv6 transition?

My suggestion:

Assign a prefix or suffix space, and dump all IPv4 there. Then any router request getting outdated numeric range requests can translate this easily, by attaching the proper prefix/suffix after hexadecimal conversion.

In the interim, use a reserved IPv4 range as a translation layer for IPv6 requests, so that IPv6 use can begin now. All IPv6 requests would be routed to this otherwise invalid IPv4 range. Yes, this may require a two-part request, numeric first for the IPv4 routers/switches, followed by passed IPv6 request (via tunneling requests).

At some point in time, have a cut-off date for datacenters, ISPs and other providers -- similar to the analog broadcast cut-off. IPv4 pre-requests would end, and IPv6-only requests would take place, with outdated IPv4 requests being translated.

Some of this seems to already be in place, or at least a variation thereof.

Design flaw?

I come across quotes like this: http://en.wikipedia.org/wiki/IPv6
To me, that sounds like a design flaw. Yes, in theory, less/smaller packets sounds great. But in reality, it's turned into little more than a technological boondoggle, from everything I've observed. Something that was already challenging was made harder. Granted, this is NOT my area of expertise.

Curious if anybody else has comments about this fiasco? :)

...because IPv4 only hosts can't communicate with IPv6 only hosts (and vice versa). Also there is a ton of hardware and software out there that is not IPv6 aware. Yes, IPv6 is a super set of IPv4. A device can have the same address in both protocols, but DNS is supposed to hide all that from the end user.

The tunneling solutions out there work. Which one is best depends on the application. The system you are describing sounds like the obsolete NAT-PT proposal. It would have allowed IPv4 only hardware and software to communicate with IPv6 only hosts using dynamically pooled IPv4 addresses.

It sounds very similar to the way the CDMA MEID system works on cell phones. (background: the pool of ESNs ran out, so longer MEIDs were created) When the phone encounters a non-MEID network it generates and uses a generated pseudo-ESN in a reserved range. There is risk of collision on the network (many MEIDs can hash to the same pESN), but its likely not going to happen.

To me, that just seems silly to create a "next generation" product that doesn't offer any kind of easy/simple reverse compatibility. I can understand the draw to all-new standards that free themselves of past flaws, but in reality it more often makes a mess. That's why computer operating systems are such spaghetti code, and imperfect in so many ways. But still, they work. That was the main reason OS X took so long to get off the ground, awaiting Rosetta emulation.

I can only guess you know more about this than I do. Networking at this scale is a little outside my current areas of expertise.

Do you think IPv6 is fine as is, or should something else have been done differently? :)

I also didn't realize IPv6 needed application/software support! See http://en.wikipedia.org/wiki/Compari...cation_support. And it appears MySQL is not IPv6 compliant? Uh-oh, that's not good long-term. I have a feeling the next decade is not going to be very fun for web developers and site owners. I'll be forced to learn something that I'd rather just avoid.