Do we really need IPv6? detractors say IPv4 address depletion is overblown; boosters point to other benefits

America's Network, Nov 1, 2003 by John (American congressional representative) Tanner

One of the traditional drivers of IPv6 adoption has always been the lack of sufficient address space in today's IPv4 in the upcoming world of ubiquitous IP-enabled, always-on devices. However, some critics of v6 have claimed that this is all v6 really has to offer in terms of technology, and that even the address shortage issue is overrated.

Geoff Huston, the chief scientist of Australian telco Telstra, became the center of controversy late last year after publishing a study that estimated that the remaining unallocated v4 address field wouldn't be exhausted for at least another 20 years--possibly as late as 2028 or 2029.

Although Huston--an otherwise respected authority on IPv6--has said himself that any number of factors could alter that projection, the study drew plenty of fire from v6 advocates.

"Huston's figures are just numerical, not considering changes in applications, trends, policy, deployment of 3G, etc." says IPv6 Forum president Latif Ladid.

Jim Bound, chairman of the North American IPv6 Task Force and the IPv6 Forum Technical Directorate, adds that Huston's metrics did not include the H-Ratio base. The H-ratio base is a concept that presumes that since no network ever uses every single available address, the exhaustion level for usable v4 addresses would occur well before the protocol literally ran out of numerical combinations. Between the H-ratio and the failure to account for factors such as the likelihood that households, business nodes and individuals will probably need at least five IP addresses, the usable v4 field will be depleted far sooner than Huston estimates, Bound says.

Bound also disputes Huston's assertion that v6 brings no meaningful technology advantages to the table. Users will not see or feel IP directly, but they will benefit from the technology advantages it will bring, such as enhanced mobility applications.

Bound is especially critical of NAT (network address translation), which is typically used with Ipv4 to dynamically allocate addresses to individual devices sitting behind a router in a LAN. Whatever NAT accomplishes in terms of address availability, it loses in delivering certain security services, such as end-to-end encryption, and supporting remote users on large networks.

In fact, says Bound, implementing NAT opens all kinds of security weaknesses in otherwise secure LANs, and interoperability isn't guaranteed when communicating with peers or applications outside of the site. These aren't insurmountable issues, but dealing with them takes time, effort and money.

Bound does concede that IPv6 has no QoS advantage, but says that's only until the IETF and industry define the standard for the flow label field in the header. "IPv6 is a means to identify traffic flows with the source and destination address," he says. "IPv4 cannot do that at all."

All-fiber networks have been a goal of the telecommunications industry since the late 1980s, but with the latest boom of interest in Passive Optical Networking, the day when fiber comes home seems almost at hand.

The breakdown: IPv6

The Benefits                       The Hindrances

Virtually unlimited IP addresses  Lack of applications
End-to-end security               How fast is v4 really being depleted?
QoS (pending IETF standard)       Maturity phase not there yet
No more NAT                       NAT as an interim fix