China disowns IPv9 hype

06-07-2004

by The Register

Evidence is growing that IPv9, hyped as the widely-adopted foundation of a next generation Internet infrastructure in China, is really a marginal project backed by few even in China.

Reports from China this week about widespread adoption of the previously unheard of Internet protocol have created bewilderment and something approaching a diplomatic incident in the system administrator community.

Vint Cerf, SVP of technology strategy at MCI, and one of chief architects of the modern Internet, was bewildered by the reports. In an e-mail sent to senior figures in the Chinese Internet community, he asked: "What could this possibly be about? As far as I know, IANA (Internet Assigned Numbers Authority) has not allocated the IPv9 designation to anyone. IPv9 is not an Internet standard. Could you please explain what is intended here? I am disturbed by the reference to root servers, 'control.' What is the 'ten digit text file' all about? Who is behind the Shanghai Jiuyao Digital Network Company?"

Professor Hualin Qian of the Computer Network Information Center of the Chinese Academy of Sciences described IPv9 as a research project that turned out to have serious practical shortcomings and little support.

"CNIC explains IPv9 is proposed by the director, Mr. Xie Jian-Ping, of the Institute of Chemical Engineering located in Shanghai. Two years ago, Chinese Academy of Sciences (CAS) invited them to introduce their idea about IPv9. According to my understanding, their proposal includes two main aspects: the first one is IPv9, the second one is Digital Domain Names," Professor Hualin wrote.

"For IPv9, they think that the address space of IPv6 (128bits long) is not enough for future use, they expanded the IP address to 256 bits. I don't think the protocols for IPv9 have major difference from IPv6 except the longer IP address. Almost all the people working on networks in CAS do not agree with their opinion, because there is not any evidence showing that the IPv6 address is not enough and using 256 bits source and destination IP address will increase the overhead of an IP packet. And when communicating with IPv4/IPv6, equipment such as NAT-PT [Network Address Translation] must be installed. This will be the bottle neck for future high capacity interconnection with IPv4 and IPv6 global Internet."

Hualin added that IPv9 is unfamiliar to network experts from Fudan University in Shanghai who "do not know any deployment of IPv9 in Shanghai" contrary to initial reports by China's official news agency, Xinhua.

Tim Chown of Southampton University, and technology adviser to the IPv6 Task Force in the UK, told the Register: "The consensus now seems to be it is one researcher or group trying to promote a 256-bit adaptation of IPv6, but it doesn't yet seem to have much traction. It is hard to tell how serious it is, or whether it is a complete non-starter in the same way as Jim Fleming's ludicrous IPv8 is. There may well be some sensible ideas behind IPv9, but IPv6 is the system that is standardised and now (very) widely implemented."

The Register and its contents are copyright 2003 Situation Publishing. Reprinted with permission.