Why Bosch rigs up IPv6?
Bosch prepares for the next version of the internet, Web 3.0: The Internet of Things and Services is a smarter web, enabling better ways to share information not only for computers, but also for even the most common things of our daily life. Predicted by technology evangelists and market researchers as well, the Internet of Things and Services will have large impact on us, society, and systems. In the coming years, more and more intelligent systems will be able to use the internet to communicate automatically with each other. In just 15 years, the resulting Internet of Things and Services will interconnect more than 50 billion components – from tiny sensors to high performance computers.
Many applications in different domains will be coming up like:
- eHealth: The future of healthcare lies in creating tighter connections between physicians, hospitals, rescue centers and patients by using modern technologies.
- eMobility: To make electric vehicles widely accepted, it will take a sophisticated charging infrastructure, an intelligent system to orchestrate this new industry with vehicles, charging service providers, and utilities.
- eProduction: In the efficient industrial production of tomorrow, decisions will be made real-time derived from event-driven supply chains and diagnoses.
- eEnergy: Energy Management targets to optimize energy demand and consumption in different ways, at utilities, at industrial, at communal or at private consumers.
Besides all the enthusiasm about the new possibilities of the next internet there is some technical homework left before this becomes a reality:
- Current address space of IPv4 is running out of free numbers (“house numbers” in the internet)
- Standard Security level of the internet needs an upgrade
- IP connectivity of devices is not common yet
- Standard interfaces to talk to devices are still missing
IPv6 in numbers
IPv6 addresses the first two issues. IPv4 has only 232 roughly 4*109 addresses whereas IPv6 has 2128 roughly 3*1038 addresses – both big numbers. Setting a relationship to people or addresses per 1 km² on the planet earth it becomes very clear that we need an extension of the address space for the Internet of Things & Services.
In 1980 the year IPv4 was released to the world there was about 1 IP address per living person on earth. According to the projection of the UN the world population for 2050 will be 9.1 billion people leaving only every second person one IPv4 address. Whereas IPv6 would allow 4.9*1028 for every person today – more than enough to run a computer, smart phone, home entertainment, and multiple Internet of Things devices.
Looking on the distribution of IPv4 addresses on the surface one can easily recognize that only eight IPv4 addresses per square km is not really enough to run infrastructures today. For example one square km in highly populated Singapore with 7,315 people/km also “runs” today 10,990 phone contracts and 1,372 vehicles. IPv6 allows about 6.7*1029 addresses.
Addressing end-to-end connectivity & privacy
Today the world relies on network address translation (NAT) in different flavors to open up local IPv4 address spaces. This strongly violates one of the key success factors and internet architecture principles: End-to-end connectivity. The end-to-end connectivity has many technical, organizational, and even political implications as it avoids the dependency of single internet participants on a central NAT instance. Many privacy activities argue that NAT allows protecting the identity of internet participants. This is not true as the knowledge who is behind the address is still known by the NAT operator. If one really tries to achieve anonymity in the internet different mechanisms like “The Onion Router” (TOR)“, “Java Anon Proxy (JAP)” or RFC 3041 are required. The huge IPv6 address space even allows completely new mechanisms for privacy like changing dynamic IP addresses which are public-private-key encrypted between the two peers.
*) Robert L. Mitchell (December 28, 2009). “Y2K: The good, the bad and the crazy”.ComputerWorld
The IPv6 migration has some similarities to the Millennium or Y2K bug. Most organization have a reactive business attitude to IT best explained with the mantra „if it ain’t broke, don’t fix it”. The Y2K touched every part of the organization worldwide which is also true for IPv6. The Y2K cost about $308 billion worldwide*) and we have to assume the IPv6 transition costs are not that much different. The main difference is though that Y2K had a “day of truth” like 01.01.2000 and no single company wanted to fail. IPv4 address space problem cannot be assigned to one single company or person. This removes the pressure to act.
Bosch and IPv6
Bosch is preparing its devices and networks to run with IPv6. This is because the address space of 128 bit is a pure necessity for the Internet of Things and Services, Sensor networks, and interconnected devices. Additional reasons for IPv6 are mobile IP support, auto configuration techniques, and build in security features of IPSec. For the time until IPv4 is completely replaced by IPv6 Bosch will use special techniques like dual-stacking, tunneling or translation to allow interoperability. Bosch assumes that this will take about 5-10 years. One example is the standard, currently in CD stage, for electromobility “ISO 15118 Road vehicles — Vehicle to grid communication interface” which relies only on IPv6.
I am looking forward to your comments. Get on board!