When archrivals Microsoft and Sun make a joint announcement, something important is clearly afoot. The news that the two companies are part of a consortium that is applying to ICANN to create a new top level domain for mobile devices is a case in point.
At first sight, the story is about convergence: the fact that more computers are portable these days, and that mobile phones now pack a considerable computing punch: contemporary mobile phones such as the Sony P800 are more powerful than the machine that ran Netcraft’s first Web Server Survey. But at another level, it is symptomatic of an even more profound change: a move from wired Internet connectivity centred on the users of a Net connection - companies or individuals - to a wireless Internet connectivity of objects, essentially independent of users.
The idea is not new: it is implicit in IPv6, the next generation of the Internet Protocol, which goes back to 1994. One of IPv6’s key changes is to increase the address space from 32 bits to 128 bits; this raises the total number of Internet addresses from 4,294,967,296 to a theoretical 340,282,366,920,938,463,463,374,607,431,768,211,456. As a useful background paper points out, this represents 665,570,793,348,866,943,898,599 addresses per square metre of the Earth’s surface. Allowing for inefficiencies in the address architecture, the actual number is likely to be between 1,564 and 3,911,873,538,269,506,102 addresses - more than enough to give an IP address to every kind of physical object at any point in the world.
The implementation of IPv6 has been proceeding rather slowly - largely thanks to the widespread adoption of Network Address Translation (NAT) as a stopgap measure to make up for the impending IPv4 address shortage. NAT has proved so good at solving the short-term problem that the longer-term benefits of IPv6 have been postponed.
An alternative approach to creating an “Internet of things” has been developed under the auspices of a group known collectively as the Auto-ID Labs, whose goal is summarised by its slogan: “identify any object anywhere automatically.” That identification will be achieved by drawing on three technologies: the Internet, Radio Frequency Identification (RFID) tags, and Electronic Product Codes (EPC).
RFID tags have a surprisingly long history, but are only now starting to become widespread as a way of tracking goods from manufacturer to distributor to retailer to customer. EPCs are generalised versions of the Universal Product Codes that are used in barcodes, but with the important difference that they can identify each product uniquely. By hooking the RFIDs encoded with EPCs into the Internet, details about any object can be retrieved from online databases while it is in transit.
The industry body behind these moves is called EPCglobal. On its site, there is information about the network, a FAQ and a page of links to technical specifications. These include the EPC tag data specification; the Object Name Service that translates EPCs into one or more URLs where further information on the object in question may be found, which are called EPC Information Services; and the Savant middleware, which carries out the processing using something called the Physical Markup Language (PML).
At the beginning of the year, EPCglobal selected VeriSign to provide the root server of the EPCglobal network. The home page of VeriSign’s EPC Network Services has links to an overview, how it all works in practice, and a white paper.
What is particularly significant about VeriSign’s involvement is the fact that it already occupies a central position in the current Internet landscape, running the registry for the .com and .net domains. Since the new Net of objects will be orders of magnitude bigger - VeriSign speaks of hundreds of billions of lookups every day - and arguably even more important in terms of its role in the global economy - its rise is likely to consolidate the company’s pivotal role in the world of domain names yet further.
Glyn Moody welcomes your comments.