2.2 Defining the Space

---

Name and number spaces are everywhere in our technology-saturated environment. Bank ATM cards and credit cards all have numbers assigned to them that must be unique within their particular technological system. Postal codes carve up countries into distinct, mutually exclusive regions. Bar codes in grocery stores are assigned to specific products. Books have their own international numbering standard (ISBN). Almost every durable good we buy has a unique serial number that is part of a number space defined by the manufacturer. The rise of the Internet and the digitization of all forms of information have fomented a great deal of research and experimentation on new ways of naming or identifying information content (Green and Bide 1997).

Depending on the technological, economic, and organizational circumstances, defining an address space can be very simple or very complex. Imagine a simple number space that starts with 1 and goes on to infinity. The first applicant would get the number 1, the next would be assigned the number 2, and so on indefinitely. Such a space would work like one of the 'take a number' machines at a crowded delicatessen but with an infinitely large roll of tickets. Such an address space architecture makes it easy to assign values but imposes other costs. A few lucky people would get short, memorable, easy-to-use identifiers; those who came later would get increasingly long, unwieldy ones. In this hypothetical system, the identifier assigned to individuals would not yield information that was useful in running a communication network. All it would tell us is the particular sequence in which people received identifiers. It would tell us nothing about where they were located or how they might communicate with other people on the network. It would also make it difficult for computers or other automated methods to process such addresses efficiently, because they would never know exactly how long the number would be.

The hypothetical example is intended to illustrate some of the choices that must be made in defining a name or address space. Should the unique name or address merely identify an item, as a serial number does, or should it locate the item, as a Web URL (Uniform Resource Locator) or a telephone number does? Or should it try to do both? Should the address space be flat or hierarchical? A flat space may have difficulty adjusting to rapid growth, but a hierarchical space may impose limits on the mobility of the addressed objects and lead to less efficient use of the space. Should the address be purely arbitrary, or should it embed some intuitively accessible information about the object? There are operational advantages and disadvantages either way.

Table 2.1 provides a summary of some common name or address spaces and their basic features.

Table 2.1: Summary of Common Name/Address Spaces and Their Features

Name	Owner/Root Administrator	Purpose	Capacity	Architecture
E-164	International Telecommunication Union (ITU), Geneva	To coordinate international telephone dialing	~1,110 country codes; permits 1 billion to 1 trillion national numbers per country	Recommends number prefixes for international calls (00) and domestic toll calls (0); assigns unique country codes (1-3 digits); fixes maximum digits for international numbers (15 digits, excl. international prefix)
'Handles' (Digital Object Identifiers)	Corporation for National Research Initiatives (CNRI), Virginia	To provide persistent unique identifiers for digital objects	No design limits on number of prefixes or suffixes	Two-part hierarchy: a prefix assigned by naming authority and a suffix created by user, separated by a slash '/'; separates location from identification to achieve permanent identifiers
Ethernet (EUI-64)	Institute of Electrical and Electronics Engineers (IEEE), Piscataway, New Jersey	To assign unique addresses for Ethernet Network Interface Cards (NICs)	16 million OUIs (Organizational Unique Identifiers); 1 trillion unique values per OUI	Two-part hierarchy: a 24-bit OUI and a 40-bit Ethernet Unique Identifier (EUI)
ISBN (International Standard Book Number)	International ISBN Agency, State Library, Berlin	To make processing and handling of books more efficient for publishers and booksellers	Number of unique IDs available depends on how much space is consumed by higher-level identifiers. Group ID max = 5 digits; publisher ID max = 7 digits; title ID max = 6 digits	10-digit number divided into 4 parts (separated by spaces or dashes) representing codes for group ID, publisher ID, title ID, and a check digit for error control Convertible into optical bar codes

The structure of an identifier can be compared to a language that the network uses to talk to itself. The switches, routers, or other machinery on a network can 'read' it to better handle the movement of information. A telephone number in North America, for example, has a syntax based on geography or function and the switching hierarchy. The number starts with a three-digit area code associated with a geographical region or special function. If the area code is 800, for example, the user knows that it is a toll-free call and the network knows to which database to go to find out how to connect the call. The area code is followed by a three-digit exchange number and a four-digit line number. The structure plays a vital role in telling the network how to route phone calls.

---