6. Size of the Internet
6.1 Computers on the Internet (Updated)
The data presented here presents the number of hosts (computers) permanently connected to the Internet globally, and in New Zealand. The information is derived from the Internet Domain Survey (IDS) which attempts to discover every host on the Internet by doing a complete search of the Domain Name System (DNS).1 The latest survey was completed in January 2001.
Changes to the number of "hosts" (computers) on the Internet is regarded as an indicator of the rate of growth of the Internet. To be counted as a host, a computer must have its own Internet address and be permanently and directly connected to the Internet. Home or small business computers which connect by dialling up to a service provider are therefore not counted. Consequently, it is reasonable to assume that there are many more Internet users than hosts, possibly by a factor of five or more. To stress this point: the information presented here refers to the number of computers permanently linked to the Internet, it does not purport to describe directly the number of Internet users.
Data relating to Internet users and usage should be treated with caution due to the evolving nature of the Internet, the variety of ways it can be accessed and the difficulty in gathering reliable data. Over time, information collected has become less reliable because an increasing proportion of organisations now restrict access to their domain data. A new survey technique was introduced in July 1997 and "adjusted" host counts were provided for earlier years of the survey to enable comparisons to be made. For a full explanation of the changes to the survey visit the Internet Software Consortium Web site at http://www.isc.org/ds/index.html.
A more serious problem with collecting this information is due to an increasing trend for corporate networks to be screened off from the Internet behind firewalls. This means that a private network with hundreds or even thousands of computers may appear to the outside world as just one or two host computers. Unfortunately, the new methodology employed since 1997 does nothing to overcome the problem of networks being screened behind firewalls.
Figure 6.1.1 shows the growth in the number of hosts connected to the Internet worldwide, from July 1991 through to January 2001 including the revised total host counts for the earlier surveys from January 1995 to January 1997.
Use of a logarithmic scale allows the number of hosts in the .nz ccTLD2 to be graphed alongside the world-wide data despite the substantial difference in the actual numbers. Note that an exponential curve plots as a straight line when a logarithmic scale is used.
Figure 6.1.1: Growth in Internet Hosts - New Zealand (.nz) and the World
The graph shows a continuous exponential increase in the number of Internet hosts worldwide. The number of hosts in New Zealand has also increased rapidly although it appeared to plateau from 1996 to 1997 and actually fell slightly in 1998.
As other information detailed below shows, use of the Internet in New Zealand is continuing to increase rapidly and the apparent fall in host numbers in 1998 indicates that this now a less useful measure of Internet activity in New Zealand. The reasons for the anomaly is not clear but it may be due at least in part to the growing use of firewalls which screen the actual number of computers on corporate Intranets from external scrutiny, and thus could be interpreted as representing a maturation of the use of the Internet in New Zealand.
Figure 6.1.2 is based on similar data to that in Figure 6.1.1 and shows the number of hosts per 1000 population for the 11 leading countries or economies (including Hong Kong). The information is based on a methodology which adjusts the figures for each country based on combining the number of hosts in each ccTLD and the number for each country in each of the gTLDs. Figures for September 1997 to March 2000 are from OECD Local Access Pricing and Ecommerce (OECD 2000b) and are based on Netsizer data (see http://www.netsizer.com/) using a methodology devised by Telcordia. Figures for March 2001 are taken directly from Netsizer. Note that Hong Kong has now entered the "top ten" and it is to be expected that a number of other Asian economies, especially Korea and Singapore, are likely to overtake others listed within the next 12 months.
Figure 6.1.2: Number of Internet Hosts per 1000 Population (Netsizer/Telcordia)
Figure 6.1.3 is based on data for OECD members as of July 1999 (OECD 2000a) derived by a different methodology from those in Figure 6.1.2 and shows ten of the same eleven countries (Hong Kong omitted) but with Switzerland added. Although the figures are probably not as robust as with the other methodology, the chart is useful because it shows the respective contributions of hosts registered in both the ccTLDs and the gTLDs.
There is quite wide variability between countries with registrations in the gTLDs. One reason for this is that, while there are no restrictions on who may register in the commercially important .com domain, there is a considerable difference with registration policies amongst the ccTLDs, for example with some restricted to residents only and others being completely unregulated.
Figure 6.1.3: Top Ten OECD Countries by Hosts per 1000 Population as of July 1999 (Internet Software Consortium)
6.2 Domain Name Registrations (Updated)
The data in this section (and section 6.3) are derived from figures compiled by Mark Davies of Victoria University of Wellington (Davies).
When an organisation connects to the Internet, it typically registers a domain name such as companyname.co.nz. The number of domain names registered provides an indication of the numbers of organisations with an Internet presence. Fourth and higher level sub-domains are not counted, thus med.govt.nz is counted but comms.med.govt.nz is not.
Note that the figures can be taken only as indicative of Internet activity rather than providing firm information about the number of organisations linking to the Internet in New Zealand. There are few restrictions on who may register in the .nz domain, especially the most popular second level domains .co.nz (commercial organisations) .org.nz and .net.nz. Some organisations have multiple registrations and some registrations are from overseas. Conversely, many New Zealand organisations have registrations in other domains, especially the generic top level domains (gTLDs), such as .com.
Figure 6.2 shows the domain names registered in .nz as a time series since January 1994.
Figure 6.2: New Zealand Internet Domains
6.3 Types of Organisation on the Internet (Updated)
Organisations are normally allocated domain names according to organisation type, with the penultimate part of a name (or second level domain), such as the "co" in <name.co.nz> used to categorise domains. The following second level domains are used in New Zealand:
| Second Level Domain | Normally Used by: |
| ac.nz | Tertiary educational institutions |
| co.nz | Companies |
| cri.nz | Crown Research Institutes |
| gen.nz | Individuals and organisations which do not fit the other categories |
| govt.nz | Central government agencies and local and regional councils |
| iwi.nz | Iwi organisations |
| mil.nz | Military organisations |
| net.nz | Internet Service Providers |
| org.nz | Non-profit organisations and incorporated societies |
| school.nz | Schools |
Note that the table represents customary usage and registering organisations free to choose between several secondary level domains. For example, ISPs have been registered in .gen.nz and .co.nz as well as .net.nz.
Figure 6.3.1 shows the split of commercial (.co.nz) to non-commercial organisations (all other second level domain names registered). Commercial registrations have greatly outnumbered non-commercial since early 1996.
Figure 6.3.1 Commercial vs Non-Commercial Domains
Figure 6.3.2 shows the growth in the non-commercial domains. The "all others" category includes .govt.nz, .ac.nz, .cri.nz which were most significant in the early years of Internet development. Also included are .iwi.nz and .mil.nz, both with very low numbers of registrations. The rapid increase in registrations in .org.nz shows increasing interest in the Internet by non-profit and community organisations, while the rise in .net.nz reflects a growing number of organisations indicating their predominant interest in the Internet itself. In recent years there has also been increasing use of the Internet by schools as shown by the growth in .school.nz (to 1244 in early March 2001).
Figure 6.3.2: Non-Commercial Registrations on the Internet in New Zealand
6.4 Organisations with World-Wide Web Sites (Updated)
Information in this section is derived from figures compiled by Mark Davies of Victoria University of Wellington (Davies) and the OECD (OECD 2000b).
The number of organisations with a World-wide Web site gives a better measure of the actual numbers of organisations actively involved in the Internet than does the number of domain names registered. Organisations will most commonly have just one Web site even if they have registered several domain names. The number of Web sites is estimated by counting all the domain names with the format: www.name.domain.nz. However, this only gives an estimate of the number of Web sites, as it over counts organisations using more than one address, excludes Web sites which do not start with "www" and does not take into account non-active Web sites.
Figure 6.4.1 shows the total number of sites in the .nz domain split into commercial (co.nz) Web sites and non-commercial Web sites estimated in this way from February 1997 to February 2000. There were 60,966 Web sites in .nz as of March 2001 of which 52,693 or 86% have been set up by commercial organisations.
Figure 6.4.1 New Zealand (.nz) WWW Sites
As noted in section 6.2, many organisations register names in the generic top level domains (such as .com), rather than country code top level domains such as .nz. Figure 6.4.2 shows the numbers of Web servers per 1000 population (by probable location) as of February 2000 in the top 14 OECD countries, along with the OECD average (OECD 2000b).
Figure 6.4.2: Web Servers per 1000 Population (OECD Countries, February 2000)
The disproportionate number of Web sites located in the US reflects a real digital divide between countries even within the OECD. While there are many factors at work here, it is probably due in some measure to businesses outside the US locating their Web content in the US where they find the best prices, services and (especially) network performance, including the availability and pricing of international bandwidth.
Electronic commerce activity is probably best measured by the number of sites which provide secure services on the Web. Figure 6.4.3 shows the "top ten" numbers of secure Web servers per 1 million population on March 2000 as reported by the OECD (OECD 2000b) based on information collected by Netcraft (http://www.netcraft.co.uk/). The data are derived from numbers of secure Web sites with third party certification suitable for encrypted credit card transactions over the Internet. Note that, unlike statistics which rely on domain names for an indication of country, the information in Figure 6.4.3 is based on the actual geographic address of the business.
New Zealand has a comparatively high level of use of secure servers by international standards.
Figure 6.4.3: Secure Web Servers per 1,000,000 Population (OECD Countries, March 2000)
Back to Top