Computer network size growth forecasting method and system

Details Computer network size growth forecasting method and system Computer network size growth forecasting method and system

1998-11-03

2001-01-23

Burgess, Glenton B. (Department: 2757)

Electrical computers and digital processing systems: multicomput

Computer network managing

Computer network monitoring

C702S128000, C702S179000

Reexamination Certificate

active

06178451

ABSTRACT:

BACKGROUND OF THE INVENTION
The present invention relates generally to estimating the size of computer networks and, more particularly, to estimating and forecasting growth in the number of Internet hosts registered in the Internet Domain Name Service (DNS).
Every computer with a permanent connection to the Internet is identified by an Internet Protocol (IP) address. As shown in
FIG. 1
, examples of computers with IP addresses are host computer
10
, server
20
, workstation
30
(connected to Internet Backbone
50
through LAN
35
), and Internet Service Provider (ISP)
40
. An IP address comprises four parts, e.g., “a.b.c.d,” which corresponds to the address of host
10
in FIG.
1
. Each part is an integer between 0 and 255, so that there are 2
8
=256 possible values for each of the four parts, and each part can be represented using 8 bits. Therefore, there are 2
32
possible IP addresses. Each part of the address becomes increasingly machine-specific. For example, the first part specifies a geographic region, the second part specifies a service provider or organization such as a university, the third part specifies a group of computers, and the fourth part specifies the machine itself.
In addition to host computers, other pieces of computer equipment have their own IP addresses, e.g., server
20
has address efg.h, and workstation
30
has address i.j.k.l. Individual computers such as PCs, however, do not have permanent IP addresses. They only receive an IP address temporarily by connecting to an ISP such as ISP
40
, which has an IP address for each modem in its modem bank.
Every computer on the Internet also has an alphanumeric name, referred to as a domain name, and the Domain Name System (DNS) contains, among other information, mappings between IP addresses and domain names. The DNS is a distributed database held by systems running name server software. There is a hierarchy of DNS servers, with servers at the lowest level containing name-to-address mappings for a group of hosts, and servers at increasingly higher levels containing data for larger groups of hosts. At the top level are root name servers that hold all data for the top-level domains, e.g., “.com,” “.org,” and geographical domains such as “.uk” and “.jp.” Whenever a user on a local computer enters a domain name, the local computer contacts a DNS server, possibly on the local computer itself. If the first DNS server cannot resolve the domain name by finding its IP address, the DNS server contacts an authoritative server higher up the hierarchy. That server, in turn, contacts an even higher-level DNS server if it cannot resolve the domain name.
For reverse mapping from IP addresses to domain names, there is a pseudo-domain called IN-ADDR.ARPA. This domain contains exactly one PTR (pointer) record for each IP address. Because the highest order of significance in the naming system is on the right, the notation for addresses is reversed in IN-ADDR.ARPA. For example, the IP address 120.76.108.14 would have a reverse domain entry of 14.108.76.120.IN-ADDR.ARPA.
The size of the Internet can be determined based only on one directly measurable quantity—the number of computers registered in DNS. One method for determining the number of computers registered is to perform an exhaustive count using DNS zone transfers. In a DNS zone transfer, a DNS server requests a download of information from another server in the same zone, or portion of the domain space. Usually a server requests a download from a server that is higher in the hierarchy. By requesting DNS zone transfers throughout the Internet, one server can actually count the number of computers registered. One problem with this approach, however, is that as the Internet grows, exhaustive surveys take longer and longer to perform, so that results for one month may not be available until at least the following month. A second problem with this approach is that zone transfers put a heavy load on servers and are considered to be intrusive. As a result, many servers have banned zone transfers, causing estimates of Internet size based on DNS zone transfers to become less and less accurate.
Furthermore, current methodologies for estimating the size of the Internet provide only historical measurements. They do not provide a forecast of future Internet growth.
It is desirable, therefore, to provide a method for accurately estimating the number of registered hosts on the Internet. It is also desirable to provide a method for forecasting the growth in the number of registered hosts on the Internet. It is also desirable to provide a method for estimating and forecasting the size of the Internet by arbitrary segment, e.g., based on the top level or second level domain name. It is even more desirable to provide a method that quickly and accurately estimates current Internet size and does not underestimate the number of hosts.
DESCRIPTION OF THE INVENTION
The present invention satisfies those desires and other desires by providing a methodology for estimating and forecasting the number of Internet hosts based on a statistical analysis of a random sampling of Internet addresses.
A method consistent with an embodiment of the present invention for estimating growth in the number of computers in a network comprises the steps of modeling network growth based on a statistical model, determining a probability from the statistical model, and estimating network growth based on the determined probability. In particular, the statistical model may be a birth rate model, and the probability may be a birth probability. A method consistent with an embodiment of the present invention, wherein the statistical model has unknown parameters, further comprises the steps of obtaining a sample of the number of computers in the network and estimating the value of the unknown parameters based on the sampled number of computers.
A method consistent with an embodiment of the present invention for sampling the number of computers in the network comprises the steps of randomly generating a plurality of potential computer addresses, determining whether each of the addresses has been allocated to a computer, and statistically analyzing the number of allocated and non-allocated addresses to estimate the number of computers.


REFERENCES:
patent: 5778069 (1998-07-01), Thomlinson et al.
patent: 5778172 (1998-07-01), Riedle et al.
patent: 5838688 (1998-11-01), Kadambi et al.
Internet Software Consortium, “Internet Domain Survey Background”, http://www.isc.org/ds
ew-survey.html, 2000.
Internet Software Consortium, Internet Domain Survey, Jan. 1998, http://www.isc.org/ds/WWW-9801/report.html, 2000.
Stephen et al., “How to Judge the Quality of a Survey”, American Demographics, p. 42, Apr. 1990.
Stillwater, “Study Puts US Online Population at 62 Million”, Newsbytes News Network, Feb. 1998.
Kendall et al., Dictionary of Statistical Terms, Hafner Publishing Company, 1960, p. 29, 1960.
Eleftheriadis et al., “Address Management and Connection Control for Multicast Communication Applications”, IEEE 1995, pp. 386-393.
Lottor, M., “Internet Growth (1981-1991)”, RFC 1296, Jan. 1992, pp. 1-9.
Wiggins R.W., “Growth of the Internet”, http://www.themesh.com/gedit12.html, Sep. 1995.

Affiliated with

Also associated with

Rating

Computer network size growth forecasting method and system does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Computer network size growth forecasting method and system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Computer network size growth forecasting method and system will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2539679