WK1 Discussion and Replies

8

The Incredible Growth

8.1 Introduction

During the years between 1980 and the 2000s, the Internet changed from a small, experimental research project into the world’s largest computer network. In 1981, the Internet connected approximately one hundred computers at research sites and universi- ties. By 2000, over seventy-two million computers were attached, and the Internet con- tinues to grow. The introduction of the smart phone changed the Internet considerably; in 2016, more users accessed the Internet through a smart phone than through a laptop or desktop computer.

This chapter chronicles the phenomenal growth of the Internet and the changes that accompanied it. It discusses steps that were taken to stimulate growth, and concludes by explaining some of the consequences and opportunities that arose from rapid adop- tion.

8.2 Stimulating Adoption

In 1980, the Internet was merely a research project. A handful of universities and research labs had copies of the TCP/IP software. By 1985, it was becoming a produc- tion network system. Experimental TCP/IP software was available for several brands of computers, and it was used every day. The Internet reached researchers at few dozen academic and industrial research labs.

Before the U.S. military could use TCP/IP for production work, however, the tech- nology needed to become more robust. The software needed to be polished and tested,

81

8

The Incredible Growth

8.1 Introduction

During the years between 1980 and the 2000s, the Internet changed from a small, experimental research project into the world’s largest computer network. In 1981, the Internet connected approximately one hundred computers at research sites and universi- ties. By 2000, over seventy-two million computers were attached, and the Internet con- tinues to grow. The introduction of the smart phone changed the Internet considerably; in 2016, more users accessed the Internet through a smart phone than through a laptop or desktop computer.

This chapter chronicles the phenomenal growth of the Internet and the changes that accompanied it. It discusses steps that were taken to stimulate growth, and concludes by explaining some of the consequences and opportunities that arose from rapid adop- tion.

8.2 Stimulating Adoption

In 1980, the Internet was merely a research project. A handful of universities and research labs had copies of the TCP/IP software. By 1985, it was becoming a produc- tion network system. Experimental TCP/IP software was available for several brands of computers, and it was used every day. The Internet reached researchers at few dozen academic and industrial research labs.

Before the U.S. military could use TCP/IP for production work, however, the tech- nology needed to become more robust. The software needed to be polished and tested,

81

K392894.indd 109 03/08/18 12:07 pmComer, Douglas E.. The Internet Book : Everything You Need to Know about Computer Networking and How the Internet Works, CRC Press LLC, 2018. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=5502790. Created from apus on 2025-01-08 05:09:18.

C op

yr ig

ht ©

2 01

8. C

R C

P re

ss L

LC . A

ll rig

ht s

re se

rv ed

.

82 The Incredible Growth Chap. 8

and the whole system needed more tuning. DARPA considered the next step in its research program carefully.

8.3 Meanwhile, Back In Computer Science

While DARPA worked on the Internet research project, another technology came from a research lab and swept the computer science community: an operating system. Although vendors use the term operating system to refer to all the software that comes with a computer or smart phone, computer scientists use the term to describe the main piece of software that manages the computer, runs apps, controls input and output de- vices, and provides file storage. Operating systems are so complex that scientists and engineers spent years in the 1960s trying to understand them. By 1970, computer ven- dors had adopted the closed paradigm for operating system software, and vendors had created a proprietary operating system for each of their computers.

In the early 1970s, a small team of computer scientists at Bell Laboratories built a new operating system called the Unix Time-sharing System. Because Bell Laboratories used a variety of computers, the researchers wanted an operating system that could run on any hardware. So, they designed the system to be general — they created the software carefully, and made it easy to move a copy onto new computers.

Bell Labs decided to allow universities to obtain copies of the Unix system for use in teaching and research. Because they were interested in measuring its portability, Bell Labs gave away copies of the code, and encouraged universities to try running the sys- tem on new computers. As a result, the Unix system became one of the first operating systems that students could study.

A group of faculty and graduate students from the University of California at Berkeley became interested in the Unix system. They wrote application programs and modified the system itself. They added new features and experimented with applica- tions that communicated over a Local Area Network. To make the work available to other universities, researchers at Berkeley established a software distribution facility. When a university wanted a copy of the software, the distribution facility mailed a mag- netic tape that contained the software. The Berkeley version of the Unix system, known as BSD Unix,† became popular at other universities.

8.4 The Internet Meets Unix

DARPA realized that the Berkeley work on operating systems reached many universities, and decided to use it to disseminate Internet software. They negotiated a research contract with Berkeley. Under the terms of the contract, DARPA gave researchers at Berkeley a copy of the TCP/IP software that had been developed as part of the Internet project. Berkeley incorporated the software into their version of the Unix system, and modified application programs to use TCP/IP.

†The acronym BSD stands for Berkeley Software Distribution.

Sec. 8.4 The Internet Meets Unix 83

When Berkeley issued its next major software distribution, most computer science departments received TCP/IP software at virtually no cost. Although only a few com- puter science departments had computers connected to the Internet, most of them had a Local Area Network or were about to install one. They knew that their students needed to study networking. They also knew that using a network would make computing easier because it would allow users to share resources like printers.

For many departments, TCP/IP was the first viable networking software they had encountered. It offered a low-cost, efficient way to provide a departmental network and a technology that could be studied in classes. Thus, in a short time, most computer sci- ence departments had TCP/IP software running on their Local Area Networks, even though most had not yet connected to the Internet. The point is:

Computer science departments in universities received TCP/IP software along with a release of Unix system software from U.C. Berkeley. Although only a few departments had computers connected to the Internet, most of them used TCP/IP on their Local Area Net- works for teaching, research, and production computing.

8.5 The U.S. Military Makes A Commitment

By the early 1980s, the Internet operated reliably. It interconnected academic and research sites. More important, the Internet demonstrated that the basic principles of in- ternetworking were sound. Convinced of the Internet’s viability, the U.S. military start- ed to connect computers to the Internet and to use TCP/IP software.

In 1982, the U.S. military chose the Internet as its primary computer communica- tion system. Consequently, a cutoff date was planned. At the beginning of 1983, the ARPANET and associated military networks stopped running old communication software and switched to TCP/IP. On the cutover date, any computer that did not understand TCP/IP could not communicate. The point is:

Although the U.S. military funded Internet research and eventually chose to use the TCP/IP software, internetworking was developed and tested at civilian sites. Only after Internet technology had been demonstrated did the military switch its computers to the new technol- ogy.

8.6 The Internet Doubled In Size In One Year

Before the U.S. military started using TCP/IP on all its computers, the Internet in- terconnected approximately two hundred computers. One year later, it had doubled in size. In retrospect, the increase seems trivial. It involved hundreds, not thousands or millions of computers. At the time, however, the increase was significant.

K392894.indd 110 03/08/18 12:07 pmComer, Douglas E.. The Internet Book : Everything You Need to Know about Computer Networking and How the Internet Works, CRC Press LLC, 2018. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=5502790. Created from apus on 2025-01-08 05:09:18.

C op

yr ig

ht ©

2 01

8. C

R C

P re

ss L

LC . A

ll rig

ht s

re se

rv ed

.

82 The Incredible Growth Chap. 8

and the whole system needed more tuning. DARPA considered the next step in its research program carefully.

8.3 Meanwhile, Back In Computer Science

While DARPA worked on the Internet research project, another technology came from a research lab and swept the computer science community: an operating system. Although vendors use the term operating system to refer to all the software that comes with a computer or smart phone, computer scientists use the term to describe the main piece of software that manages the computer, runs apps, controls input and output de- vices, and provides file storage. Operating systems are so complex that scientists and engineers spent years in the 1960s trying to understand them. By 1970, computer ven- dors had adopted the closed paradigm for operating system software, and vendors had created a proprietary operating system for each of their computers.

In the early 1970s, a small team of computer scientists at Bell Laboratories built a new operating system called the Unix Time-sharing System. Because Bell Laboratories used a variety of computers, the researchers wanted an operating system that could run on any hardware. So, they designed the system to be general — they created the software carefully, and made it easy to move a copy onto new computers.

Bell Labs decided to allow universities to obtain copies of the Unix system for use in teaching and research. Because they were interested in measuring its portability, Bell Labs gave away copies of the code, and encouraged universities to try running the sys- tem on new computers. As a result, the Unix system became one of the first operating systems that students could study.

A group of faculty and graduate students from the University of California at Berkeley became interested in the Unix system. They wrote application programs and modified the system itself. They added new features and experimented with applica- tions that communicated over a Local Area Network. To make the work available to other universities, researchers at Berkeley established a software distribution facility. When a university wanted a copy of the software, the distribution facility mailed a mag- netic tape that contained the software. The Berkeley version of the Unix system, known as BSD Unix,† became popular at other universities.

8.4 The Internet Meets Unix

DARPA realized that the Berkeley work on operating systems reached many universities, and decided to use it to disseminate Internet software. They negotiated a research contract with Berkeley. Under the terms of the contract, DARPA gave researchers at Berkeley a copy of the TCP/IP software that had been developed as part of the Internet project. Berkeley incorporated the software into their version of the Unix system, and modified application programs to use TCP/IP.

†The acronym BSD stands for Berkeley Software Distribution.

Sec. 8.4 The Internet Meets Unix 83

When Berkeley issued its next major software distribution, most computer science departments received TCP/IP software at virtually no cost. Although only a few com- puter science departments had computers connected to the Internet, most of them had a Local Area Network or were about to install one. They knew that their students needed to study networking. They also knew that using a network would make computing easier because it would allow users to share resources like printers.

For many departments, TCP/IP was the first viable networking software they had encountered. It offered a low-cost, efficient way to provide a departmental network and a technology that could be studied in classes. Thus, in a short time, most computer sci- ence departments had TCP/IP software running on their Local Area Networks, even though most had not yet connected to the Internet. The point is:

Computer science departments in universities received TCP/IP software along with a release of Unix system software from U.C. Berkeley. Although only a few departments had computers connected to the Internet, most of them used TCP/IP on their Local Area Net- works for teaching, research, and production computing.

8.5 The U.S. Military Makes A Commitment

By the early 1980s, the Internet operated reliably. It interconnected academic and research sites. More important, the Internet demonstrated that the basic principles of in- ternetworking were sound. Convinced of the Internet’s viability, the U.S. military start- ed to connect computers to the Internet and to use TCP/IP software.

In 1982, the U.S. military chose the Internet as its primary computer communica- tion system. Consequently, a cutoff date was planned. At the beginning of 1983, the ARPANET and associated military networks stopped running old communication software and switched to TCP/IP. On the cutover date, any computer that did not understand TCP/IP could not communicate. The point is:

Although the U.S. military funded Internet research and eventually chose to use the TCP/IP software, internetworking was developed and tested at civilian sites. Only after Internet technology had been demonstrated did the military switch its computers to the new technol- ogy.

8.6 The Internet Doubled In Size In One Year

Before the U.S. military started using TCP/IP on all its computers, the Internet in- terconnected approximately two hundred computers. One year later, it had doubled in size. In retrospect, the increase seems trivial. It involved hundreds, not thousands or millions of computers. At the time, however, the increase was significant.

82 The Incredible Growth Chap. 8

and the whole system needed more tuning. DARPA considered the next step in its research program carefully.

8.3 Meanwhile, Back In Computer Science

While DARPA worked on the Internet research project, another technology came from a research lab and swept the computer science community: an operating system. Although vendors use the term operating system to refer to all the software that comes with a computer or smart phone, computer scientists use the term to describe the main piece of software that manages the computer, runs apps, controls input and output de- vices, and provides file storage. Operating systems are so complex that scientists and engineers spent years in the 1960s trying to understand them. By 1970, computer ven- dors had adopted the closed paradigm for operating system software, and vendors had created a proprietary operating system for each of their computers.

In the early 1970s, a small team of computer scientists at Bell Laboratories built a new operating system called the Unix Time-sharing System. Because Bell Laboratories used a variety of computers, the researchers wanted an operating system that could run on any hardware. So, they designed the system to be general — they created the software carefully, and made it easy to move a copy onto new computers.

Bell Labs decided to allow universities to obtain copies of the Unix system for use in teaching and research. Because they were interested in measuring its portability, Bell Labs gave away copies of the code, and encouraged universities to try running the sys- tem on new computers. As a result, the Unix system became one of the first operating systems that students could study.

A group of faculty and graduate students from the University of California at Berkeley became interested in the Unix system. They wrote application programs and modified the system itself. They added new features and experimented with applica- tions that communicated over a Local Area Network. To make the work available to other universities, researchers at Berkeley established a software distribution facility. When a university wanted a copy of the software, the distribution facility mailed a mag- netic tape that contained the software. The Berkeley version of the Unix system, known as BSD Unix,† became popular at other universities.

8.4 The Internet Meets Unix

DARPA realized that the Berkeley work on operating systems reached many universities, and decided to use it to disseminate Internet software. They negotiated a research contract with Berkeley. Under the terms of the contract, DARPA gave researchers at Berkeley a copy of the TCP/IP software that had been developed as part of the Internet project. Berkeley incorporated the software into their version of the Unix system, and modified application programs to use TCP/IP.

†The acronym BSD stands for Berkeley Software Distribution.

Sec. 8.4 The Internet Meets Unix 83

When Berkeley issued its next major software distribution, most computer science departments received TCP/IP software at virtually no cost. Although only a few com- puter science departments had computers connected to the Internet, most of them had a Local Area Network or were about to install one. They knew that their students needed to study networking. They also knew that using a network would make computing easier because it would allow users to share resources like printers.

For many departments, TCP/IP was the first viable networking software they had encountered. It offered a low-cost, efficient way to provide a departmental network and a technology that could be studied in classes. Thus, in a short time, most computer sci- ence departments had TCP/IP software running on their Local Area Networks, even though most had not yet connected to the Internet. The point is:

Computer science departments in universities received TCP/IP software along with a release of Unix system software from U.C. Berkeley. Although only a few departments had computers connected to the Internet, most of them used TCP/IP on their Local Area Net- works for teaching, research, and production computing.

8.5 The U.S. Military Makes A Commitment

By the early 1980s, the Internet operated reliably. It interconnected academic and research sites. More important, the Internet demonstrated that the basic principles of in- ternetworking were sound. Convinced of the Internet’s viability, the U.S. military start- ed to connect computers to the Internet and to use TCP/IP software.

In 1982, the U.S. military chose the Internet as its primary computer communica- tion system. Consequently, a cutoff date was planned. At the beginning of 1983, the ARPANET and associated military networks stopped running old communication software and switched to TCP/IP. On the cutover date, any computer that did not understand TCP/IP could not communicate. The point is:

Although the U.S. military funded Internet research and eventually chose to use the TCP/IP software, internetworking was developed and tested at civilian sites. Only after Internet technology had been demonstrated did the military switch its computers to the new technol- ogy.

8.6 The Internet Doubled In Size In One Year

Before the U.S. military started using TCP/IP on all its computers, the Internet in- terconnected approximately two hundred computers. One year later, it had doubled in size. In retrospect, the increase seems trivial. It involved hundreds, not thousands or millions of computers. At the time, however, the increase was significant.

K392894.indd 111 03/08/18 12:07 pmComer, Douglas E.. The Internet Book : Everything You Need to Know about Computer Networking and How the Internet Works, CRC Press LLC, 2018. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=5502790. Created from apus on 2025-01-08 05:09:18.

C op

yr ig

ht ©

2 01

8. C

R C

P re

ss L

LC . A

ll rig

ht s

re se

rv ed

.

84 The Incredible Growth Chap. 8

Anyone who has written a computer program knows that the program has specific size limits built into it. For example, some parts of the TCP/IP software use lists of computers and the addresses used to access them. When the Internet only contained dozens of computers, programmers chose maximum sizes that seemed huge (e.g., 300). As new computers joined the Internet, the list of computers exceeded the limits, and the software had to be revised to accommodate longer lists. At first, researchers made small increments to the software. They increased the capacity by ten or twenty percent. Soon, they found that it was insufficient, and further increases were needed. As the In- ternet continued to grow, the process of changing the software kept pace.

In addition to uncovering limitations in the software, the Internet growth revealed limits in manual and clerical procedures. For example, each time a new computer was added to the Internet, several people had to take action. Someone had to review the rea- sons for the connection and its relationship to the project before approving the connec- tion. Someone else had to assign a name to the computer, and then enter it in a data- base. Finally, someone had to make a physical connection between the computer and the network.

During the period of rapid growth, researchers were busy updating the software and had little spare time to help with manual procedures like registration; the duties be- gan to pass to a professional staff. We can summarize what happened:

As new computers were added to the Internet, it doubled in size in a single year. The rapid growth forced researchers to tune administra- tive procedures as well as the communication software.

8.7 Internet For Every Computer Science Department

In the late 1970s, many computer scientists recognized the importance of network- ing. A small group of researchers proposed a networking project to the National Sci- ence Foundation (NSF)† with a goal of devising a computer network to connect all computer science researchers in the U.S.

After reviewing the proposal, the National Science Foundation funded a project to build the Computer Science Network. The project, which also had support from DAR- PA, became known by the acronym CSNET. To reach all computer scientists in the country, CSNET had to contend with the problem of providing network service to small universities in rural areas as well as major universities in metropolitan areas. DARPA encouraged CSNET to run TCP/IP software and connect researchers to the Internet. For smaller institutions that could not afford direct connections, CSNET devised ways to provide limited network services at much lower cost.

By the time the U.S. military selected the Internet as a primary computer commun- ication system, many of the top computer science groups in industry and academia were already using it. Over the next few years, CSNET worked to provide Internet connec- tions to the remaining computer science departments. As a result, by the mid-1980s, most computer scientists had Internet access.

†NSF, a U.S. federal agency, is responsible for funding research and education in science and engineer- ing.

Sec. 8.8 Graduate Student Volunteers Contribute 85

8.8 Graduate Student Volunteers Contribute

Connecting computer science researchers to the Internet had an interesting effect. Although some computer scientists work in industrial research labs, many are professors who work in universities, where they also teach classes and advise students. The pro- fessors talked to students about the Internet project, the technology and software that it used, its success, and the remaining research problems. The professors’ enthusiasm was contagious.

Students became interested in learning more about TCP/IP and the Internet. Gra- duate students who were searching for research topics began to investigate the technical details of TCP/IP software. They studied ways to extend the Internet technology, and devised experiments to measure its capabilities. They considered new applications, and found ways to extend the functionality. The result was synergistic: students gained valuable knowledge and experience with computer networks, while their creative ener- gies helped advance Internet technologies.

8.9 Internet Governance: The IAB And IETF

Scientists and engineers working on the Internet held regular meetings to discuss new ideas, review the technology, share discoveries, and exchange technical informa- tion. DARPA originally named the group the Internet Advisory Board. With the Inter- net growing rapidly, DARPA decided that the group of scientists should have a more formal structure and more responsibility for coordinating TCP/IP research and Internet development. It renamed the group the Internet Activities Board. Following military tradition, the board became known by its acronym, IAB.

DARPA appointed a chairman of the IAB, who was given the informal title Inter- net Architect, even though the Internet was already growing too rapidly for a single per- son to be responsible for an architectural plan. Another member of the IAB was desig- nated to be the RFC Editor, and given responsibility for reviewing and editing all RFCs before they were published. Other scientists on the IAB were each assigned a specific problem to investigate.

To study an assigned problem, each member of the IAB gathered volunteers from the research community to serve on a task force. Each task force held meetings to dis- cuss ideas, resolve issues, generate new approaches, and report on experiments. If a task force reached a consensus on a new approach, members would build prototype software to demonstrate how their ideas worked in practice, and then would generate and submit a specification as an RFC.

The IAB guided the development of the Internet for several years. In 1989, it was reorganized to add more representatives from commercial companies. The IAB’s duties and interactions with other groups were reorganized again in 1992, when it became part of the Internet Society. At the time of its second reorganization, the IAB kept the acro- nym, but changed its name to the Internet Architecture Board. The IAB divested most

K392894.indd 112 03/08/18 12:07 pmComer, Douglas E.. The Internet Book : Everything You Need to Know about Computer Networking and How the Internet Works, CRC Press LLC, 2018. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=5502790. Created from apus on 2025-01-08 05:09:18.

C op

yr ig

ht ©

2 01

8. C

R C

P re

ss L

LC . A

ll rig

ht s

re se

rv ed

.

84 The Incredible Growth Chap. 8

Anyone who has written a computer program knows that the program has specific size limits built into it. For example, some parts of the TCP/IP software use lists of computers and the addresses used to access them. When the Internet only contained dozens of computers, programmers chose maximum sizes that seemed huge (e.g., 300). As new computers joined the Internet, the list of computers exceeded the limits, and the software had to be revised to accommodate longer lists. At first, researchers made small increments to the software. They increased the capacity by ten or twenty percent. Soon, they found that it was insufficient, and further increases were needed. As the In- ternet continued to grow, the process of changing the software kept pace.

In addition to uncovering limitations in the software, the Internet growth revealed limits in manual and clerical procedures. For example, each time a new computer was added to the Internet, several people had to take action. Someone had to review the rea- sons for the connection and its relationship to the project before approving the connec- tion. Someone else had to assign a name to the computer, and then enter it in a data- base. Finally, someone had to make a physical connection between the computer and the network.

During the period of rapid growth, researchers were busy updating the software and had little spare time to help with manual procedures like registration; the duties be- gan to pass to a professional staff. We can summarize what happened:

As new computers were added to the Internet, it doubled in size in a single year. The rapid growth forced researchers to tune administra- tive procedures as well as the communication software.

8.7 Internet For Every Computer Science Department

In the late 1970s, many computer scientists recognized the importance of network- ing. A small group of researchers proposed a networking project to the National Sci- ence Foundation (NSF)† with a goal of devising a computer network to connect all computer science researchers in the U.S.

After reviewing the proposal, the National Science Foundation funded a project to build the Computer Science Network. The project, which also had support from DAR- PA, became known by the acronym CSNET. To reach all computer scientists in the country, CSNET had to contend with the problem of providing network service to small universities in rural areas as well as major universities in metropolitan areas. DARPA encouraged CSNET to run TCP/IP software and connect researchers to the Internet. For smaller institutions that could not afford direct connections, CSNET devised ways to provide limited network services at much lower cost.

By the time the U.S. military selected the Internet as a primary computer commun- ication system, many of the top computer science groups in industry and academia were already using it. Over the next few years, CSNET worked to provide Internet connec- tions to the remaining computer science departments. As a result, by the mid-1980s, most computer scientists had Internet access.

†NSF, a U.S. federal agency, is responsible for funding research and education in science and engineer- ing.

Sec. 8.8 Graduate Student Volunteers Contribute 85

8.8 Graduate Student Volunteers Contribute

Connecting computer science researchers to the Internet had an interesting effect. Although some computer scientists work in industrial research labs, many are professors who work in universities, where they also teach classes and advise students. The pro- fessors talked to students about the Internet project, the technology and software that it used, its success, and the remaining research problems. The professors’ enthusiasm was contagious.

Students became interested in learning more about TCP/IP and the Internet. Gra- duate students who were searching for research topics began to investigate the technical details of TCP/IP software. They studied ways to extend the Internet technology, and devised experiments to measure its capabilities. They considered new applications, and found ways to extend the functionality. The result was synergistic: students gained valuable knowledge and experience with computer networks, while their creative ener- gies helped advance Internet technologies.

8.9 Internet Governance: The IAB And IETF

Scientists and engineers working on the Internet held regular meetings to discuss new ideas, review the technology, share discoveries, and exchange technical informa- tion. DARPA originally named the group the Internet Advisory Board. With the Inter- net growing rapidly, DARPA decided that the group of scientists should have a more formal structure and more responsibility for coordinating TCP/IP research and Internet development. It renamed the group the Internet Activities Board. Following military tradition, the board became known by its acronym, IAB.

DARPA appointed a chairman of the IAB, who was given the informal title Inter- net Architect, even though the Internet was already growing too rapidly for a single per- son to be responsible for an architectural plan. Another member of the IAB was desig- nated to be the RFC Editor, and given responsibility for reviewing and editing all RFCs before they were published. Other scientists on the IAB were each assigned a specific problem to investigate.

To study an assigned problem, each member of the IAB gathered volunteers from the research community to serve on a task force. Each task force held meetings to dis- cuss ideas, resolve issues, generate new approaches, and report on experiments. If a task force reached a consensus on a new approach, members would build prototype software to demonstrate how their ideas worked in practice, and then would generate and submit a specification as an RFC.

The IAB guided the development of the Internet for several years. In 1989, it was reorganized to add more representatives from commercial companies. The IAB’s duties and interactions with other groups were reorganized again in 1992, when it became part of the Internet Society. At the time of its second reorganization, the IAB kept the acro- nym, but changed its name to the Internet Architecture Board. The IAB divested most

84 The Incredible Growth Chap. 8

Anyone who has written a computer program knows that the program has specific size limits built into it. For example, some parts of the TCP/IP software use lists of computers and the addresses used to access them. When the Internet only contained dozens of computers, programmers chose maximum sizes that seemed huge (e.g., 300). As new computers joined the Internet, the list of computers exceeded the limits, and the software had to be revised to accommodate longer lists. At first, researchers made small increments to the software. They increased the capacity by ten or twenty percent. Soon, they found that it was insufficient, and further increases were needed. As the In- ternet continued to grow, the process of changing the software kept pace.

In addition to uncovering limitations in the software, the Internet growth revealed limits in manual and clerical procedures. For example, each time a new computer was added to the Internet, several people had to take action. Someone had to review the rea- sons for the connection and its relationship to the project before approving the connec- tion. Someone else had to assign a name to the computer, and then enter it in a data- base. Finally, someone had to make a physical connection between the computer and the network.

During the period of rapid growth, researchers were busy updating the software and had little spare time to help with manual procedures like registration; the duties be- gan to pass to a professional staff. We can summarize what happened:

As new computers were added to the Internet, it doubled in size in a single year. The rapid growth forced researchers to tune administra- tive procedures as well as the communication software.

8.7 Internet For Every Computer Science Department

In the late 1970s, many computer scientists recognized the importance of network- ing. A small group of researchers proposed a networking project to the National Sci- ence Foundation (NSF)† with a goal of devising a computer network to connect all computer science researchers in the U.S.

After reviewing the proposal, the National Science Foundation funded a project to build the Computer Science Network. The project, which also had support from DAR- PA, became known by the acronym CSNET. To reach all computer scientists in the country, CSNET had to contend with the problem of providing network service to small universities in rural areas as well as major universities in metropolitan areas. DARPA encouraged CSNET to run TCP/IP software and connect researchers to the Internet. For smaller institutions that could not afford direct connections, CSNET devised ways to provide limited network services at much lower cost.

By the time the U.S. military selected the Internet as a primary computer commun- ication system, many of the top computer science groups in industry and academia were already using it. Over the next few years, CSNET worked to provide Internet connec- tions to the remaining computer science departments. As a result, by the mid-1980s, most computer scientists had Internet access.

†NSF, a U.S. federal agency, is responsible for funding research and education in science and engineer- ing.

Sec. 8.8 Graduate Student Volunteers Contribute 85

8.8 Graduate Student Volunteers Contribute

Connecting computer science researchers to the Internet had an interesting effect. Although some computer scientists work in industrial research labs, many are professors who work in universities, where they also teach classes and advise students. The pro- fessors talked to students about the Internet project, the technology and software that it used, its success, and the remaining research problems. The professors’ enthusiasm was contagious.

Students became interested in learning more about TCP/IP and the Internet. Gra- duate students who were searching for research topics began to investigate the technical details of TCP/IP software. They studied ways to extend the Internet technology, and devised experiments to measure its capabilities. They considered new applications, and found ways to extend the functionality. The result was synergistic: students gained valuable knowledge and experience with computer networks, while their creative ener- gies helped advance Internet technologies.

8.9 Internet Governance: The IAB And IETF

Scientists and engineers working on the Internet held regular meetings to discuss new ideas, review the technology, share discoveries, and exchange technical informa- tion. DARPA originally named the group the Internet Advisory Board. With the Inter- net growing rapidly, DARPA decided that the group of scientists should have a more formal structure and more responsibility for coordinating TCP/IP research and Internet development. It renamed the group the Internet Activities Board. Following military tradition, the board became known by its acronym, IAB.

DARPA appointed a chairman of the IAB, who was given the informal title Inter- net Architect, even though the Internet was already growing too rapidly for a single per- son to be responsible for an architectural plan. Another member of the IAB was desig- nated to be the RFC Editor, and given responsibility for reviewing and editing all RFCs before they were published. Other scientists on the IAB were each assigned a specific problem to investigate.

To study an assigned problem, each member of the IAB gathered volunteers from the research community to serve on a task force. Each task force held meetings to dis- cuss ideas, resolve issues, generate new approaches, and report on experiments. If a task force reached a consensus on a new approach, members would build prototype software to demonstrate how their ideas worked in practice, and then would generate and submit a specification as an RFC.

The IAB guided the development of the Internet for several years. In 1989, it was reorganized to add more representatives from commercial companies. The IAB’s duties and interactions with other groups were reorganized again in 1992, when it became part of the Internet Society. At the time of its second reorganization, the IAB kept the acro- nym, but changed its name to the Internet Architecture Board. The IAB divested most

K392894.indd 113 03/08/18 12:07 pmComer, Douglas E.. The Internet Book : Everything You Need to Know about Computer Networking and How the Internet Works, CRC Press LLC, 2018. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=5502790. Created from apus on 2025-01-08 05:09:18.

C op

yr ig

ht ©

2 01

8. C

R C

P re

ss L

LC . A

ll rig

ht s

re se

rv ed

.

86 The Incredible Growth Chap. 8

of its technical responsibilities, passing more control to subordinate groups, and leaving the board as the ultimate arbiter of policies and standards.

Among the subordinate groups, one stood out: the Internet Engineering Task Force (IETF). The IETF has survived reorganizations, and has taken over responsibility for generating new Internet technologies, revising older technologies, and creating, revising, and publishing standards documents. Most RFCs now originate within the IETF from committees, which are called working groups.

The IETF is partitioned into areas of interest, with an area director assigned to coordinate groups within each area. The IETF holds open meetings approximately three times per year. When it holds a meeting, thousands of people attend, most from com- mercial companies. Attendees are volunteers who attend to hear about the latest developments and participate in efforts to refine and improve the software.

We can summarize:

The group responsible for guiding the research and development of the Internet is known as the Internet Architecture Board (IAB). The primary subgroup responsible for technical matters is known as the Internet Engineering Task Force (IETF).

8.10 NSF Led Internet Expansion

During the years following the military adoption of TCP/IP, rapid growth contin- ued. Government agencies, such as NASA (the National Aeronautics And Space Ad- ministration) connected to the Internet. By the mid-1980s, the National Science Foun- dation (NSF) realized that eminence in science would soon demand computer communi- cation. Before computer networks, scientists exchanged ideas by publishing them in scientific journals, which took many months, sometimes years. Computer communica- tion was about to change the way scientists did research. A computer network makes it possible to share data as an experiment proceeds, making it possible for many other scientists to analyze the results without traveling to the site of the experiment.

Recognizing how important the Internet was becoming to science, NSF decided to fund further Internet growth. In 1985, NSF announced that it intended to connect researchers at 100 U.S. universities to the Internet. NSF advised the U.S. Congress of the plan, and received additional funds to support networking. NSF consulted experts in the field, devised a plan, and began a program that resulted in major changes to the Internet.

Scientists often use sophisticated, high-speed computers called supercomputers to analyze data from their experiments. Because supercomputers were expensive, NSF had established five supercomputer centers around the country. A scientist working on an NSF project had to travel to the nearest supercomputer center to process their data.

As the first step of Internet expansion, NSF built a Wide Area Network that inter- connected its five supercomputer centers. The network used TCP/IP, and provided a

Sec. 8.10 NSF Led Internet Expansion 87

connection to the Internet. Named NSFNET, the network was initially much smaller, and not any faster, than the ARPANET. Scientists found the network useful, but not exciting.

8.11 NSF Target: All Of Science And Engineering

Spurred by the success of CSNET and the initial NSFNET, NSF launched a new program to keep the U.S. competitive. The program had an ambitious goal:

NSF decided that the U.S. would not remain competitive unless every science and engineering researcher had Internet access.

To achieve the goal, NSF decided to use its funds to create a major new Internet that had significantly more capacity than the existing Internet. After examining avail- able technologies and reviewing its budget, NSF decided that it could not afford to pay for the entire project. Instead, it decided to offer partial support, in the form of federal grants. Companies and other organizations submitted written proposals to NSF to re- quest funding to work on the project.

NSF divided the grants into two types. First, NSF funded a group that wanted to build and operate a new high-speed Wide Area Network to connect parts of the Internet. The new WAN had to replace parts of the ARPANET as well as the original NSFNET. Second, NSF funded groups wanted to interconnect computers in a small region of the country and attach them to the new WAN. For example, NSF thought that each state might choose to apply as a group. Originally, the groups were referred to as NSF Re- gional Networks. Later, when it became clear that some of the groups spanned large geographic areas, NSF began referring to them as NSF Mid-Level Networks, but most professionals still called them regionals.

Because most universities or companies already had LANs connecting their com- puters, NSF decided to use its funds to help pay for long-distance connections; indivi- dual companies and schools paid for their own internal networks.

8.12 The NSFNET Backbone

NSF used a competitive bidding process when it awarded a grant for the new Inter- net WAN, which became known as the NSFNET backbone.† In 1987, NSF asked for proposals and used a panel of scientists to help assess them. After considering the alter- natives, NSF selected a joint proposal from three organizations: IBM, a computer manufacturer; MCI, a long-distance telephone company; and MERIT, an organization that built and operated a network connecting schools in Michigan.

The three groups cooperated to establish a new Wide Area Network that became the backbone of the Internet in the summer of 1988. MCI provided long-distance

†The term backbone is used as an analogy to a human spine that forms a central structure to which many other bones attach.

K392894.indd 114 03/08/18 12:07 pmComer, Douglas E.. The Internet Book : Everything You Need to Know about Computer Networking and How the Internet Works, CRC Press LLC, 2018. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=5502790. Created from apus on 2025-01-08 05:09:18.

C op

yr ig

ht ©

2 01

8. C

R C

P re

ss L

LC . A

ll rig

ht s

re se

rv ed

.

86 The Incredible Growth Chap. 8

of its technical responsibilities, passing more control to subordinate groups, and leaving the board as the ultimate arbiter of policies and standards.

Among the subordinate groups, one stood out: the Internet Engineering Task Force (IETF). The IETF has survived reorganizations, and has taken over responsibility for generating new Internet technologies, revising older technologies, and creating, revising, and publishing standards documents. Most RFCs now originate within the IETF from committees, which are called working groups.

The IETF is partitioned into areas of interest, with an area director assigned to coordinate groups within each area. The IETF holds open meetings approximately three times per year. When it holds a meeting, thousands of people attend, most from com- mercial companies. Attendees are volunteers who attend to hear about the latest developments and participate in efforts to refine and improve the software.

We can summarize:

The group responsible for guiding the research and development of the Internet is known as the Internet Architecture Board (IAB). The primary subgroup responsible for technical matters is known as the Internet Engineering Task Force (IETF).

8.10 NSF Led Internet Expansion

During the years following the military adoption of TCP/IP, rapid growth contin- ued. Government agencies, such as NASA (the National Aeronautics And Space Ad- ministration) connected to the Internet. By the mid-1980s, the National Science Foun- dation (NSF) realized that eminence in science would soon demand computer communi- cation. Before computer networks, scientists exchanged ideas by publishing them in scientific journals, which took many months, sometimes years. Computer communica- tion was about to change the way scientists did research. A computer network makes it possible to share data as an experiment proceeds, making it possible for many other scientists to analyze the results without traveling to the site of the experiment.

Recognizing how important the Internet was becoming to science, NSF decided to fund further Internet growth. In 1985, NSF announced that it intended to connect researchers at 100 U.S. universities to the Internet. NSF advised the U.S. Congress of the plan, and received additional funds to support networking. NSF consulted experts in the field, devised a plan, and began a program that resulted in major changes to the Internet.

Scientists often use sophisticated, high-speed computers called supercomputers to analyze data from their experiments. Because supercomputers were expensive, NSF had established five supercomputer centers around the country. A scientist working on an NSF project had to travel to the nearest supercomputer center to process their data.

As the first step of Internet expansion, NSF built a Wide Area Network that inter- connected its five supercomputer centers. The network used TCP/IP, and provided a

Sec. 8.10 NSF Led Internet Expansion 87

connection to the Internet. Named NSFNET, the network was initially much smaller, and not any faster, than the ARPANET. Scientists found the network useful, but not exciting.

8.11 NSF Target: All Of Science And Engineering

Spurred by the success of CSNET and the initial NSFNET, NSF launched a new program to keep the U.S. competitive. The program had an ambitious goal:

NSF decided that the U.S. would not remain competitive unless every science and engineering researcher had Internet access.

To achieve the goal, NSF decided to use its funds to create a major new Internet that had significantly more capacity than the existing Internet. After examining avail- able technologies and reviewing its budget, NSF decided that it could not afford to pay for the entire project. Instead, it decided to offer partial support, in the form of federal grants. Companies and other organizations submitted written proposals to NSF to re- quest funding to work on the project.

NSF divided the grants into two types. First, NSF funded a group that wanted to build and operate a new high-speed Wide Area Network to connect parts of the Internet. The new WAN had to replace parts of the ARPANET as well as the original NSFNET. Second, NSF funded groups wanted to interconnect computers in a small region of the country and attach them to the new WAN. For example, NSF thought that each state might choose to apply as a group. Originally, the groups were referred to as NSF Re- gional Networks. Later, when it became clear that some of the groups spanned large geographic areas, NSF began referring to them as NSF Mid-Level Networks, but most professionals still called them regionals.

Because most universities or companies already had LANs connecting their com- puters, NSF decided to use its funds to help pay for long-distance connections; indivi- dual companies and schools paid for their own internal networks.

8.12 The NSFNET Backbone

NSF used a competitive bidding process when it awarded a grant for the new Inter- net WAN, which became known as the NSFNET backbone.† In 1987, NSF asked for proposals and used a panel of scientists to help assess them. After considering the alter- natives, NSF selected a joint proposal from three organizations: IBM, a computer manufacturer; MCI, a long-distance telephone company; and MERIT, an organization that built and operated a network connecting schools in Michigan.

The three groups cooperated to establish a new Wide Area Network that became the backbone of the Internet in the summer of 1988. MCI provided long-distance

†The term backbone is used as an analogy to a human spine that forms a central structure to which many other bones attach.

86 The Incredible Growth Chap. 8

of its technical responsibilities, passing more control to subordinate groups, and leaving the board as the ultimate arbiter of policies and standards.

Among the subordinate groups, one stood out: the Internet Engineering Task Force (IETF). The IETF has survived reorganizations, and has taken over responsibility for generating new Internet technologies, revising older technologies, and creating, revising, and publishing standards documents. Most RFCs now originate within the IETF from committees, which are called working groups.

The IETF is partitioned into areas of interest, with an area director assigned to coordinate groups within each area. The IETF holds open meetings approximately three times per year. When it holds a meeting, thousands of people attend, most from com- mercial companies. Attendees are volunteers who attend to hear about the latest developments and participate in efforts to refine and improve the software.

We can summarize:

The group responsible for guiding the research and development of the Internet is known as the Internet Architecture Board (IAB). The primary subgroup responsible for technical matters is known as the Internet Engineering Task Force (IETF).

8.10 NSF Led Internet Expansion

During the years following the military adoption of TCP/IP, rapid growth contin- ued. Government agencies, such as NASA (the National Aeronautics And Space Ad- ministration) connected to the Internet. By the mid-1980s, the National Science Foun- dation (NSF) realized that eminence in science would soon demand computer communi- cation. Before computer networks, scientists exchanged ideas by publishing them in scientific journals, which took many months, sometimes years. Computer communica- tion was about to change the way scientists did research. A computer network makes it possible to share data as an experiment proceeds, making it possible for many other scientists to analyze the results without traveling to the site of the experiment.

Recognizing how important the Internet was becoming to science, NSF decided to fund further Internet growth. In 1985, NSF announced that it intended to connect researchers at 100 U.S. universities to the Internet. NSF advised the U.S. Congress of the plan, and received additional funds to support networking. NSF consulted experts in the field, devised a plan, and began a program that resulted in major changes to the Internet.

Scientists often use sophisticated, high-speed computers called supercomputers to analyze data from their experiments. Because supercomputers were expensive, NSF had established five supercomputer centers around the country. A scientist working on an NSF project had to travel to the nearest supercomputer center to process their data.

As the first step of Internet expansion, NSF built a Wide Area Network that inter- connected its five supercomputer centers. The network used TCP/IP, and provided a

Sec. 8.10 NSF Led Internet Expansion 87

connection to the Internet. Named NSFNET, the network was initially much smaller, and not any faster, than the ARPANET. Scientists found the network useful, but not exciting.

8.11 NSF Target: All Of Science And Engineering

Spurred by the success of CSNET and the initial NSFNET, NSF launched a new program to keep the U.S. competitive. The program had an ambitious goal:

NSF decided that the U.S. would not remain competitive unless every science and engineering researcher had Internet access.

To achieve the goal, NSF decided to use its funds to create a major new Internet that had significantly more capacity than the existing Internet. After examining avail- able technologies and reviewing its budget, NSF decided that it could not afford to pay for the entire project. Instead, it decided to offer partial support, in the form of federal grants. Companies and other organizations submitted written proposals to NSF to re- quest funding to work on the project.

NSF divided the grants into two types. First, NSF funded a group that wanted to build and operate a new high-speed Wide Area Network to connect parts of the Internet. The new WAN had to replace parts of the ARPANET as well as the original NSFNET. Second, NSF funded groups wanted to interconnect computers in a small region of the country and attach them to the new WAN. For example, NSF thought that each state might choose to apply as a group. Originally, the groups were referred to as NSF Re- gional Networks. Later, when it became clear that some of the groups spanned large geographic areas, NSF began referring to them as NSF Mid-Level Networks, but most professionals still called them regionals.

Because most universities or companies already had LANs connecting their com- puters, NSF decided to use its funds to help pay for long-distance connections; indivi- dual companies and schools paid for their own internal networks.

8.12 The NSFNET Backbone

NSF used a competitive bidding process when it awarded a grant for the new Inter- net WAN, which became known as the NSFNET backbone.† In 1987, NSF asked for proposals and used a panel of scientists to help assess them. After considering the alter- natives, NSF selected a joint proposal from three organizations: IBM, a computer manufacturer; MCI, a long-distance telephone company; and MERIT, an organization that built and operated a network connecting schools in Michigan.

The three groups cooperated to establish a new Wide Area Network that became the backbone of the Internet in the summer of 1988. MCI provided long-distance

†The term backbone is used as an analogy to a human spine that forms a central structure to which many other bones attach.

K392894.indd 115 03/08/18 12:07 pmComer, Douglas E.. The Internet Book : Everything You Need to Know about Computer Networking and How the Internet Works, CRC Press LLC, 2018. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=5502790. Created from apus on 2025-01-08 05:09:18.

C op

yr ig

ht ©

2 01

8. C

R C

P re

ss L

LC . A

ll rig

ht s

re se

rv ed

.

88 The Incredible Growth Chap. 8

transmission lines, IBM provided the dedicated computers and software used in the WAN, and MERIT operated the network. Most people referred to the new backbone using the same name applied to its predecessor, NSFNET.

8.13 On To The ANS Backbone

Eventually, as traffic on the new WAN reached capacity, NSF approved reorganiz- ing the network slightly and tripled the capacity of each transmission line. By the end of 1991, it became clear that the Internet was growing so fast that the NSFNET back- bone would soon be completely saturated. NSF realized that the federal government could not afford to pay for the Internet indefinitely. They wanted private industry to as- sume some responsibility. To solve the problem, IBM, MERIT, and MCI formed a nonprofit company named Advanced Networks and Services (ANS).

During 1992, ANS built a new Wide Area Network to serve as the Internet back- bone. Known as ANSNET, the WAN used transmission lines with 30 times the capacity of the NSFNET backbone it replaced. Figure 8.1 illustrates the ANSNET connections.

End-User Site MCI Point of Presence

Figure 8.1 The backbone of the Internet in 1995, known as ANSNET. Funding came from NSF, IBM, MCI, and MERIT.

Sec. 8.14 Commercialization 89

8.14 Commercialization

The move to ANSNET and associated regional networks represented a major shift in the Internet. For the first time, the Internet had become commercial. When DARPA and NSF provided Internet funding, they had to adhere to government rules. In particu- lar, the Internet had an Acceptable Use Policy (AUP), that allowed scientists and en- gineers to use it for research and teaching, but not for money-making activities. The le- gal rules changed when ANS, not the federal government, owned the transmission lines and computers that constituted the network. As ownership began to transfer to private companies, the Internet took its first steps toward commercialization and privatization.

NSF played a role in creating another aspect of the Internet that we now take for granted: Internet Service Providers (ISPs). In the early days, the Internet grew chaoti- cally. When a new site wanted to join, the site paid for a transmission line to the nearest Internet site. When NSF decided to fund regional networks that each provided service to a group of subscribers in an area, the connectivity paradigm changed. When a site wanted to join the Internet, the site contacted the appropriate regional network for service. When NSF transferred ownership of regional networks to the private sector, they each became an ISP.

8.15 Exponential Growth

As NSF connected scientists and engineers, the Internet grew incredibly fast. In 1983, the Internet connected 562 hosts. Ten years later, it connected over 1,200,000 hosts and was still growing quickly. Such staggering growth can best be understood by considering individual hosts:

By 1999, the Internet was growing so fast that, on the average, a host was added to the Internet every second. By 2006, the average exceed- ed ten hosts per second, and by 2016, the use of smart devices made growth impossible to measure.

Although the Internet did not grow at exactly the same rate in all years and most of the hosts were added in recent years, a trend of doubling can be identified. In round numbers, the Internet has experienced sustained growth of approximately 10 percent per month, doubling in size approximately every 10 months. Mathematicians call such growth exponential. The table in Figure 8.2 illustrates growth from 1983 through 2018.

Exponential growth has some interesting properties. For example, although the In- ternet has been around for many years, exponential growth means that approximately half the people connected to the Internet have gained access in the past year. Interest- ingly, that same statement could have been made in any previous year. In fact, the fol- lowing summarizes the incredible growth:

K392894.indd 116 03/08/18 12:07 pmComer, Douglas E.. The Internet Book : Everything You Need to Know about Computer Networking and How the Internet Works, CRC Press LLC, 2018. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=5502790. Created from apus on 2025-01-08 05:09:18.

C op

yr ig

ht ©

2 01

8. C

R C

P re

ss L

LC . A

ll rig

ht s

re se

rv ed

.

88 The Incredible Growth Chap. 8

transmission lines, IBM provided the dedicated computers and software used in the WAN, and MERIT operated the network. Most people referred to the new backbone using the same name applied to its predecessor, NSFNET.

8.13 On To The ANS Backbone

Eventually, as traffic on the new WAN reached capacity, NSF approved reorganiz- ing the network slightly and tripled the capacity of each transmission line. By the end of 1991, it became clear that the Internet was growing so fast that the NSFNET back- bone would soon be completely saturated. NSF realized that the federal government could not afford to pay for the Internet indefinitely. They wanted private industry to as- sume some responsibility. To solve the problem, IBM, MERIT, and MCI formed a nonprofit company named Advanced Networks and Services (ANS).

During 1992, ANS built a new Wide Area Network to serve as the Internet back- bone. Known as ANSNET, the WAN used transmission lines with 30 times the capacity of the NSFNET backbone it replaced. Figure 8.1 illustrates the ANSNET connections.

End-User Site MCI Point of Presence

Figure 8.1 The backbone of the Internet in 1995, known as ANSNET. Funding came from NSF, IBM, MCI, and MERIT.

Sec. 8.14 Commercialization 89

8.14 Commercialization

The move to ANSNET and associated regional networks represented a major shift in the Internet. For the first time, the Internet had become commercial. When DARPA and NSF provided Internet funding, they had to adhere to government rules. In particu- lar, the Internet had an Acceptable Use Policy (AUP), that allowed scientists and en- gineers to use it for research and teaching, but not for money-making activities. The le- gal rules changed when ANS, not the federal government, owned the transmission lines and computers that constituted the network. As ownership began to transfer to private companies, the Internet took its first steps toward commercialization and privatization.

NSF played a role in creating another aspect of the Internet that we now take for granted: Internet Service Providers (ISPs). In the early days, the Internet grew chaoti- cally. When a new site wanted to join, the site paid for a transmission line to the nearest Internet site. When NSF decided to fund regional networks that each provided service to a group of subscribers in an area, the connectivity paradigm changed. When a site wanted to join the Internet, the site contacted the appropriate regional network for service. When NSF transferred ownership of regional networks to the private sector, they each became an ISP.

8.15 Exponential Growth

As NSF connected scientists and engineers, the Internet grew incredibly fast. In 1983, the Internet connected 562 hosts. Ten years later, it connected over 1,200,000 hosts and was still growing quickly. Such staggering growth can best be understood by considering individual hosts:

By 1999, the Internet was growing so fast that, on the average, a host was added to the Internet every second. By 2006, the average exceed- ed ten hosts per second, and by 2016, the use of smart devices made growth impossible to measure.

Although the Internet did not grow at exactly the same rate in all years and most of the hosts were added in recent years, a trend of doubling can be identified. In round numbers, the Internet has experienced sustained growth of approximately 10 percent per month, doubling in size approximately every 10 months. Mathematicians call such growth exponential. The table in Figure 8.2 illustrates growth from 1983 through 2018.

Exponential growth has some interesting properties. For example, although the In- ternet has been around for many years, exponential growth means that approximately half the people connected to the Internet have gained access in the past year. Interest- ingly, that same statement could have been made in any previous year. In fact, the fol- lowing summarizes the incredible growth:

88 The Incredible Growth Chap. 8

transmission lines, IBM provided the dedicated computers and software used in the WAN, and MERIT operated the network. Most people referred to the new backbone using the same name applied to its predecessor, NSFNET.

8.13 On To The ANS Backbone

Eventually, as traffic on the new WAN reached capacity, NSF approved reorganiz- ing the network slightly and tripled the capacity of each transmission line. By the end of 1991, it became clear that the Internet was growing so fast that the NSFNET back- bone would soon be completely saturated. NSF realized that the federal government could not afford to pay for the Internet indefinitely. They wanted private industry to as- sume some responsibility. To solve the problem, IBM, MERIT, and MCI formed a nonprofit company named Advanced Networks and Services (ANS).

During 1992, ANS built a new Wide Area Network to serve as the Internet back- bone. Known as ANSNET, the WAN used transmission lines with 30 times the capacity of the NSFNET backbone it replaced. Figure 8.1 illustrates the ANSNET connections.

End-User Site MCI Point of Presence

Figure 8.1 The backbone of the Internet in 1995, known as ANSNET. Funding came from NSF, IBM, MCI, and MERIT.

Sec. 8.14 Commercialization 89

8.14 Commercialization

The move to ANSNET and associated regional networks represented a major shift in the Internet. For the first time, the Internet had become commercial. When DARPA and NSF provided Internet funding, they had to adhere to government rules. In particu- lar, the Internet had an Acceptable Use Policy (AUP), that allowed scientists and en- gineers to use it for research and teaching, but not for money-making activities. The le- gal rules changed when ANS, not the federal government, owned the transmission lines and computers that constituted the network. As ownership began to transfer to private companies, the Internet took its first steps toward commercialization and privatization.

NSF played a role in creating another aspect of the Internet that we now take for granted: Internet Service Providers (ISPs). In the early days, the Internet grew chaoti- cally. When a new site wanted to join, the site paid for a transmission line to the nearest Internet site. When NSF decided to fund regional networks that each provided service to a group of subscribers in an area, the connectivity paradigm changed. When a site wanted to join the Internet, the site contacted the appropriate regional network for service. When NSF transferred ownership of regional networks to the private sector, they each became an ISP.

8.15 Exponential Growth

As NSF connected scientists and engineers, the Internet grew incredibly fast. In 1983, the Internet connected 562 hosts. Ten years later, it connected over 1,200,000 hosts and was still growing quickly. Such staggering growth can best be understood by considering individual hosts:

By 1999, the Internet was growing so fast that, on the average, a host was added to the Internet every second. By 2006, the average exceed- ed ten hosts per second, and by 2016, the use of smart devices made growth impossible to measure.

Although the Internet did not grow at exactly the same rate in all years and most of the hosts were added in recent years, a trend of doubling can be identified. In round numbers, the Internet has experienced sustained growth of approximately 10 percent per month, doubling in size approximately every 10 months. Mathematicians call such growth exponential. The table in Figure 8.2 illustrates growth from 1983 through 2018.

Exponential growth has some interesting properties. For example, although the In- ternet has been around for many years, exponential growth means that approximately half the people connected to the Internet have gained access in the past year. Interest- ingly, that same statement could have been made in any previous year. In fact, the fol- lowing summarizes the incredible growth:

K392894.indd 117 03/08/18 12:07 pmComer, Douglas E.. The Internet Book : Everything You Need to Know about Computer Networking and How the Internet Works, CRC Press LLC, 2018. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=5502790. Created from apus on 2025-01-08 05:09:18.

C op

yr ig

ht ©

2 01

8. C

R C

P re

ss L

LC . A

ll rig

ht s

re se

rv ed

.

90 The Incredible Growth Chap. 8

At any time from 1983 through 2007, approximately half the Internet growth occurred in the previous 10 to 12 months.

What happened to growth starting in 2007? The answer is the smart phone hap- pened. The numbers listed in Figure 8.2 are a count of host computers with permanent IP addresses. As Chapter 26 explains, the cell phone system issues temporary IP ad- dresses to smart phones, making it impossible to obtain an accurate count. Thus, numbers in Figure 8.2 for years after 2007 are low, and in January of 2018, fewer hosts had permanent addresses than the year before.

Year Computers Year Computers 1981 213 1982 235 1983 562 1984 1,024 1985 1,961 1986 2,308 1987 5,089 1988 28,174 1989 80,000 1990 313,000 1991 535,000 1992 727,000 1993 1,313,000 1994 2,217,000 1995 4,852,000 1996 9,472,000 1997 16,146,000 1998 29,670,000 1999 43,230,000

2000 72,398,092 2001 109,574,429 2002 147,344,723 2003 171,638,297 2004 233,101,481 2005 317,646,084 2006 394,991,609 2007 433,193,199 2008 541,677,360 2009 625,226,456 2010 732,740,444 2011 818,374,269 2012 888,239,420 2013 963,518,598 2014 1,010,251,829 2015 1,012,706,608 2016 1,048,766,623 2017 1,062,660,523 2018 1,003,604,363

Figure 8.2 Internet hosts with permanent IP addresses each year from 1983 through 2018. The counts for years 2007 on are low because smart phones cannot be counted accurately.

Sec. 8.16 When Will Growth End? 91

8.16 When Will Growth End?

At various times in the past, people have predicted the imminent collapse of the In- ternet by observing that some small piece of the technology was reaching its limit. By 1990, for example, someone had predicted that the Internet could not survive past March of 1993. In 1995, a group predicted that the Internet would collapse in the sum- mer of 1997. Then in 1999, another group predicted collapse in 2004. The predictions of doom have been incorrect, and the Internet keeps growing. Each time the traffic has approached the capacity of a backbone network, a new backbone technology has been developed and deployed with significantly more capacity. When the traffic approached the capacity of the systems that forward data across the Internet, faster systems have been created. At one time, a group observed that Internet growth must be curtailed be- cause it was about to overtake the worldwide production of computers. However, the group focused on PCs, and was surprised when tablets and smart phones came along.

Another group calculated the end of growth by carefully estimating the world po- pulation growth and the rate at which users were being added to the Internet. They con- fidently predicted a date when Internet growth would stop because every person on earth would have a computer hooked to the Internet. Since that prediction two things occurred. First, smart devices mean many people have multiple devices for use in their business and personal lives. Second, as Chapter 24 explains, the latest Internet expan- sion is occurring because users are connecting many small devices to the Internet.

The point is that both technology and the way we use the Internet keeps changing, making accurate prediction difficult.

Although researchers agree that growth cannot continue unchecked forever, The Internet has persisted in growing beyond predictions of its end.

EXERCISES

8.1 Various groups estimate the number of Internet users. Search online to obtain an estimate of how many users worldwide access the Internet on an average day.

8.2 Extend the previous exercise and find out what percentage of the world’s population has In- ternet access.

8.3 Search the Internet to find programs and projects that are being undertaken to deliver Inter- net services to the few groups of people who do not yet have access.

K392894.indd 118 03/08/18 12:07 pmComer, Douglas E.. The Internet Book : Everything You Need to Know about Computer Networking and How the Internet Works, CRC Press LLC, 2018. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=5502790. Created from apus on 2025-01-08 05:09:18.

C op

yr ig

ht ©

2 01

8. C

R C

P re

ss L

LC . A

ll rig

ht s

re se

rv ed

.

90 The Incredible Growth Chap. 8

At any time from 1983 through 2007, approximately half the Internet growth occurred in the previous 10 to 12 months.

What happened to growth starting in 2007? The answer is the smart phone hap- pened. The numbers listed in Figure 8.2 are a count of host computers with permanent IP addresses. As Chapter 26 explains, the cell phone system issues temporary IP ad- dresses to smart phones, making it impossible to obtain an accurate count. Thus, numbers in Figure 8.2 for years after 2007 are low, and in January of 2018, fewer hosts had permanent addresses than the year before.

Year Computers Year Computers 1981 213 1982 235 1983 562 1984 1,024 1985 1,961 1986 2,308 1987 5,089 1988 28,174 1989 80,000 1990 313,000 1991 535,000 1992 727,000 1993 1,313,000 1994 2,217,000 1995 4,852,000 1996 9,472,000 1997 16,146,000 1998 29,670,000 1999 43,230,000

2000 72,398,092 2001 109,574,429 2002 147,344,723 2003 171,638,297 2004 233,101,481 2005 317,646,084 2006 394,991,609 2007 433,193,199 2008 541,677,360 2009 625,226,456 2010 732,740,444 2011 818,374,269 2012 888,239,420 2013 963,518,598 2014 1,010,251,829 2015 1,012,706,608 2016 1,048,766,623 2017 1,062,660,523 2018 1,003,604,363

Figure 8.2 Internet hosts with permanent IP addresses each year from 1983 through 2018. The counts for years 2007 on are low because smart phones cannot be counted accurately.

Sec. 8.16 When Will Growth End? 91

8.16 When Will Growth End?

At various times in the past, people have predicted the imminent collapse of the In- ternet by observing that some small piece of the technology was reaching its limit. By 1990, for example, someone had predicted that the Internet could not survive past March of 1993. In 1995, a group predicted that the Internet would collapse in the sum- mer of 1997. Then in 1999, another group predicted collapse in 2004. The predictions of doom have been incorrect, and the Internet keeps growing. Each time the traffic has approached the capacity of a backbone network, a new backbone technology has been developed and deployed with significantly more capacity. When the traffic approached the capacity of the systems that forward data across the Internet, faster systems have been created. At one time, a group observed that Internet growth must be curtailed be- cause it was about to overtake the worldwide production of computers. However, the group focused on PCs, and was surprised when tablets and smart phones came along.

Another group calculated the end of growth by carefully estimating the world po- pulation growth and the rate at which users were being added to the Internet. They con- fidently predicted a date when Internet growth would stop because every person on earth would have a computer hooked to the Internet. Since that prediction two things occurred. First, smart devices mean many people have multiple devices for use in their business and personal lives. Second, as Chapter 24 explains, the latest Internet expan- sion is occurring because users are connecting many small devices to the Internet.

The point is that both technology and the way we use the Internet keeps changing, making accurate prediction difficult.

Although researchers agree that growth cannot continue unchecked forever, The Internet has persisted in growing beyond predictions of its end.

EXERCISES

8.1 Various groups estimate the number of Internet users. Search online to obtain an estimate of how many users worldwide access the Internet on an average day.

8.2 Extend the previous exercise and find out what percentage of the world’s population has In- ternet access.

8.3 Search the Internet to find programs and projects that are being undertaken to deliver Inter- net services to the few groups of people who do not yet have access.

90 The Incredible Growth Chap. 8

At any time from 1983 through 2007, approximately half the Internet growth occurred in the previous 10 to 12 months.

What happened to growth starting in 2007? The answer is the smart phone hap- pened. The numbers listed in Figure 8.2 are a count of host computers with permanent IP addresses. As Chapter 26 explains, the cell phone system issues temporary IP ad- dresses to smart phones, making it impossible to obtain an accurate count. Thus, numbers in Figure 8.2 for years after 2007 are low, and in January of 2018, fewer hosts had permanent addresses than the year before.

Year Computers Year Computers 1981 213 1982 235 1983 562 1984 1,024 1985 1,961 1986 2,308 1987 5,089 1988 28,174 1989 80,000 1990 313,000 1991 535,000 1992 727,000 1993 1,313,000 1994 2,217,000 1995 4,852,000 1996 9,472,000 1997 16,146,000 1998 29,670,000 1999 43,230,000

2000 72,398,092 2001 109,574,429 2002 147,344,723 2003 171,638,297 2004 233,101,481 2005 317,646,084 2006 394,991,609 2007 433,193,199 2008 541,677,360 2009 625,226,456 2010 732,740,444 2011 818,374,269 2012 888,239,420 2013 963,518,598 2014 1,010,251,829 2015 1,012,706,608 2016 1,048,766,623 2017 1,062,660,523 2018 1,003,604,363

Figure 8.2 Internet hosts with permanent IP addresses each year from 1983 through 2018. The counts for years 2007 on are low because smart phones cannot be counted accurately.

Sec. 8.16 When Will Growth End? 91

8.16 When Will Growth End?

At various times in the past, people have predicted the imminent collapse of the In- ternet by observing that some small piece of the technology was reaching its limit. By 1990, for example, someone had predicted that the Internet could not survive past March of 1993. In 1995, a group predicted that the Internet would collapse in the sum- mer of 1997. Then in 1999, another group predicted collapse in 2004. The predictions of doom have been incorrect, and the Internet keeps growing. Each time the traffic has approached the capacity of a backbone network, a new backbone technology has been developed and deployed with significantly more capacity. When the traffic approached the capacity of the systems that forward data across the Internet, faster systems have been created. At one time, a group observed that Internet growth must be curtailed be- cause it was about to overtake the worldwide production of computers. However, the group focused on PCs, and was surprised when tablets and smart phones came along.

Another group calculated the end of growth by carefully estimating the world po- pulation growth and the rate at which users were being added to the Internet. They con- fidently predicted a date when Internet growth would stop because every person on earth would have a computer hooked to the Internet. Since that prediction two things occurred. First, smart devices mean many people have multiple devices for use in their business and personal lives. Second, as Chapter 24 explains, the latest Internet expan- sion is occurring because users are connecting many small devices to the Internet.

The point is that both technology and the way we use the Internet keeps changing, making accurate prediction difficult.

Although researchers agree that growth cannot continue unchecked forever, The Internet has persisted in growing beyond predictions of its end.

EXERCISES

8.1 Various groups estimate the number of Internet users. Search online to obtain an estimate of how many users worldwide access the Internet on an average day.

8.2 Extend the previous exercise and find out what percentage of the world’s population has In- ternet access.

8.3 Search the Internet to find programs and projects that are being undertaken to deliver Inter- net services to the few groups of people who do not yet have access.

K392894.indd 119 03/08/18 12:07 pmComer, Douglas E.. The Internet Book : Everything You Need to Know about Computer Networking and How the Internet Works, CRC Press LLC, 2018. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=5502790. Created from apus on 2025-01-08 05:09:18.

C op

yr ig

ht ©

2 01

8. C

R C

P re

ss L

LC . A

ll rig

ht s

re se

rv ed

.

Inside The Internet An Explanation Of The Underlying Technology And Basic Capabilities Of The Infrastructure

K392894.indd 120 03/08/18 12:07 pmComer, Douglas E.. The Internet Book : Everything You Need to Know about Computer Networking and How the Internet Works, CRC Press LLC, 2018. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/apus/detail.action?docID=5502790. Created from apus on 2025-01-08 05:09:18.

C op

yr ig

ht ©

2 01

8. C

R C

P re

ss L

LC . A

ll rig

ht s

re se

rv ed

.