OSI: Internet, which was not
From the translator: This is a translation of the OSI: The Internet article that was not Andrew L. Russell, originally published in the IEEE Spectrum magazine.
If everything went according to plan , the Internet that we know would never have arisen. This plan, developed 35 years ago, envisaged the creation of a holistic set of standards for computer networks Open Systems Interconnection, OSI.
Its creators were a separate group of computer industry representatives from the United Kingdom, France and the United States of America. They imagined a complete, open and multi-layered system that would allow users around the world to easily exchange data and thereby open up new opportunities for developing cooperation and commerce.
')
At that time, their vision seemed the only correct one. Thousands of engineers and legislators around the world have become involved in the process of becoming OSI standards. Soon they had the support of all stakeholders: computer manufacturers, telephone companies, regulators, national governments, international standards agencies, academic researchers, even the US Department of Defense. By the mid-1980s, world recognition of OSI was obvious.
However, by the early 1990s, the project had almost died out, faced with a cheaper and flexible, albeit less complete, alternative: the Internet stack consisting of the Transmission Control Protocol and the Internet Protocol. When OSI's positions weakened, one of the leading supporters of the Internet, Einar Stefferud, said with satisfaction: “OSI is a beautiful dream, and TCP / IP is already a reality!” (“OSI is a beautiful dream, and TCP / IP is living it! ”) .
What happened to the “beautiful dream”? While the triumphant history of the Internet is well documented by its creators and historians who worked with them, OSI has been forgotten by everyone, with the exception of only a handful of war veterans of Internet-OSI standards. To understand the reason, we need to plunge into the early history of computer networks, a time when the annoying problems of digital unity and global connectivity agitated the minds of computer scientists, engineers from telecommunications companies, lawmakers and corporate boards. And in order to better understand this story, you will have to fold everything that you already know about the Internet for a while. Try to imagine if you can, that the Internet never existed.
The story begins in the 1960s . Erected the Berlin Wall. Free Speech Movement flourishes at Berkeley. US soldiers fighting in Vietnam. A system of interconnection of digital computers are still in infancy and are subjected to intensive wide-ranging studies, with dozens (and soon hundreds) of people in academic circles, industry and government engaged in large research projects.
The most promising of all research projects included a new approach to data transmission, called packet switching. Invented independently by Paul Baran (Paul Baran) in Rand Corp. in the United States and Donald Davies at the National Physical Laboratory in England, the packet switching method involved breaking up messages into separate blocks, or packets, that could be routed independently over the many available network channels. The host computer would collect the packets back to their original form. Baran and Davis both believed that packet switching could be more reliable and efficient than circuit switching, an old technology used in telephone systems that requires a dedicated channel for each connection.
Researchers funded by the Advanced Research Projects Agency (ARPA) of the US Department of Defense created the first packet-switched network, called ARPANET , in 1969. Soon, other organizations, in particular the computer giant IBM and a number of telephone monopolists in Europe, began to carry out ambitious plans to create packet-switched networks. Considering the possibility of digital unity between computers and communications, these companies are still more concerned about maintaining the level of income derived from their existing business. As a result, IBM and telephone monopolies preferred packet switching to rely on “virtual circuits” —a design that replicates the technical and organizational methods of circuit-switched systems.
With such a number of ideas from stakeholders, everyone agreed on the need for switching packets of some form of international standardization. The first attempts began in 1972, when the International Network Working Group (INWG) was formed . Vint Cerf was its first chairman; other active participants were Alex McKenzie in the United States, Donald Davis and Roger Scantlebury in England, and also Luis Pouzin and Hubert Zimmermann in France.
The task of INWG was to promote the idea of ​​packet switching based on "datagrams" developed by Pouzan. As he explained to me during our meeting in Paris in 2012, “The essence of the datagrams is the absence of a connection. This means that no relationship is created between the sender and receiver. They are independent of each other, like photons. ” This was a radical proposition, especially when compared to virtual connections preferred by IBM and telecom engineers.
The INWG regularly collected and exchanged technical articles in order to bring the development into line with the idea of ​​datagrams, especially the transport protocol, a key mechanism for exchanging packets between networks of various types. After several years of debate and discussion, the group finally reached an agreement in 1975, and Cerf and Pouzan sent the documentation on their protocol to the international telecommunications standards supervision organization, the International Telegraph and Telephone Consultative Commitee by its French acronym, CCITT) .
The committee, which was dominated by telecom engineers, rejected the INWG proposal as too risky and not run-in. The surf and colleagues were very disappointed. Puzan, the leader of Cyclades, a French research project on packet networks , sarcastically remarked that CCITT is "not against packet switching if it also looks like circuit switching." And when Puzan at conferences complained about the tactic of “wringing hands” by the “state monopolies”, everyone understood what he said about the French regulator [CCITT]. French bureaucrats did not like the sincerity of their compatriot, which led to the gradual depletion of Cyclades public funding from 1975 to 1978, after which Puzan left the project.
For his part, Cerf was so discouraged by this experience of creating international standards that by the end of 1975 he left the post of chairman of the INWG. He also left the department at Stanford and accepted the offer of Robert Kahn to work at ARPA. Surf and Kahn had already developed Puzan’s development and published the details of their “ transfer management program ” a year earlier in IEEE Transactions on Communications. This development became the technical foundation of the Internet, a term that was later adopted to refer to a network of networks that used the TCP / IP ARPA stack. In the following years, the two led the development of Internet protocols in their environment: a small community of ARPA contractors.
Serf's departure marked a break in the INWG. While Cerf and ARPA contractors eventually formed the backbone of the Internet in the eighties, many of the remaining INWG veterans regrouped and joined the international alliance that was formed under the banner of OSI, and these two camps became hostile.
OSI was developed by the committee , but this fact alone is not enough to bury the project — after all, many successful standards begin in this way. However, it is important to understand what happened next.
In 1977, representatives of the British computer industry proposed the creation of a new standards committee dedicated to packet-switched networks within the framework of the International Organization for Standardization (ISO) , an independent non-governmental association established after World War II. Unlike CCITT, ISO did not deal exclusively with telecommunications: the breadth of the topics included the technical committee TC1 engaged in thread thread standards and TC17 working with steel. Also, unlike CCITT, ISO already had computer standards committees and could be more receptive to the ideas of datagrams without connections.
The British proposal, which was supported by representatives from the United States and France, called for the creation of "network standards necessary for open interaction." These standards, according to the British, could offer an alternative to the "closed" closed "systems" of traditional computers, developed "without regard to the possibility of their joint work." The concept of open interaction was both strategic and technical, and spoke about their desire to create competition to market leaders, specifically IBM and telecommunications monopolies.
As expected, ISO agreed with the British proposal and appointed Charles Bachman, the US database expert, as the head of the committee. Highly respected in computer computing circles, Buckman received four years earlier the prestigious Turing Award for his work on the Integrated Data Store database management system.
When I interviewed Buckman in 2011, he described the “architectural vision” he brought to OSI, the vision was inspired by his work with databases in general and the architecture of the IBM Systems Network Architecture in particular. He began by defining a reference model that divided the various tasks of communication between computers into multiple levels. For example, the physical transmission medium (such as copper cables) gets into the first level, transport protocols for moving data fall into the fourth level, and applications (such as email and file transfer) are at the seventh level. After the approval of the architecture based on the levels, it was possible to start developing specific protocols.
Buckman's development was significantly different from Systems Network Architecture: while IBM built terminal-to-computer architecture, Buckman connected peer-to-peer computers. This approach has made the project very attractive for companies like General Motors, a leading OSI supporter in the eighties. GM had dozens of factories and hundreds of suppliers using a mixture of low-compatible software and hardware systems. The Buckman scheme would allow computers and networks of different proprietary types to be integrated - if they followed OSI standards.
The layered reference OSI model also provided an important organizational opportunity: modularity. That is, the division into levels made it possible to divide the work on protocols between the committees. Obviously, the Buckman model was just the beginning. To become an international standard, each proposal had to go through four stages, starting with a working draft, then a draft of the proposed international standard, then a draft of the international standard and, finally, an international standard. Creating a consensus around the OSI reference model and related standards required an extraordinary number of plenary meetings and committee meetings.
The first plenary session of OSI lasted three days, from February 28 to March 2, 1978. Many delegates from ten countries and observers from four international organizations gathered. All participants had their own market interests and ready-made developments. At the same time, delegates from one country could have completely different goals. Many of the attendees were INWG veterans who remained cautiously optimistic about the possibility of snatching the future of networks from IBM and the telecommunications monopolies, which, quite obviously, planned to dominate the new market.
At the same time, IBM representatives, led by Joseph De Blasi , a very capable director of standards at the company, skillfully guided the discussion, keeping the development of OSI within IBM's business interests. Informatics John Day, who developed the protocols for ARPANET, was a key member of the US delegation. In his 2008 book, Patterns in Network Architecture, ed. Prentice Hall , Day recalls that IBM representatives skillfully intervened in disputes between delegates who fought for a piece of cake ... IBM turned them about as they wanted. It was a fascinating sight. ”
Despite the sticks being inserted into the wheels, Buckman's leadership drove OSI along a perilous path from concept to reality. Buckman and Hubert Zimmerman (veteran Cyclades and INWG) made an alliance with telecommunications engineers from CCITT. But this partnership with difficulty overcame the fundamental incompatibility of their views. Zimmerman and his computer science colleagues, inspired by Puzan’s datagram design, fought for connectionless protocols, while telecommunications professionals insisted on virtual circuits. Instead of resolving the dispute, they agreed to include both options within OSI, thereby increasing its size and complexity.
This complex alliance of informatics and communications specialists published an international standard for the OSI reference model in 1984. Separate OSI standards for transport protocols, email, electronic directories, network management, and many other functions soon followed. OSI was beginning to show signs of its inevitability. Leading computer companies, such as Digital Equipment Corp., Honeywell, and IBM, were by then very interested in OSI, along with the European Economic Community, governments in Europe, North America and Asia.
Even the US government - the leading sponsor of Internet protocols that are incompatible with OSI - has joined the popular project. The Department of Defense formally adopted a conclusion from the recommendations of the National Research Council (1985) of 1985 on the transition from TCP / IP to OSI. At the same time, the Department of Commerce issued a mandate in 1988 prescribing the use of OSI standards on all computers procured by the US government after August 1990.
Although such decrees look like the work of distorting bureaucrats, it should be recalled that in the eighties the Internet was just a research network: the network grew rapidly, but its administrators did not allow commercial traffic or third-party providers to the government-funded highway until 1992. For companies and other large organizations that wanted to exchange data between different types of computers or networks, OSI remained the only option.
This, of course, does not end the story . By the end of the eighties, resentment at the slow development of OSI had reached a critical point. During an assembly in Europe in 1989, OSI advocate Brian Carpenter gave a speech entitled “Is OSI Too Late Late?”. That time, as he recalls in his recent memoirs , "was the only time in my life" when he was honored with a "standing ovation at a technical conference." Two years later, a French network expert and former member of the INWG, Puzan , in an essay entitled “Ten Years of OSI — Maturity or Infancy?” (“Ten Years of OSI — Maturity or Infancy?”) Outlined growing uncertainty: OSI is constantly recommended as a major solution. However, it is much easier and faster to implement a homogeneous network based on proprietary architectures or to combine heterogeneous systems using TCP-based products. ” Even for OSI supporters, the Internet looked increasingly attractive.
The feeling of doom was exacerbated, progress slowed down and in the mid-nineties, OSI's beautiful dream finally ended. The fatal flaw of the process, no matter how ironic it was, was a commitment to openness. The formal rules of international standardization gave any interested party the right to participate in the development, which led to general tension, incompatibility of views and gave way to subversive activities.
OSI’s first chairman, Buckman, had been expecting these problems from the start. In a speech at a conference in 1978, he expressed concern about OSI’s chances of success: "The [organizational challenge] is great. Technical problems are more than anything previously known in information systems. And political problems will make the most clever statesman sweat Can you imagine an attempt to bring together in the foreseeable future representatives of a dozen leading competing computer corporations, and a dozen telephone companies, and PTT [public telecommunications monitors Olii - prim.avtora], and technical experts from ten different states? "
Despite the efforts of Buckman and other people, the burden of organizational problems did not subside. Hundreds of engineers attended the meetings of various committees and OSI working groups, and the bureaucratic procedures used to structure the discussion prevented the creation of standards quickly. They argued about everything: even about the trivial nuances of the language (for example, about the difference between the turns of “you obey” and “you should obey”). Much more complicated questions separated experts from computer science and telecommunications, whose technical and business plans were in conflict. Thus, openness and modularity - the key principles of project coordination - have led to the collapse of OSI.
At the same time, the Internet flourished . With sufficient funding from the US government, Cerf and Kahn and their colleagues were protected from the forces of international politics and economics. ARPA and the Defense Communications Agency (Defense Communications Agency) accelerated the introduction of the Internet in the early eighties, funding scientists to implement Internet protocols in popular operating systems, such as the Unix variant from the University of California, Berkeley. Then, on January 1, 1983, ARPA stopped supporting the ARPANET inter-node protocol, thereby forcing contractors to use TCP / IP if they wanted to stay connected; This day is known as the " birthday of the Internet ."
Thus, while many users were still waiting for OSI to become a solution for global networking, an increasing number of people began to use TCP / IP as a practical, albeit temporary, means of networking.
Engineers who joined the Internet community in the 1980s often did not understand OSI, mocking its erroneous monstrosity created by incomprehensible European bureaucrats. Engineer Marshall Rose wrote in his 1990 textbook that “The Internet community is trying hard to ignore the OSI community. OSI technologies are generally ugly compared to Internet technologies. "
Unfortunately, the negative attitude of the Internet community has also led to the rejection of good technical ideas OSI. A classic example is the "palace coup" in 1992. , OSI, (Internet Activities Board) (Internet Engineering Task Force, IETF), . 1992 . , , TCP IP, , , OSI. - , .
TCP/IP , : . , OSI, , , ISO , . (Mark Levilion), IBM, 2012 OSI TCP/IP : „ , - , . , , , , . , ?“
, - . OSI, „“ . , „ “ , .
, OSI „ “ . , , , OSI: ( ) , , .
„“ „“, OSI , , . , „“. OSI - . OSI , . , , ?
“'Rough Consensus and Running Code' and the Internet-OSI Standards War” 2006 , IEEE Annals of the History of Computing. Open Standards and the Digital Age: History, Ideology, and Networks (Cambridge University Press, 2014).
(Janet Abbate) Inventing the Internet (MIT Press, 1999) , , .
“ INWG and the Conception of the Internet: An Eyewitness Account ” , IEEE Annals of the History of Computing 2011 , International Networking Working Group, ().
- Entrepreneurial Capitalism and Innovation: A History of Computer Communications, 1968–1988 (James Pelkey) , 1980- , , OSI . Ethernet .
How TCP / IP surpassed Open Systems Interconnection standards, becoming a protocol for global computer networks.
If everything went according to plan , the Internet that we know would never have arisen. This plan, developed 35 years ago, envisaged the creation of a holistic set of standards for computer networks Open Systems Interconnection, OSI.
Its creators were a separate group of computer industry representatives from the United Kingdom, France and the United States of America. They imagined a complete, open and multi-layered system that would allow users around the world to easily exchange data and thereby open up new opportunities for developing cooperation and commerce.
')
Photo: INRIA
Just plug in : Explorer Hubert Zimmerman [left - approx. author] talks about computer networks to representatives of the French government at a meeting in 1974. Zimmerman will subsequently play a key role in the development of OSI standards.
At that time, their vision seemed the only correct one. Thousands of engineers and legislators around the world have become involved in the process of becoming OSI standards. Soon they had the support of all stakeholders: computer manufacturers, telephone companies, regulators, national governments, international standards agencies, academic researchers, even the US Department of Defense. By the mid-1980s, world recognition of OSI was obvious.
However, by the early 1990s, the project had almost died out, faced with a cheaper and flexible, albeit less complete, alternative: the Internet stack consisting of the Transmission Control Protocol and the Internet Protocol. When OSI's positions weakened, one of the leading supporters of the Internet, Einar Stefferud, said with satisfaction: “OSI is a beautiful dream, and TCP / IP is already a reality!” (“OSI is a beautiful dream, and TCP / IP is living it! ”) .
Photo: INRIA
1961: Paul Baran (Paul Baran) in Rand Corp. begins to develop its concept of "switching message blocks" as a method of transmitting data over computer networks.
What happened to the “beautiful dream”? While the triumphant history of the Internet is well documented by its creators and historians who worked with them, OSI has been forgotten by everyone, with the exception of only a handful of war veterans of Internet-OSI standards. To understand the reason, we need to plunge into the early history of computer networks, a time when the annoying problems of digital unity and global connectivity agitated the minds of computer scientists, engineers from telecommunications companies, lawmakers and corporate boards. And in order to better understand this story, you will have to fold everything that you already know about the Internet for a while. Try to imagine if you can, that the Internet never existed.
The story begins in the 1960s . Erected the Berlin Wall. Free Speech Movement flourishes at Berkeley. US soldiers fighting in Vietnam. A system of interconnection of digital computers are still in infancy and are subjected to intensive wide-ranging studies, with dozens (and soon hundreds) of people in academic circles, industry and government engaged in large research projects.
1965: Donald W. Davies, working independently of Baran, develops his packet-switching network.
The most promising of all research projects included a new approach to data transmission, called packet switching. Invented independently by Paul Baran (Paul Baran) in Rand Corp. in the United States and Donald Davies at the National Physical Laboratory in England, the packet switching method involved breaking up messages into separate blocks, or packets, that could be routed independently over the many available network channels. The host computer would collect the packets back to their original form. Baran and Davis both believed that packet switching could be more reliable and efficient than circuit switching, an old technology used in telephone systems that requires a dedicated channel for each connection.
Researchers funded by the Advanced Research Projects Agency (ARPA) of the US Department of Defense created the first packet-switched network, called ARPANET , in 1969. Soon, other organizations, in particular the computer giant IBM and a number of telephone monopolists in Europe, began to carry out ambitious plans to create packet-switched networks. Considering the possibility of digital unity between computers and communications, these companies are still more concerned about maintaining the level of income derived from their existing business. As a result, IBM and telephone monopolies preferred packet switching to rely on “virtual circuits” —a design that replicates the technical and organizational methods of circuit-switched systems.
1969 : ARPANET, the first packet-switched network, is established in the United States.
1970 : Estimated profit in the US computer networking market: US $ 46 million.
1971 : Cyclades packet network project launched in France.
With such a number of ideas from stakeholders, everyone agreed on the need for switching packets of some form of international standardization. The first attempts began in 1972, when the International Network Working Group (INWG) was formed . Vint Cerf was its first chairman; other active participants were Alex McKenzie in the United States, Donald Davis and Roger Scantlebury in England, and also Luis Pouzin and Hubert Zimmermann in France.
The task of INWG was to promote the idea of ​​packet switching based on "datagrams" developed by Pouzan. As he explained to me during our meeting in Paris in 2012, “The essence of the datagrams is the absence of a connection. This means that no relationship is created between the sender and receiver. They are independent of each other, like photons. ” This was a radical proposition, especially when compared to virtual connections preferred by IBM and telecom engineers.
The INWG regularly collected and exchanged technical articles in order to bring the development into line with the idea of ​​datagrams, especially the transport protocol, a key mechanism for exchanging packets between networks of various types. After several years of debate and discussion, the group finally reached an agreement in 1975, and Cerf and Pouzan sent the documentation on their protocol to the international telecommunications standards supervision organization, the International Telegraph and Telephone Consultative Commitee by its French acronym, CCITT) .
1972 : The International Network Working Group (INWG) meets to develop an international standard for packet-switched networks, including [from left to right - author] Louis Puzan, Vint Cerf, Alex McKinsey, Hubert Zimmerman and Donald Davis.
The committee, which was dominated by telecom engineers, rejected the INWG proposal as too risky and not run-in. The surf and colleagues were very disappointed. Puzan, the leader of Cyclades, a French research project on packet networks , sarcastically remarked that CCITT is "not against packet switching if it also looks like circuit switching." And when Puzan at conferences complained about the tactic of “wringing hands” by the “state monopolies”, everyone understood what he said about the French regulator [CCITT]. French bureaucrats did not like the sincerity of their compatriot, which led to the gradual depletion of Cyclades public funding from 1975 to 1978, after which Puzan left the project.
1974 : Screw Cerf and Robert Kahn (Robert Kahn) publish the article “A Protocol for Packet Network Intercommunication,” [ Protocol for Packet Network Interconnection - approx. Translator] in IEEE Transactions on Communications.
For his part, Cerf was so discouraged by this experience of creating international standards that by the end of 1975 he left the post of chairman of the INWG. He also left the department at Stanford and accepted the offer of Robert Kahn to work at ARPA. Surf and Kahn had already developed Puzan’s development and published the details of their “ transfer management program ” a year earlier in IEEE Transactions on Communications. This development became the technical foundation of the Internet, a term that was later adopted to refer to a network of networks that used the TCP / IP ARPA stack. In the following years, the two led the development of Internet protocols in their environment: a small community of ARPA contractors.
Serf's departure marked a break in the INWG. While Cerf and ARPA contractors eventually formed the backbone of the Internet in the eighties, many of the remaining INWG veterans regrouped and joined the international alliance that was formed under the banner of OSI, and these two camps became hostile.
OSI was developed by the committee , but this fact alone is not enough to bury the project — after all, many successful standards begin in this way. However, it is important to understand what happened next.
In 1977, representatives of the British computer industry proposed the creation of a new standards committee dedicated to packet-switched networks within the framework of the International Organization for Standardization (ISO) , an independent non-governmental association established after World War II. Unlike CCITT, ISO did not deal exclusively with telecommunications: the breadth of the topics included the technical committee TC1 engaged in thread thread standards and TC17 working with steel. Also, unlike CCITT, ISO already had computer standards committees and could be more receptive to the ideas of datagrams without connections.
The British proposal, which was supported by representatives from the United States and France, called for the creation of "network standards necessary for open interaction." These standards, according to the British, could offer an alternative to the "closed" closed "systems" of traditional computers, developed "without regard to the possibility of their joint work." The concept of open interaction was both strategic and technical, and spoke about their desire to create competition to market leaders, specifically IBM and telecommunications monopolies.
Layered Approach : The OSI Reference Model (left) divides the connection between computers into seven levels, from physics on the first to applications on the seventh. Although less rigorous, the TCP / IP approach can also be represented as levels, as shown on the right.
As expected, ISO agreed with the British proposal and appointed Charles Bachman, the US database expert, as the head of the committee. Highly respected in computer computing circles, Buckman received four years earlier the prestigious Turing Award for his work on the Integrated Data Store database management system.
When I interviewed Buckman in 2011, he described the “architectural vision” he brought to OSI, the vision was inspired by his work with databases in general and the architecture of the IBM Systems Network Architecture in particular. He began by defining a reference model that divided the various tasks of communication between computers into multiple levels. For example, the physical transmission medium (such as copper cables) gets into the first level, transport protocols for moving data fall into the fourth level, and applications (such as email and file transfer) are at the seventh level. After the approval of the architecture based on the levels, it was possible to start developing specific protocols.
1974 : IBM launches a packet-switched network called Systems Network Architecture.
1975 : INWG sends a proposal to the CCITT, the committee rejects it. Surf leaves office at INWG.
1976 : CCITT publishes Recommendation X.25, a packet switching standard using “virtual circuits”.
Buckman's development was significantly different from Systems Network Architecture: while IBM built terminal-to-computer architecture, Buckman connected peer-to-peer computers. This approach has made the project very attractive for companies like General Motors, a leading OSI supporter in the eighties. GM had dozens of factories and hundreds of suppliers using a mixture of low-compatible software and hardware systems. The Buckman scheme would allow computers and networks of different proprietary types to be integrated - if they followed OSI standards.
The layered reference OSI model also provided an important organizational opportunity: modularity. That is, the division into levels made it possible to divide the work on protocols between the committees. Obviously, the Buckman model was just the beginning. To become an international standard, each proposal had to go through four stages, starting with a working draft, then a draft of the proposed international standard, then a draft of the international standard and, finally, an international standard. Creating a consensus around the OSI reference model and related standards required an extraordinary number of plenary meetings and committee meetings.
The first plenary session of OSI lasted three days, from February 28 to March 2, 1978. Many delegates from ten countries and observers from four international organizations gathered. All participants had their own market interests and ready-made developments. At the same time, delegates from one country could have completely different goals. Many of the attendees were INWG veterans who remained cautiously optimistic about the possibility of snatching the future of networks from IBM and the telecommunications monopolies, which, quite obviously, planned to dominate the new market.
1977 : ISO Committee on Open Systems Interconnection The Open Systems Interconnection is formed headed by Charles Buckman [left]; among other active participants are Hubert Zimmerman [center] and John Day (John Day) [right].
1980 : The US Department of Defense publishes the Internet Standards for Transmission Control Protocol and Protocol (“Standards for the Internet Protocol and Transmission Control Protocol”).
At the same time, IBM representatives, led by Joseph De Blasi , a very capable director of standards at the company, skillfully guided the discussion, keeping the development of OSI within IBM's business interests. Informatics John Day, who developed the protocols for ARPANET, was a key member of the US delegation. In his 2008 book, Patterns in Network Architecture, ed. Prentice Hall , Day recalls that IBM representatives skillfully intervened in disputes between delegates who fought for a piece of cake ... IBM turned them about as they wanted. It was a fascinating sight. ”
Despite the sticks being inserted into the wheels, Buckman's leadership drove OSI along a perilous path from concept to reality. Buckman and Hubert Zimmerman (veteran Cyclades and INWG) made an alliance with telecommunications engineers from CCITT. But this partnership with difficulty overcame the fundamental incompatibility of their views. Zimmerman and his computer science colleagues, inspired by Puzan’s datagram design, fought for connectionless protocols, while telecommunications professionals insisted on virtual circuits. Instead of resolving the dispute, they agreed to include both options within OSI, thereby increasing its size and complexity.
This complex alliance of informatics and communications specialists published an international standard for the OSI reference model in 1984. Separate OSI standards for transport protocols, email, electronic directories, network management, and many other functions soon followed. OSI was beginning to show signs of its inevitability. Leading computer companies, such as Digital Equipment Corp., Honeywell, and IBM, were by then very interested in OSI, along with the European Economic Community, governments in Europe, North America and Asia.
Even the US government - the leading sponsor of Internet protocols that are incompatible with OSI - has joined the popular project. The Department of Defense formally adopted a conclusion from the recommendations of the National Research Council (1985) of 1985 on the transition from TCP / IP to OSI. At the same time, the Department of Commerce issued a mandate in 1988 prescribing the use of OSI standards on all computers procured by the US government after August 1990.
Although such decrees look like the work of distorting bureaucrats, it should be recalled that in the eighties the Internet was just a research network: the network grew rapidly, but its administrators did not allow commercial traffic or third-party providers to the government-funded highway until 1992. For companies and other large organizations that wanted to exchange data between different types of computers or networks, OSI remained the only option.
January 1983 : The US Department of Defense (US Department of Defense) requirement to use TCP / IP in ARPANET marks “The birth of the Internet.”
May 1983 : ISO publishes the international standard "ISO 7498: Basic Reference Model for Open Systems Interconnection".
1985 : The US National Research Council recommends that the Defense Ministry gradually migrate from TCP / IP to OSI.
1988 : Revenue in the US computer communications market: $ 4.9 billion.
This, of course, does not end the story . By the end of the eighties, resentment at the slow development of OSI had reached a critical point. During an assembly in Europe in 1989, OSI advocate Brian Carpenter gave a speech entitled “Is OSI Too Late Late?”. That time, as he recalls in his recent memoirs , "was the only time in my life" when he was honored with a "standing ovation at a technical conference." Two years later, a French network expert and former member of the INWG, Puzan , in an essay entitled “Ten Years of OSI — Maturity or Infancy?” (“Ten Years of OSI — Maturity or Infancy?”) Outlined growing uncertainty: OSI is constantly recommended as a major solution. However, it is much easier and faster to implement a homogeneous network based on proprietary architectures or to combine heterogeneous systems using TCP-based products. ” Even for OSI supporters, the Internet looked increasingly attractive.
The feeling of doom was exacerbated, progress slowed down and in the mid-nineties, OSI's beautiful dream finally ended. The fatal flaw of the process, no matter how ironic it was, was a commitment to openness. The formal rules of international standardization gave any interested party the right to participate in the development, which led to general tension, incompatibility of views and gave way to subversive activities.
OSI’s first chairman, Buckman, had been expecting these problems from the start. In a speech at a conference in 1978, he expressed concern about OSI’s chances of success: "The [organizational challenge] is great. Technical problems are more than anything previously known in information systems. And political problems will make the most clever statesman sweat Can you imagine an attempt to bring together in the foreseeable future representatives of a dozen leading competing computer corporations, and a dozen telephone companies, and PTT [public telecommunications monitors Olii - prim.avtora], and technical experts from ten different states? "
1988 : US Department of Commerce orders government organizations to purchase only OSI-compliant products.
1989 : While OSI begins to fall apart, computer scientist Brain Kapenter gives a speech “Has OSI already been late?” With a closing ovation.
1991 : Tim Benes-Lee announces public release of the WorldWideWeb application.
1992 : The US National Science Foundation revises its policy on allowing commercial traffic on the Internet.
Despite the efforts of Buckman and other people, the burden of organizational problems did not subside. Hundreds of engineers attended the meetings of various committees and OSI working groups, and the bureaucratic procedures used to structure the discussion prevented the creation of standards quickly. They argued about everything: even about the trivial nuances of the language (for example, about the difference between the turns of “you obey” and “you should obey”). Much more complicated questions separated experts from computer science and telecommunications, whose technical and business plans were in conflict. Thus, openness and modularity - the key principles of project coordination - have led to the collapse of OSI.
At the same time, the Internet flourished . With sufficient funding from the US government, Cerf and Kahn and their colleagues were protected from the forces of international politics and economics. ARPA and the Defense Communications Agency (Defense Communications Agency) accelerated the introduction of the Internet in the early eighties, funding scientists to implement Internet protocols in popular operating systems, such as the Unix variant from the University of California, Berkeley. Then, on January 1, 1983, ARPA stopped supporting the ARPANET inter-node protocol, thereby forcing contractors to use TCP / IP if they wanted to stay connected; This day is known as the " birthday of the Internet ."
Photo: John Day
What is in the name? : At a meeting in July 1986 in Newport, RI, representatives from France, Germany, the United Kingdom and the United States considered the issue of the OSI reference model with critical naming and addressing functions on the network.
Thus, while many users were still waiting for OSI to become a solution for global networking, an increasing number of people began to use TCP / IP as a practical, albeit temporary, means of networking.
Engineers who joined the Internet community in the 1980s often did not understand OSI, mocking its erroneous monstrosity created by incomprehensible European bureaucrats. Engineer Marshall Rose wrote in his 1990 textbook that “The Internet community is trying hard to ignore the OSI community. OSI technologies are generally ugly compared to Internet technologies. "
Unfortunately, the negative attitude of the Internet community has also led to the rejection of good technical ideas OSI. A classic example is the "palace coup" in 1992. , OSI, (Internet Activities Board) (Internet Engineering Task Force, IETF), . 1992 . , , TCP IP, , , OSI. - , .
1992 : « , - ISO ConnectionLess Network Protocol, IP 4.
1996 : - IP 6.
2013 : IPv6 1 .
TCP/IP , : . , OSI, , , ISO , . (Mark Levilion), IBM, 2012 OSI TCP/IP : „ , - , . , , , , . , ?“
, - . OSI, „“ . , „ “ , .
, OSI „ “ . , , , OSI: ( ) , , .
„“ „“, OSI , , . , „“. OSI - . OSI , . , , ?
Want to know more?
“'Rough Consensus and Running Code' and the Internet-OSI Standards War” 2006 , IEEE Annals of the History of Computing. Open Standards and the Digital Age: History, Ideology, and Networks (Cambridge University Press, 2014).
(Janet Abbate) Inventing the Internet (MIT Press, 1999) , , .
“ INWG and the Conception of the Internet: An Eyewitness Account ” , IEEE Annals of the History of Computing 2011 , International Networking Working Group, ().
- Entrepreneurial Capitalism and Innovation: A History of Computer Communications, 1968–1988 (James Pelkey) , 1980- , , OSI . Ethernet .
Source: https://habr.com/ru/post/376709/
All Articles