Section 1.6. The European Invasion


1.6. The European Invasion

The Network Working Group set the tone for the development of most protocols and procedures associated with the Internet and TCP/IP to this day. From its beginning as a group of graduate students and staffers working at the first four ARPANET sites, worried that their junior status and audacity at suggesting protocol solutions would offend the "real" protocol designers, Internet protocol development has been open, driven from the bottom up, with anyone allowed to participate in the process. RFCs were their means of documentation.

"When the RFCs were first produced," Vint Cerf writes, "they had an almost nineteenth-century character to themletters exchanged in public debating the merits of various design choices for protocols in the ARPANET."[27]

[27] Vint Cerf, "RFCsThe Great Conversation," in RFC Editor et al., "30 Years of RFCs," RFC 2555, April 1999.

The NWG grew large over the years, and in 1986 they became the Internet Engineering Task Force (IETF). Over the years, the IETF has grown to become the de facto standards body for IP and related protocols. Yet in spite of this, the IETF has never had an official charter; it still operates as a loose organization of kindred spirits dedicated to creating and improving Internet and networking protocols. And although vendors often play a big role in protocol development these days, they still must participate in the IETF working groups, come to the meetings, and present their work for review and criticism just like everyone else.

But the ad hoc spirit of the IETF did not sit well with everyone. Governments, which are (with a few disastrous exceptions) highly structured bureaucracies, prefer that highly structured bureaucracies set their standards. The IETF was hardly that. But, the International Organization for Standardization (ISO) was.[28] The ISO, headquartered in Geneva, was founded in 1946 by the union of two earlier standardization bodies. Its mission is to enhance international manufacturing and trade through agreement on industrial standards, and it sets standards for industries ranging from construction and agriculture to medicine and electronics. ISO standards define screw threads, credit cards, tool safety, and the pictorial symbols seen in airports, along highways, and on toilet doors worldwide.

[28] You might notice that the acronym ISO does not match the English name. That is because it is not actually an acronym; it is a name. ISO is derived from the Greek isos, meaning "equal." This was chosen so that the acronym would not vary from language to language. The organization is always ISO, in any country.

As experimental packet switching networks began cropping up not only in the United States but in Europe in the 1970s, many governments began to see the need for standardization but wanted the standards to come from a more official body than the IETF. In 1977, the British Standards Institute proposed that the ISO standardize an architecture for communications infrastructures. The ISO set up Subcommittee 16 under Technical Committee 97, and in 1978 the subcommittee proposed the establishment of a reference model to be called the Open Systems Interconnection (OSI) model.[29] The American representative to the ISO, the American National Standards Institute (ANSI), was tasked with developing proposals for the OSI reference model.

[29] ISO/TC97/SC16, "Provisional Model of Open Systems Architecture," Doc. N34, March 1978.

At about this same time, Mike Canepa and Charlie Bachman at Honeywell had been working on an architecture for distributed databases. Drawing from work done by IBM in 1974 on their Systems Network Architecture, which had seven layers, the Honeywell team proposed in 1978 a seven-layer architecture for interconnecting computers called the Honeywell Distributed Systems Architecture (HDSA). When ANSI met in Washington in March 1978 to consider proposals for the OSI model, Canepa and Bachman presented HDSA. It was the only proposal presented, and ANSI adopted their model as the now-famous seven layer OSI reference model.

Work then began to establish protocols compliant to the OSI model. European governments and the European Commission put their weight behind the ISO work rather than the IETF; in the United States, although the military had adopted TCP/IP, the civilian sector under the Department of Commerce was expected to support the ISO protocols. And when the ISO finally produced a preliminary set of protocols in 1988, the U.S. government adopted the OSI protocols even though TCP/IP had been in operation for five years and the OSI protocols had yet to be implemented.[30]

[30] The Government Open Systems Interconnection Profile (GOSIP) was outlined in the Federal Information Processing Standard (FIPS #146), developed by the National Institute of Standards and Technology (NIST).

With massive government backing, it seemed inevitable that the OSI protocols would replace TCP/IP. Throughout the late 1980s and early 1990s, the OSI community viewed TCP/IP as nothing more than an academic experiment. Yet in the end TCP/IP became the Internet standard, and OSI died on the vine.

The prevalence of TCP/IP can be partly attributed to the widespread popularity of the UNIX operating system, versions of which began including TCP/IP as early as 1981. But the biggest reason for the acceptance of TCP/IP has to do with the personalities of the IETF and the ISO. The IETF was interested in practical solutions, whereas the ISO was trying to develop protocols that would fit into a predetermined reference model. As the IETF merrily surged ahead under its credo of "rough consensus and running code," implementing first and then standardizing what worked and abandoning what didn't, the ISO committees plodded along developing standards before getting to implementation. Throughout Europe, supposedly the bastion of OSI, universities became impatient with waiting for the ISO and increasingly adopted TCP/IP.

However, one protocol arose out of the OSI work that remains important to the Internet to this day: IS-IS.




OSPF and IS-IS(c) Choosing an IGP for Large-Scale Networks
OSPF and IS-IS: Choosing an IGP for Large-Scale Networks: Choosing an IGP for Large-Scale Networks
ISBN: 0321168798
EAN: 2147483647
Year: 2006
Pages: 111
Authors: Jeff Doyle

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net