|
Author note: Thanks is due to Vint Cerf, who provided additional detail and comment, although not necessarily joining all the views presented. More than any other person, he deserves credit for the IETF’s formation and early success. Bob Kahn should also receive praise for the success of the IETF.
The Internet Engineering Task Force (IETF) is a collaborative body that has developed internetworking specifications for more than five decades, successfully shaping the global marketplace of digital network equipment and services. Beginning as a kind of distributed think tank among network researchers in 1969, it evolved to become one of the world’s most influential standards bodies.
This article describes, through an analysis of the IETF’s own meeting records, a history of the evolution of the participants in the organization that reflects adaptation to the ever-changing technologies, marketplace, and diversity of participants. It attempts to explore key questions like—who were the IETF constituents—the parties who expended the considerable money and resources to participate in the meetings and reflected in the participant registration pattern in the graph below.
These participatory patterns reflect the perceived usefulness of the IETF to its constituents in a continually changing technology and marketplace ecosystem where the costs of supporting someone to participate meaningfully are both significant and increasingly competitive vis-a-vis other standards bodies. As the IETF continues to evolve to accommodate change, understanding the past constituent adaptive history is helpful in providing increased diversity and effective inclusivity of new sectors and participants as it faces these new developments.
The intent here is not to provide an authoritative history of the IETF—which is unusually complex and highly dependent on observer perspective. Links to the source IETF proceedings and RFCs are provided throughout to allow readers to access original material. The objective is rather to encourage further analysis and understanding of the IETF’s constituents; how those constituents have evolved; and encourage a dialogue on how its organizational attributes and processes might evolve going forward to best serve those constituents and the global marketplace.
What is now known as the IETF began its existence as an informal group among ARPA network researchers with a document known as Request for Comments (RFC 1) distributed on 7 April 1969. RFCs evolved over the past 50 years to become the IETF’s principal product consisting of specifications for a broad variety of network protocols, applications for what it describes as a “the Internet [as a] pervasive global construct.” The loose ensemble led by Steve Crocker called themselves the “Network Working Group” (NWG) that included researchers at two private companies and two universities plus the head of ARPA’s Information Processing Techniques Office (IPTO) Program Manager.
The NWG meetings continued to grow. By March 1971, they consisted of DARPA researchers at seven private companies and seven university researcher ensembles. As time went on, subgroups emerged for treating data management, network graphics, mail, and various new protocols. However, for the next fifteen years, the NWG continued to exist as a relatively small and confined group for DARPA researchers to exchange ideas and announce specifications for DARPA’s own closed research network that were published as RFCs.
Notable contractors were BBN (Bolt, Beranek & Newman), which operated the ARPANET network), SRI (Stanford Research Institute), which operated the ARPANET Network Information Center, and ISI (Information Science Institute), which provided protocol specification support. Pursuant to a 1982 directive from the U.S. Undersecretary of Defense, the Defense Communications Agency (DCA) was responsible for management, policy, and service of the network that was detailed Internet Protocol Transition Workbook and supplemented with the Internet Protocol Implementation Guide.
Over the period, several important new bodies emerged. The Internet Configuration Control Board (ICCB) was established around 1980 by the ARPANET Program Manager Vint Cerf which became responsible for ARPA Internet Protocol policy. The ICCB was subsequently replaced by the Internet Advisory Board (IAB), which spawned 10 task forces, one of which was the IETF in 1986. The Internet Advisory Board then became the Internet Activities Board which subsequently became the Internet Architecture Board. The histories of the IAB and the IETF are closely entwined and complex.
In 1986, the NWG began to acquire definitive structure and significantly enhanced support. The change was initiated with the proposed creation of an Internet Engineering Task Force at the 4th meeting of the DARPA Gateway Algorithms and Data Structures Task Force in San Diego in January 1986 (referred to as IETF#1). The purpose of the IETF under its charter was “to identify and resolve engineering issues in the near-term planning and operation of the DoD Internet” with the agenda set by “the operational agencies and their contractors.” The new organization’s meetings, published proceedings, and participants were to be supported by U.S. government agencies and their contractors. The meeting was attended by 21 people from 13 different U.S. government agencies and contractors from private companies and academia. The top participant organisations at the IETF#1 meeting reflected its mission:
Organizations (Followed by No. of Participants) |
---|
BBN (Bolt, Beranek & Newman – which operated the ARPANET network) (4) |
SRI (Stanford Research Institute – which operated the ARPANET NIC) (3) |
MIT (3) |
DCA (Defense Communications Agency – now DISA which was responsible for policy) (2) |
The newly formed DARPA Internet Engineering Task Force met in April 1986 (referred to as the IETF#2 meeting) at the U.S. Ballistics Research Laboratory (BRL). It consisted of largely the same 21 people from 15 different organizations—DOD contractors from private companies and academia, plus government agencies. The top participant organizations were:
Organizations (Followed by No. of Participants) |
---|
BBN (Bolt, Beranek & Newman – which operated the ARPANET network) (4) |
SRI (Stanford Research Institute – which operated the ARPANET NIC) (3) |
MIT (2) |
BRL (Ballistics Research Laboratory – the meeting host) (2) |
The IETF immediately found itself in a whirlwind of multiple major developments and changes. Ultimately, the most important of these during the 1986-1994 period was the allocation of $ 2.5 billion by the U.S. Congress through the National Science Foundation (NSF) to a vast array of companies and academic institutions for networking research, specification development, and demonstration networks and application demonstrations that included and were facilitated by the IETF. NSF, NASA, and DoE committed to the TCP/IP protocol and applications.
Additionally, the DOD, together with most commercial network companies and Europe, had committed to Open Systems Interconnection (OSI) internet protocols and standards which NIST initially published as the U.S. Government OSI Profile (GOSIP) and submitted jointly to the ISO and CCITT (now ITU-T). At the time, multiple major network communities and companies had also developed and championed their own internet protocols and applications. Even the banking community had developed its own internet and trademarked the term. The pursuit of internetworking platforms was elevated to strategically important levels by companies and national governments alike. All of these developments and the related parties and technical challenges began appearing at IETF meetings.
During 1986, The Internet Activities Board also held the first TCP/IP Vendors Workshop in Monterey, California, in cooperation with DARPA. The event later became the Interop trade show. Similar to the ITU Telecom and the GSMA Mobile World Congress, the Interop provided a close binding among respective industry communities and their standards making activities.
What was especially notable about the IETF#2 meeting was the report on gateways to the DOD internet and papers on the use of it as a common bridge among the many other existing internets at the time. The ability to accommodate technical and institutional diversity would ultimately prove a key to the subsequent success of the IETF and market acceptance of its work.
By the time of the IETF#6 meeting in July 1987, efforts to expand the role of IETF beyond its original DOD internet mission were seen in the expanding commercial company participants, additional government agencies, as well as the engagement with other internet standards bodies. Especially significant were the presence of U.S. ISO standards body X3S3.3 participants who were advancing the OSI internet protocol known as CLNP, security and management capabilities, and applications in international standards bodies. The meeting was dedicated to presentations of ongoing work by multiple other bodies about their work. Also notable was a 1987 workshop of the NSF Federal Coordinating Council, which controlled project funding. Its seven subgroups several months earlier treated such key subjects as Internet Concepts, Networking Requirements and Future Alternatives, Future Standards and Services Requirements, Security Issues, and the Government Role in Networking. The attendees at IETF#6 consisted of 80 people from 46 different organizations. The top participant organizations reflected the rapidly changing IETF constituent diversity:
Organizations (Followed by No. of Participants) | ||
---|---|---|
BBN (8) | DEC (4) | MITRE (7) |
NBS,nowNIST (4) | SRI (6) | Tandem (6) |
Unisys (3) |
Among the many other organizations present were significant internetworking platform stakeholders, including Bellcore, ACC, Cisco, H-P, IBM, and NASA. Bellcore was both the principal U.S. telecom provider research organization at the time, as well as the secretariat for telecom industry’s standards—domestic and international. ACC, Cisco, H-P, and IBM had entered the growing internet network equipment marketplace. NASA became a major user and developer of internetworking capabilities.
After the IETF#6 meeting in July 1987, the internetworking market sector continued to grow rapidly. Especially significant was the buildout of the U.S. National Science Foundation Network (NSFNET) TCP/IP internet backbone across the country with major hubs on the East and West coasts that were used for international connections. Most of the parties involved used the IETF meetings to collaborate on technology, operations, new applications, and ways to make all the current and emerging network protocols interoperate. The backbone, international connections, research, and IETF activity were all funded by the $2.5 billion NSF Congressional allocation approved the year before.
Thus, a year later, at the IETF#10 meeting in Annapolis in June 1988, all the government, academic, and commercial parties engaged in operating, developing and using NSFNET and ARPANET, including new protocols and products, were active participants. Dedicated working groups created at earlier meetings expanded and began tackling key needs related to network management, authentication, security, DNS domains, ISO internet protocol integration, host requirements, and routing. The growth of working groups was noted at IETF#9 and led to established IETF internal structures and processes. The meeting was also marked by the IETF Secretariat responsibilities being assumed the Corporation for National Research Initiatives (NRI/CNRI) that was created by the former DARPA networking chief Bob Kahn and funded by NSF. The meeting was attended by 112 participants from 58 different organizations. The attendees at IETF#9 in March 1988 included a participant from Canada and Germany (the DFVLR)—the first recorded non-North American participant. MITRE, as a major network security research and development arm of the U.S. government, had a significant presence and multiple roles, including the IETF at this time. The top 25 participant organizations at IETF#10 were:
Organizations (Followed by No. of Participants) | ||
---|---|---|
BBN (5) | Carnegie Mellon U (2) | Cisco (3) |
DCA (4) | DCEC (DCA Eng) (4) | Dept of Energy (3) |
Dept of HHS (2) | IBM (3) | Lawrence Liv NL (3) |
M/A-COM (2) | Merit/NSFNET (2) | MIT (3) |
MITRE (14) | NASA/NAS (2) | NCAR (2) |
NSF (3) | NYNEX (2) | Pittsburg S C (2) |
Proteon (3) | SRI (3) | UC Berkeley (2) |
Unisys (2) | Univ of MD (2) | US Naval Acad (4) |
Wellfleet Comm (2) |
The IETF#13 meeting at Cocoa Beach in April 1989 was the occasion for announcing the decommissioning of the ARPANET. It was also notable for the introduction of the far-reaching NSA SDNS (Secure Data Network System) initiative for network security which had been announced publicly in August 1986 and been undertaken by a number of agencies and private sector contractors who participated at IETF. SDNS was subsequently treated at the 17th meeting in 1990, the 23rd and 24th meetings in 1992. It was intended to be a core part of the public internet infrastructure but ultimately implemented only as components.
The year 1990 placed the IETF on the world standards-making stage. In February 1990, the world’s principal network standards bodies convened an international summit at Fredericksburg, Virginia, known as ITSC#1. The purpose was to consider a rapidly changing technology and market ecosystem and the inadequacies of the major industry standards bodies. The ITU Secretary-General’s representative chaired the policy panel, and Vint Cerf—as Chair of the Internet Activities Board—was asked to describe the IETF and its processes. The timing coincided with the IAB developing a new standards process at the IETF#16 meeting, including the relationships with the U.S. Government OSI Profile (GOSIP) standards. The IETF was portrayed as a model of participant openness, process transparency, and document availability. Several months later, prominent IETF participants implemented a project known as “Bruno” to make all ITU technical standards publicly accessible via FTP on the TCP/IP internet.
The following year in September 1991, the IETF again at ITSG#2 in Sophia Antipolis was even more prominent as a model for other bodies—especially in light of the Bruno project. In subsequent years, the ITSG Conference grew to become the Global Standards Collaboration (GSC) organisation. Its annual meetings focused on improving working relationships among the many constituent standards bodies, as well as common problems relating to transparency, diversity, IPR, and anticompetitive behaviour. Although it attended some of its early meetings, the IETF became more insular and discontinued engagement.
In many ways, the year 1992 was a point of inflection for the IETF in much the same way as 1986. A new U.S. Administration had been elected that championed the NSFNET together with the TCP/IP internet protocols and applications developed in U.S.-funded academic centers. The providers of commercial public data services were also offering NSFNET gateways. Perhaps most notably, the efforts by DOD, public telecom providers, and Europe to employ the IETF to implement OSI internet protocols and applications began to fail. Much of the turmoil played out at rancorous meetings of the IETF and its management activities. A new foundational IETF document - “The Internet Standards Process” - was adopted. It articulated what constituted Internet Standards - together with the IETF structure, bodies, powers and procedures.
The IETF#23 meeting held in early 1992 in San Diego was the occasion of announcement of fundamental change in the IETF described at length at the beginning of the meeting proceedings. It describes the origins as a forum for DARPA and then NSFNET contractor coordination, but noting that it “has grown into a large open international community of network designers, operators, vendors, and researchers concerned with the evolution of the Internet protocol architecture and the smooth operation of the Internet.” A new mission statement was articulated with extensive structure and a designated secretariat with support staff. Special mention was given to the increase of attendees from 360 to 529, with 46 working groups and internationalization that included 43 non-U.S. attendees from 14 other countries. The rapidly increased involvement of major providers of computer and telecommunication systems and services was also evident.
Organizations (Followed by No. of Participants) | ||
---|---|---|
3Com (11) | ANS (11) | Apple (10) |
AT&T Bell Labs (4) | Ballistics Res Lab (3) | BBN (14) |
Bellcore (15) | Cabletron (5) | Carnegie Mellon Univ (5) |
CERFnet (4) | Cisco (5) | CNRI (8) |
Cornell Univ (5) | DEC (21) | DISA (4) |
EDUCOM (3) | Epilogue Tech (3) | FTP Software (5) |
Hewlett-Packard (7) | IBM (9) | Intel (4) |
ISI (7) | Korea IST (3) | Lawrence Livermore (4) |
MCNC (5) | Merit Network (11) | Microsoft (4) |
MITRE (8) | MITRE (9) | NASA (9) |
NCSC (4) | Novell (6) | Pacer Software (3) |
San Diego S C (11) | Sprint (4) | Sun Microsystems (6) |
SynOptics (6) | Univ of Michigan (5) | Xerox (4) |
Xylogics (4) |
The IETF#25 meeting was intended as a celebratory occasion held in late 1992 in Washington DC and reflected a significant shifting to major private sector participation in the U.S. The event was hosted by Sprint, which provided international connections to the NSFNET backbone from other countries and U.S. government agencies at the time. There were 633 participants from 266 different organizations. The first European IETF meeting was announced. The 652-page proceedings covered nine different areas, including 73 different groups, and introduced SIP (Session Internet Protocol) work. Presentations included both U.S. and nineteen participants attended from 14 other countries. The increasing engagement of commercial vendors is apparent. Noteworthy was participation of the Corporation for Open Systems International that was a major consortium established at the time to work with the U.S. government and industry to bring about global OSI internet standards and undertake performance testing.
Organizations (Followed by No. of Participants) | ||
---|---|---|
3Com (8) | ACC Systems (3) | ANS (11) |
Apple (13) | AT&T Bell Labs (11) | BBN (12) |
Bell Northern (3) | BellCore (23) | Boeing (3) |
Cabletron Systems (5) | Carnegie Mellon Univ (5) | Cisco (9) |
CNRI (9) | Corp for Open Systems Int’l (5) | Datability (3) |
DEC (14) | EDUCOM (5) | FTP Software (5) |
Hewlett-Packard (5) | IBM (22) | ISI (7) |
Lawrence Livermore NL (9) | MCNC (8) | Merit Network (12) |
Microcom (3) | MIT (5) | MITRE (18) |
NASA (12) | NIH (4) | NIST (6) |
Penril DataComm (4) | Sprint (19) | Sun Microsystems (11) |
SURAnet (4) | SynOptics (13) | Univ of California (7) |
Univ of Maryland (5) | Univ of Michigan (7) | Wellfleet Communications (6) |
Xerox (4) |
IETF#25 was especially notable as a fundamental point of inflection in the role and scope of the IETF organization. Although the work at the meeting still included the “OSI Integration Area,” the POISED (Process for Organisation of Internet Standards) report was presented. The report culminated months of turmoil resulting from the efforts to transition to the OSI protocols—a previous principal mission of the IETF established by the U.S. government. The result was a change of leadership, control, and process of their IETF under the aegis of the Internet Society.
The year 1992 was also notable for Cerf and Kahn forming the Internet Society “to facilitate and support the technical evolution of the Internet as a research and education infrastructure, and to stimulate the involvement of the scientific community, industry, government and others in the evolution of the Internet.” The Society’s role subsequently expanded quickly and resulted in a change of leadership, control, and process of the IETF under the aegis of the Internet Society. As Cerf notes, “the IETF resisted being drawn into ISOC and only after considerable debate did it recognize the need to be part of a formally constituted body with protection from potential standards-making risks.”
The POISED changes were ultimately captured in fundamental changes to the IETF’s core document two years later in 1994—The Internet Standards Process—Revision 2. It still included the statement that “the Internet has been evolving towards the support of multiple protocol suites, especially the Open Systems Interconnection (OSI) suite.” Two years later, in 1996, The Internet Standards Process—Revision 3 eliminated the statement entirely and applied the process to all protocols “in the Internet context.” With some updates, it remains the core IETF document today.
The continuously evolving IETF constituency was reflected in the top 40 participants in the years following IETF#25. In 1994, the IETF#31 meeting drew 1079 participants from 441 different organisations in 18 different countries. The top 40 participants still included U.S. government contractors and organizations pursuing OSI protocols. As the OSI protocol integration work came to an end over the next four years, the associated constituency disappeared.
Organizations (Followed by No. of Participants) | ||
---|---|---|
3Com (10) | Ameritech (5) | Apple (19) |
AT&T (13) | Bay Networks (21) | BBN (17) |
Bell Northern Research (6) | Bellcore (22) | Bunyip (5) |
China MPT (9) | Cisco (32) | CNRI (10) |
DEC (10) | FTP Software (8) | Fujitsu (7) |
Hewlett-Packard (41) | Hitachi (5) | IBM (28) |
Intel (5) | ISI (15) | Lawrence Berkeley Lab (7) |
Lawrence Livermore (9) | MCI (18) | Merit Network (13) |
Microsoft (5) | MIT (8) | MITRE (13) |
NASA (20) | Nat’l Ctr for Supercomputing Apps (5) | Naval Research Lab (7) |
Naval Surface Warfare Ctr (6) | Novell (6) | Rutgers University (5) |
Sprint (14) | Stanford University (7) | Sterling Software (6) |
Sun (21) | SunSoft (6) | TGV (5) |
Xerox (7) |
The year 1994 also ended with a document that, more than any other, came to characterize the IETF’s new existence—The Tao of IETF.” The organization is defined as “a loosely self-organized group of people who make technical and other contributions to the engineering and evolution of the Internet and its technologies.” Over the years, the Tao evolved. In 2001, the definition was changed to simply “Internet technologies.” In 2006, it was further evolved to “a loosely self-organized group of people who contribute to the engineering and evolution of Internet technologies,” and since 2012 published and evolved as a web page.
It was also near the end of the 1990-1998 period that the Internet Corporation for Assigned Names and Numbers (ICANN) came into existence through the efforts of the U.S. Dept of Commerce. ICANN assumed the secretariat role for IETF standards identifiers and ultimately funding an array of IETF-related activities through the imposition of charges for those identifiers.
The period between 1998 and 2007 represented peak years for the IETF in terms of participants, industry engagement and development of new marketplace platforms. It coincided with the growth of TCP/IP-based platforms and revenue opportunities—especially for public voice telephony. The most significant intervening market development during the period was the “dot-com and telecom bubbles” which ended the viability of many companies and their IETF participation, as well as enormous growth in the TCP/IP Domain Name System marketplace and new uses that included telephone number resolution.
The year 1999 was also notable for the IETF as an ancillary party to a landmark Intellectual Property agreement signed by the Internet Society and CNRI to end a long-running legal controversy with Internet, Inc. over use of the term “internet.” In July 1999, in a proceeding before the U.S. Patent and Trademark Office Trial and Appeal Board, the parties reached a settlement on all versions of the use of the term “Internet.” They were obligated to recognize that the term was “generic,” and that it represented “a large and growing global information system… which is used by government agencies, communities, commercial organizations, professional associations and individuals for commercial and non-commercial services.”
In 1998 approaching the peak of the industry bubbles, the IETF#43 meeting drew 1551 participants from 546 different organisations in 27 countries. The IETF only needed to hold meetings, and participants came. Commercial equipment vendors and service providers increased significantly—especially those who were prominent in the TCP/IP internet marketplace. U.S. government agency contractors and participants were significantly diminished. The prospective use of TCP/IP to support public telephone services also brought many more telecom providers into the meetings.
Organizations (Followed by No. of Participants) | ||
---|---|---|
3Com (22) | Alcatel (19) | Ascend (12) |
ATT (23) | Bay Networks (41) | BBN (7) |
Bell-Labs (9) | Bellcore (19) | BT (12) |
Cisco (70) | CMU (9) | Ericsson (31) |
Fore.com (8) | France Telecom (22) | Fujitsu (17) |
Hitachi (7) | HP (13) | IBM (53) |
Intel (8) | ISI (15) | Lucent (19) |
MCI (12) | MERIT (13) | Microsoft (24) |
MIT (9) | Motorola (9) | NASA (10) |
Nat’l Univ Singapore (7) | NEC (10) | Netscape (8) |
Nokia (13) | Nortel (16) | Nortel (44) |
Novell (10) | NTT (9) | SBC (8) |
Siemens (28) | Sprint (14) | Sun (35) |
Telia (9) |
In early 2001 as the dot-com and telecom bubbles were occurring, the IETF#50 meeting drew nearly 2100 people from 794 different organisations. That year reflected the zenith of IETF involvement as the communications network industry was at peak turmoil. The growing focus on IP telephony is reflected in the substantial involvement of that industry sector and the network equipment vendors who leveraged IETF standards. The meeting also began to reflect a divergence in view between providers and academicians concerning widespread use of “middle boxes” and their effects on legacy host-to-host architecture assumptions that arose in the 1980s and remains as one of the principal areas of technical contention.
Organizations (Followed by No. of Participants) | 3Com (10) | ACM (8) | Alcatel (47) | ATT (28) | Axiowave (7) | Bell-Labs (8) | BT (15) | Cisco (104) | Erickson (8) | Ericsson (38) | France Telecom (18) | Fujitsu (12) | Genuity (7) | Hitachi (10) | IBM (27) | Intel (26) | ISI (14) | Juniper (16) | Lucent (51) | Micromuse (10) | Microsoft (19) | Motorola (21) | NEC (30) | NeuStar (7) | Nexthop (10) | NIST (8) | Nokia (48) | Nortel (49) | NTT (22) | Siemens (31) | Sprint (10) | Sun (20) | Telcordia (13) | Telia (11) | Tellabs (8) | Toshiba (9) | U.S. Navy (9) | WIDE (8) | World Com (9) | Yahoo (8) |
---|
In 2006—after the 2001 peak - the IETF#65 meeting drew 1258 people from 437 different organisations. The IETF proceedings during this period provided no country-related information, but it was apparent that the non-U.S. attendance was also rapidly increasing and ultimately leveled off at about 70%. The increasing internationalization at the end of this timeframe is also marked by the engagement of Chinese companies and organizations in the IETF work as they became more active in all standards bodies in the sector as a strategic investment similar to North American and European countries and expanded their global market share.
Organizations (Followed by No. of Participants) | ||
---|---|---|
Alcatel (22) | ATT (13) | Avaya (9) |
Boeing (8) | BT (8) | Cisco (150) |
Comcast (6) | Deutsche Telekom (10) | Ericsson (25) |
ETRI (8) | France Telecom (23) | Fujitsu (10) |
FutureWei (8) | Hewlett-Packard (9) | Huawei (20) |
IBM (15) | ISC (6) | Juniper (17) |
Keio University (6) | Lucent (15) | Microsoft (14) |
MITRE (6) | Motorola (13) | NEC (19) |
Network Appliance (7) | NIST (7) | Nokia (33) |
Nortel (27) | NTT (38) | Panasonic (18) |
Qualcomm (9) | Samsung (9) | Siemens (12) |
Sparta (8) | Sun (12) | Toshiba (7) |
USC/ISI (8) | VeriSign (8) | Verizon (9) |
Yokogawa Electric (7) |
The 1998 to 2007 period is also marked by the onset of a kind of self-asserted political activism in the form of an IETF Mission Statement adopted in 2004 that remains as a fundamental continuing constraint on the IETF’s work and evolution.
The IETF mission further states that the Internet isn’t value-neutral, and neither is the IETF. The IETF wants the Internet to be useful for communities that share our commitment to openness and fairness. The IETF embraces technical concepts such as decentralized control, edge-user empowerment, and sharing of resources because those concepts resonate with the core values of the IETF community.
The same Mission Statement also adopted a second foundational constraint—the IETF is only constituted by individual participants. Both of the constraints subsequently became included in The Tao of IETF.
From 2007 onwards, after two decades of turmoil and growth, the IETF participation continued in a kind of stable mode that evolved to accommodate different organizations who leveraged it as a venue to advance their TCP/IP internet products and services in the marketplace. Cisco was joined by Huawei as the dominant attendee organization, followed by a relatively stable, small set of other commercial marketplace participants.
Globally, during this period, the engagement by participants from China increased significantly to constitute the largest country participant in the IETF, as described in the international section below. In 2010, IETF#79 was hosted in Beijing—the only meeting to date occurring in China.
As the public internet infrastructure began to shift significantly toward personalized content delivery and social media services, new and rapidly growing companies like Google became active in the IETF. Both the Internet Society, which was provided revenue by ICANN as well as ICANN itself, began global outreach efforts to make the IETF more globally inclusive. A structured legal aegis in the form of the IETF Trust was finally created in December 2005 to assume ownership of the IPR and the informal institutional roles of the Internet Society and CNRI.
However, as the mobile communications market expanded exponentially during the same period, the preponderance of the industry participants carried that work out in 3GPP—which grew to become the largest and most active of the network communications industry body with its own internet implementations. The 3GPP participant and contribution exponential growth and an enormous number of work items accelerated even further after 2015 when the global planning for 5G emerged. Specialized TCP/IP internet standards tasks were outsourced to the IETF.
In 2011, the IETF#80 meeting drew 1307 people from 446 different organisations in 62 different countries. The top 40 participants are shown below.
Organizations (Followed by No. of Participants) | ||
---|---|---|
Aalto Univ (10) | Alcatel-Lucent (44) | ARIN (4) |
AT&T (9) | BBN Technologies (5) | BT (5) |
BUPT (4) | CATR (15) | CESNET (6) |
China Mobile (9) | Cisco (99) | Comcast (7) |
CZ.NIC (11) | Deutsche Telekom (18) | Ericsson (39) |
ETRI (6) | France Telecom (17) | Fujitsu (5) |
Google (15) | Hitachi (10) | Huawei (49) |
ICANN (6) | Internet Society (18) | ISC (14) |
Juniper (19) | Microsoft (10) | NEC (10) |
NetApp (8) | Nokia (11) | Nokia Siemens (10) |
NTT (19) | Oracle (8) | RIPE NCC (6) |
Siemens (6) | Skype (5) | Telecom Italia (7) |
Tsinghua Univ. (5) | Verisign (5) | Verizon (6) |
ZTE (31) |
In 2014, the IETF#91 meeting drew 1396 people from 441 different organisations in 67 different countries. The period was also one of significant IETF turmoil over material released by Snowden reflected in the IETF mission and work and the adverse effects on network protection and constituent participants. Those pursuing passionate causes found a home in the IETF which did not exist elsewhere. The top 40 participants are shown below.
Organizations (Followed by No. of Participants) | ||
---|---|---|
Alcatel-Lucent (21) | Apple (5) | ARM (5) |
AT&T (8) | Broadcom (8) | Brocade (8) |
BUPT (7) | CableLabs (6) | China Mobile (9) |
China NIC (5) | Cisco (120) | Comcast (5) |
Deutsche Telekom (9) | Ericsson (40) | ETRI (10) |
Google (22) | Huawei (63) | ICANN (10) |
Internet Society (25) | ISC (6) | JPRS (7) |
Juniper (31) | Keio Univ (7) | Meetecho (6) |
Microsoft (22) | Mozilla (11) | NEC (6) |
NIST (9) | Nokia (8) | NSA (5) |
NTT (18) | Oracle (7) | Orange (9) |
RIPE NCC (7) | Tsinghua Univ (8) | TTA (6) |
Verilan (8) | Verisign (15) | Verizon (6) |
ZTE (9) |
In 2017, the IETF#100 meeting drew 1525 people (526 remote) from 526 different organisations in 88 different countries. The top 40 participants are shown below.
Organizations (Followed by No. of Participants) | ||
---|---|---|
Acunetix (5) | AFRINIC (6) | Akamai (13) |
APNIC (8) | Apple (10) | AT&T (5) |
China Mobile (8) | Cisco (79) | Comcast (7) |
Cosmo Software (6) | Dell EMC (5) | Deutsche Telekom (8) |
Ericsson (31) | ETRI (11) | Facebook (9) |
FutureWei (7) | Google (33) | Huawei (88) |
ICANN (15) | IMT Atlantique (6) | INRIA (7) |
Intel (7) | Internet Society (25) | ISC (8) |
JPRS (6) | Juniper (19) | Meetecho (5) |
Microsoft (9) | Mozilla (8) | Netflix Inc (5) |
NICT (7) | NIST (11) | Nokia (37) |
NTT (10) | Oracle (6) | Orange (8) |
Tsinghua Univ (10) | Univ de Chile (7) | Verisign (9) |
ZTE (10) |
In 2019, the IETF#106 meeting drew 1634 people (611 remote) from 561 different organisations in 88 different countries. The top 40 participants are shown below.
Organizations (Followed by No. of Participants) | ||
---|---|---|
Akamai (14) | AMS (7) | APNIC (6) |
Apple (14) | Arrcus (6) | CAICT (10) |
Cisco (74) | Cyberstorm (7) | Dell (6) |
Deutsche Telekom (10) | Dresden Tech Univ (6) | Ericsson (24) |
ETRI (7) | Facebook (12) | Fastly (6) |
FutureWei (17) | Google (43) | Huawei (64) |
ICANN (11) | INRIA (8) | Intel (6) |
Internet Society (21) | ISC (7) | Juniper (38) |
Meetecho (6) | Microsoft (14) | MISA (7) |
Mozilla (7) | NIST (9) | NLnet Labs (6) |
Nokia (36) | NTT (13) | RIPE NCC (9) |
Siemens (7) | Tsinghua Univ (9) | Univ of Accra (20) |
UPSA (9) | Verisign (8) | Verizon (6) |
ZTE (6) |
In 2021, the IETF#110e meeting drew 1329 people (all remote) from 498 different organisations in 68 different countries. The top 40 participants are shown below. The #110 meeting metrics are especially useful to understand the effects of a fully virtual IETF. It appears that neither the top participant nor international metrics changed significantly using virtual rather than physical meetings. As noted below, however, virtual meetings may result in participant organizations having multiple employees attending IETF meetings—likely because of the significantly lower travel costs.
Organizations (Followed by No. of Participants) | ||
---|---|---|
Akamai (14) | AMS (10) | Apple (10) |
AT&T (5) | BT (5) | CAICT (7) |
China Mobile (12) | China Telecom (11) | Ciena (14) |
Cisco (62) | Cloudflare (11) | Comcast (6) |
Deutsche Telekom (11) | Ericsson (30) | ETRI (7) |
Facebook (6) | Fastly (8) | Futurewei (16) |
Google (32) | Huawei (57) | ICANN (6) |
INRIA (5) | Intel (5) | Internet Society (11) |
JHU/APL (8) | JPNIC (6) | Juniper (33) |
Meetecho (7) | Microsoft (15) | Mozilla (6) |
NIST (7) | NLnet Labs (5) | Nokia (21) |
NSA (7) | Orange (7) | RIPE NCC (7) |
Salesforce (9) | Tencent (6) | Verisign (13) |
ZTE (6) |
While the Top-40 participant metrics discussed above are useful in understanding the IETF’s most engaged participant organizations, it is also especially useful to examine the metrics for all the individual actor attendees. After migrating away from the DOD and NSF in the 1990s, the IETF rather uniquely among industry standards bodies, enabled individual actors to exercise significant influence and power.
Although viewed as a desirable attribute by many within the IETF and part of its Tao adopted in 1994, individual actors often have with vague institutional affiliation and unknown funding or motive. It creates significant transparency and antitrust challenges within a standards body that controls what can be deployed as a product or service in the multi-billion dollar internet marketplace. This rather large and constantly churning group is typically diverse: computer scientists and individual consultants who are often funded to be present to gain intelligence or influence outcomes, academic researchers introducing their ideas or job searching, and deserving individuals in remote world locations who are provided stipends to attend a meeting that would otherwise be unattainable. The IETF also attracts what an Oxford Internet Institute research paper described as self-styled technical policy advocates.
The individual actor attendance in the graph includes all those who were either the only person from a participant group or otherwise independent. The numbers were relatively low until the millennium and then increased significantly over nearly the next two decades to reach 50 percent, i.e., half the attendees. As noted below, there is also considerable churn among individual attendees.
Over the past two years, however, the percentage has decreased—especially with recent fully virtual meetings. As noted above, a plausible explanation may be the effects of travel costs. With virtual meetings, two people from the same organization can attend without incurring significant costs and away time.
Because the IETF continues to rely on the actions of individual participants over substantial time periods to introduce and advance standards or attain decision-making positions, it is useful to examine the patterns of individual participant recurrence at meetings over a period of time. For example, an examination of eight meetings over the three-year period between 2017 and 2019 shows that nearly two-thirds of the participants only attended one meeting and that only 5 percent continued over the entire set of meetings.
There appears to be considerable churn in the IETF attendance that suggests fewer organizations today willing to invest the resources in allocating and funding an employee to participate continuously in the IETF. The recurrence metrics also explain the individual participant metrics where one person attends only a single meeting. By contrast, in all other standards bodies, a participant organization can obtain one or even several annual memberships and send multiple different employees as required.
Prior to the changes that were instituted in 1992, the IETF fulfilled its original DOD mission. To the extent people and organizations outside the U.S. participated, it was limited to a few institutions in other countries connected to the DARPA and NSF network backbones. After 1992, the non-U.S. participation increased significantly. The most rapid increase occurred between 1995 and 2010 when the percentage of non-U.S. participants increased to a steady state of around 40 percent, as shown in the above graph for the top 20 countries.
Most of the non-U.S. participation consisted of those from a number of European countries plus Japan. After 2010, Chinese participants became the largest non-U.S. percentage—maintaining a steady rate of around 9%. In general, the small percentages from non-U.S. countries over the past two decades have remained very steady at levels of a few percent, notwithstanding attempts to expand the diversity and inclusivity such as holding meetings in different world regions and subsidies.
The large U.S. percentages follow in considerable measure the IETF’s provenance as well as perceived incentives or disincentives of participation by those in other countries—which is also driven by the Top 40 participating organizations who consistently send large numbers of employees to meetings to further their marketplace interests.
Over the years, the motivations of IETF internationalization evolved. During the 1980s, the academic network researchers—some funded by companies like IBM who created Bitnet—collaborated and sought ways to establish gateways. When the rather massive NSFNET funding was instituted in 1986, substantial money was made available for non-U.S. connections directly through to the U.S. backbone directly or through protocol converter gateways.
Multiple organisations for both operations and standards development existed, and the academic communities met globally at INET Conferences. As the years progressed, the INETs became associated with the Internet Society and included operational workshops, and helped bring non-U.S. participants to IETF meetings.
In the 1990s, as the TCP/IP internet began transitioning from a U.S. government activity to a public commercial offering, and as alternative internet implementations subsided, commercial vendors and equipment providers worldwide—as well as government-sponsored communities in Asia—began appearing in large numbers. Singapore, Japan, Hong Kong, and Korea were among the first, followed by China. IETF’s also became an opportunity for the operational community to gather. Internationalization also became significant as the IETF sought to establish itself as a peer international standards organization and give their standards recognition.
The motivations over the past quarter-century have changed little - which explains the surprising consistency in the country metrics despite considerable efforts by different parties to increase the numbers. There is also a cultural barrier, as the IETF’s emphasis on individual actors and prevalence of aggressive behavior in groups and lists are not tolerated in many cultures.
Most collaborative organizations attempt to examine their constituent participants from time to time and attempt to improve the organization’s practices and mission. Indeed, this recently occurred in the IETF’s Internet Architecture Board initiative Fifty Years of RFCs. In addition, the IETF Director recently undertook a consultation process to solicit views concerning the efficacy of the organization and needed adjustments.
The kind of analysis here is useful for that purpose and should be institutionalized by the IETF. The internetworking architectures, services, and markets are undergoing changes today that equal or exceed any over the past fifty years. In addition, the venues for the work have increased in number and become in some measure competitive. All organizations have their own virtues, but inevitably, participating commercial organizations expend their money where there is perceived benefit, and nation-states engage where there is strategic value.
The IETF participant metrics clearly show a maximum that occurred twenty years ago, followed by a drop a few years later to a steady state of half that number. The principal participating organizations also remain a consistently small number of dominant vendor product providers combined with large numbers of academics and individual actors that make the IETF complex and challenging to manage.
The IETF’s reliance on an “individual” actor paradigm that necessitates a person’s participation over many meetings may also today represent a significant, critical organization liability in today’s ecosystem array of global standards bodies. A commercial company that has devoted enormous resources to the development of IPR in the products and services, may not be willing to risk the investment participating in a standards body: 1) that rests on the behavior and actions of individual participants who are interpreting and acting to implement legacy IETF Mission Statement mandates with which the company does not agree, or 2) may restrain their products in the marketplace, or 3) where important work item decisions occur with limited transparency, or 4) producing a specification can take years to achieve even marginally useful specifications. Even a conversation about these matters is difficult to pursue within the IETF, and although it may be regarded as disparaging, it needs to occur for the IETF’s own good.
To the extent the IETF, as a collaborative standards body competing with other bodies, wishes to adapt to evolve its mission and viability, it seems helpful to examine past the successes. What adaptations brought about an increase in the number and diversity of participants? Or did not. What kinds of cultural behavior and stridency—especially in discussion lists and creating new working groups—enhanced diversity and dialogue? Or not. What changes in mission can bring about a perceived value proposition? Or not. What makes the IETF open and welcome to the broadest array of valued participants? What turns them away? What kind of organization does the IETF want to be in the standards ecosystem? What is viable? What avoids the antitrust “alligators” or standards that are untenable for service providers in the real world where content delivery architectures prevail and for whom “end-user empowerment” is not a sine qua non?
The metrics of organizational affiliations and trends are uniquely important within the IETF for other essential reasons, namely transparency and anticompetitive considerations. Without any explicit membership and a reliance on vague judgemental decision-making and unfettered list-based work item shaping activity, the IETF becomes potentially vulnerable to subtle takeover or manipulation of its activities to further aims and strategies, including unfair competition. The significant number of participants with vague or unknown affiliations exacerbates the challenge. The only way to mitigate these vulnerabilities is to enhance the transparency of affiliations and provide continuing analytical metrics directed at showing proportions and patterns of affiliations. It is due diligence and encouraged by recent antitrust enforcement actions.
Three decades ago, when the IETF was young, welcomed change, and sought recognition by older standards bodies, it engaged in common collaborative activity with those bodies via mechanisms like the ITSC/GSC to address common challenges. Now that the IETF is itself older, it is facing an enhanced need today shared by all bodies—to avoid institutional ossification, better serve the constituents, improve transparency and mitigate anticompetitive behavior. It may be time to again proactively collaborate among the bodies to achieve these common goals.
Sponsored byDNIB.com
Sponsored byVerisign
Sponsored byIPv4.Global
Sponsored byRadix
Sponsored byCSC
Sponsored byWhoisXML API
Sponsored byVerisign
It is correct that the IETF has perhaps uniquely informal decision making and has somewhat informal affiliation declarations. However, participation only by individuals is not so unusual. In particular, all IEEE 802 Working Groups, including 802.1 that specifies bridging and L2 architecture, 802.3 that specifies Ethernet including all speeds of copper, fiber, etc., 802.11 that specifies Wi-Fi, 802.15 that specifies Bluetooth, Zigbee, etc., all of these have individual membership and control, albeit more formally than the IETF through a defined membership and voting. It is also true that with individual membership, a large organization/entity can generally gain influence by sponsoring many individuals to be active and attend. On the other hand, with organizational/entity membership and control a large organization can gain influence by having many smaller organizational members that it influences as customers, suppliers, contracts with, etc. There is probably no perfect system of governance.