The Future of IP Standardization Activity

By: Scott Bradner

Once upon a time as far as the traditional standards bodies were concerned TCP/IP was not worth taking, even for free. In 1975 TCP/IP was offered to the International Telecommunication Union (ITU) as a technology submission and was rejected. The world of the Internet might well be very different if the offer had been accepted.

The Internet Engineering Task Force (IETF) has been the primary organization defining standards for the Internet since the mid 1980s. The IETF is a bottoms up organization where most of the initiatives to work on specific topics come from the attendees. The work of the IETF is done in working groups. These working groups are organized within areas with each area having one or two area directors. The area directors, along with the chair of the IETF, form the IETF management group known as the Internet Engineering Steering Group (IESG). Almost always working groups are formed when a few individuals get together and decide that there is a technology which should be standardized. They then approach the area director in the relevant area with their plan. If the area director approves the proposal for a working group is presented to the IESG. IESG approval of the proposal creates the working group.

This is different than the process in a number of the more traditional standards bodies where long term work plans are developed by the management of the standards organization and the equivalent of IETF working groups are formed to address the individual work items that make up the work plans.

The difference in the style of standards bodies is often reflected in the resulting technology. The IETF tends to develop standards which follow the stupid-network -- smart-edges philosophy which I talked about in my last column. A network following the IETF standards tends to have few significant resources embedded in the network itself, preferring to assume that most needed services will be provided on the hosts or be distributed widely around the network. The domain name system (DNS) is an example of this. The main work of the DNS is preformed by small DNS servers run by each organization rather than by some very large central directory service. On the other hand the traditional standards bodies tend to develop technologies which assume that significant services will be provided by the network. X.400 email is an example of this philosophy. X.400 assumes that there will be large vendor supplied servers embedded in the network.

I bring up these differences in approach and resulting technology because the world of the standardization of IP technology is changing and these differences will make some of the transition to new standardization processes quite difficult.

Not too long ago the assumption among many government officials, telecommunications "experts" and traditional standards bodies was that the Global Information Infrastructure (GII) would replace the Internet and that a new set of protocol standards would be developed to support it. But the Internet overcame their assumptions and now almost everybody in the data world now accept that the Internet is what was once known as the GII. Some of the telecommunications experts still think that a replacement is pending but the number of people that think that way is shrinking quickly.

The Internet's assumption of the mantle of the GII has raised its profile in the traditional standards organizations consciousness and they are now getting very interested in working on standards for the Internet. At the same time people in the IETF are finding technologies which would one have been within the scope of the traditional standards bodies such as the ITU and the International Standardization Organization (ISO) interesting and now Internet-related topics. Thus there is beginning to be a significant amount of overlap between the IETF and the world's other standards organizations.

This overlap in topics of interest has raised the importance of the standards organizations having good communications channels between each other. ITU and ISO, as well as some of the other traditional standards organizations, have established such communications channels over many years, in some cases, over decades. Since the Internet is a very new, at least in the perception of the traditional standards organizations, force, these communications channels are only now starting to be established. This is not to say that some communication has not been going on for quite a while. The IETF established liaisons with specific ISO study groups a number of years ago but up to now the liaisons have been lightly used. Also, the Internet Society (the umbrella organization for the IETF) exchanged memberships with the ITU a few years ago. But this, until lately, was a little used communications channel. In the last month the communication between the IETF and the ITU has been significantly strengthened, with new, widely publicized communications procedures established.

In an ideal world all of the future standards for the Internet would be developed in a cooperative way among the standards organizations. An example of this kind of cooperation is the arrangement between the IETF and the World Wide Web Consortium (W3C). We have agreed to split up the standardization tasks with the IETF working on transport protocols such as http and the W3C dealing with standards dealing with presentation such as HTML. This sort of split is harder to do with the ITU and ISO since they and the IETF have a number of areas of standardization activity in common.

Where it can be done future standardization work will now be coordinated between the IETF and the ITU. Sometimes it will be decided that the basic work on a particular topic should be done in one of the organizations and sometimes the work should be done using some form of joint effort. Last year's Internet FAX work was an example of the latter. Both the IETF FAX working group and the ITU Study Group worked on the same set of documents which wound up being published by the IETF.

But, due to the basic architectural assumptions noted above, there will be times when the IETF and the ITU have fundamentally different approaches to solving the same problem and work will proceed in parallel in the two organizations. We hope that this case will be the exception rather than the rule but it will happen and when it does the marketplace will decide which architectural view succeeds.