"Options For making Standards"

Options for Making Standards

By: Scott Bradner

A problem some people see with traditional standards organizations, which even holds true with the IETF, is that they do not have control over who is part of team. This is seen as a problem because a too-broad a membership can mean that standards take longer to develop with more cooks in the kitchen. It can also mean that the standards will not be focused on a specific problem but instead will be forced to try and attack a wider range of issues.

One reaction to this is that some companies decide to forgo all appearances of an open standards process by getting other companies to agree to use its technology. Some companies also follow this path because they think they have valuable technology that no one else has protected by patents so they think they will be able to make some money licensing the technology.

Another tactic that is gaining in popularity is the use of forums. This is not a new idea. The Frame Relay Forum (http://www.frforum.com/) has been around since 1991 and claims to be the first of its kind. Their charter says "The Forum's technical committee takes existing standards, which are necessary but not sufficient for full interoperability, and creates Implementation Agreements (IAs). These IAs represent an agreement by all members of the Frame Relay community as to the specific manner in which standards will be applied, thus helping to ensure interoperability. At the same time, the Forum's marketing committees are chartered with worldwide market development through education as to the benefits of Frame Relay technology."

Forums can be designed, like the Frame Relay Forum, to come to agreements on how to make use of existing standards or, like the ATM Forum (http://www.atmforum.com/), they can be designed to create standards (often called "recommendations" or "specifications" to avoid the use of the term "standard" due to perceived legal risks.)

A standards developing forum can be created because no other group is working on the specific problems, because a group of vendors think that the work of all existing standards organizations has problems that can not be solved within those organizations, or because a group of vendors wants to be in control of their own destiny or attack a narrowly defined problem.

New forums seem to be springing up almost every week these days. It is hard to be sure from their press releases what they are all about but quite a few of them seem to be in the standards business and want to preempt the work of other standards organizations without quite saying so. It is also quite hard to tell what the actual level of support is for some of these groups. Many of them have impressive lists of "members" with the major companies like Cisco showing up as members on multiple forums with sometimes competing areas of interest. In some cases it turns out that some low-level marketing person at such a company agreed to "join" but that may not mean that the company technical people agree that the effort is technically valid or useful.

Two of the new forums that have received quite a bit of press coverage are the QoS Forum and the Multiservice Switching Forum.

The QoS Forum has not quite formed yet and does not seem to have a web page but got some publicity in conjunction with the iBAND conference in November (http://www.stardust.com/iband/) where an organizing meeting was held. According to the organizer, the QoS Forum will not be a standards setting organization. He wrote me to say that he wanted the forum to focus on:

1. The Need - Business Solutions & Applications enabled by QoS technologies

2. The Hurdles - What do we need to do to increase awareness & accelerate deployment?

3. The Technology - Which layers/standards/protocols etc should this forum embrace?

This forum could be a very useful complement to the standards development effort now underway in the IETF since the first two issues are outside of the scope for the IETF and any standards body will assume the answer to the third point is their own technology when that may not be the ideal answer.

The Multiservice Switching Forum (http://www.msforum.org/) was formed last November and already has created two large technical proposals. The forum's mission statement is:

"The mission of the Multiservice Switching Forum (MSF) is to define:

(1) an architecture that separates the control and user/data plane aspects of an ATM-capable Multiservice Switching System and establishes a framework which can be easily extended to support new user/data plane and control functions

(2) a set of open intra-switch interfaces and promote implementation agreements for these interfaces that allow service providers to deploy ATM-capable Multiservice Switching Systems composed of best-of-breed components from multiple vendors."

The forum has already produced an impressive architecture document and virtual switch interface specification which are on their web site. Strangely enough the pubic architecture document is labeled "MSF Conidential".

The targets of these two forums are quite different - the QoS Forum is addressing the Internet and the Multiservice Switching Forum is addressing interfaces within an ATM switch. The QoS forum is following the wave of assumption that IP and the Internet will soon transport the world's telecommunications traffic and the MSF seems to be assuming that at least some of this wave is overstated.

I do not think that ATM switches and ATM-based services are going to be irrelevant in the future but I'm not sure just how they will fit into the picture in five years. There are perfectly logical scenarios in which the provision of a wide range of services will be done by networks of ATM switches but with the advent of IP-based QoS there are other perfectly logical scenarios in which the services are provisioned over IP networks and whatever ATM switches there are in the network will be dedicated to transporting IP and only differentiating between services based on the markings in IP packet headers.

It will be quite interesting to see which of these Forums, if any, will be playing a major role in the data communications world of 2005.