A challenge for Researchers by John Day of the Pouzin Society

In the science Computer Science (as opposed to mathematical CS), we face a rather special challenge: We Build What We Measure. We don’t really have Nature to test our theories for the structure of systems. This makes it very difficult to know what is principle and what is an artifact of the engineering decisions.

This has lead to a somewhat “any thing that works is good” attitude and has contributed to the master craftsman approach that seems to dominate recently. History indicates that artisan approaches come to rely on tradition and ultimately stagnate. We are already seeing this.

This is not science. In science, we construct and disprove theories. Those theories not yet disproven are working models. In software systems it is possible to make almost anything “work.” The question then is how do we choose our theories without nature to test against? Is it a popularity contest? Which company has the most money to put behind it? This too is not science. What criteria should we use? After much thought on this problem looking for a neutral basis for criteria, I was drawn to Newton’s Regulae Philosophandi of 1726 (as paraphrased by Gerald Holton):

  1. Nature is essentially simple; therefore, we should not introduce more hypotheses than are sufficient and necessary for the explanation of observed facts. This is a hypothesis, or rule, of simplicity and verae causae.
  2. Hence, as far as possible, similar effects must be assigned to the same cause. This is a principle of uniformity of nature.
  3. Properties common to all those bodies within reach of our experiments are assumed (even if only tentatively) as pertaining to all bodies in general. This is a reformulation of the first two hypotheses and is needed for forming universals.
  4. Propositions in science obtained by wide induction are to be regarded as exactly or approximately true until phenomena or experiments show that they may be corrected or are liable to exceptions. This principle states that propositions induced on the basis of experiment should not be confuted merely by proposing contrary hypotheses.

Or as a corollary, what problems or insights has the model produced that were not considered in its initial formulation. What predictions does it make? (Taking the view of Robert MacArthur, the great analytical biologist, to describe what is without making predictions or explaining issues not yet uncovered is Natural History, not Science.)

For the last decade or more, the research community has been considering the problem of a new theory or architecture for networking. They seem no closer now that they were when they started. To focus the thinking, we issue this challenge:

The IPC Model developed by Patterns in Network Architecture appears to fulfill Newton’s criteria and has made not only predictions, but unexpected predictions about the nature of networking. At present, there appears to be no other theory or architecture that is simpler and has the explanatory breadth of the IPC Model that is not isomorphic to it.

The challenge is to:

  • Either disprove this claim or limit its scope, or
  • Improve the model to broaden its scope
  • Find a simpler or more universal model
  • Further explore the properties of the IPC model


A challenge for Business by Chris Williams of the Pouzin Society

For the better part of 20 years, the Internet (capital I) and its underlying design and specifications has been the de facto “data dial tone” of the planet. The large global ecology of businesses which provide products, services, employment and derive profits from supporting this infrastructure, combined with the social and funding dynamics of the research community, has resulted in the ossification of worldwide network infrastructure at a time so early in its adoption that new applications are becoming increasingly expensive and time consuming, and in some cases impossible, to deploy.

A new approach to understanding and designing networks has emerged, called RINA, as described in Patterns in Network Architecture by John Day. It offers a clear framework within which network architectures and implementations can be analyzed and designed, replacing the present ad-hoc and piecemeal design methodologies currently being practiced. In order to fully realize the vision of this new approach, network architectures must be designed that do not rest upon the current fabric of the Internet Protocol (IP) that is considered by many to be immutable and universal. RINA has serious and positive implications for the design of highly secure and configurable networks that support a robust hierarchy of product and service vendors.

History provides some perspective. The late 19th century saw an infrastructure transition from DC to AC power transmission that involved technical, social and business factors. The well-funded, business savvy and ruthless Thomas Edison was initially successful in convincing interested parties to deploy his DC power transmission technology. Nikola Tesla, hampered by a combination of Edison’s early start and awareness of business issues, and his own social awkwardness and idealism, watched his more versatile AC power transmission designs languish until the technology eventually spoke for itself.

More than 100 years later, we have greater awareness of the coupling between the scientific, engineering, business and social forces that result in the adoption or languishing of technical innovations. It is often difficult to see how both personal and institutional motives influence the direction of both research and product development, but we owe it to ourselves to stand on the shoulders of our predecessors in order to both see how we got to where we are, and to see the options of where we can go.

The challenge to business is to:

  • Suspend disbelief long enough to analyze, implement and test the architecture
  • Ask your colleagues to take a fresh look at their assumptions about networking
  • Deploy the architecture to support emerging needs that stress the old IP network
  • Build a new ecology of vendors developing products that leverage RINA