This is a continuation and conclusion to a post from last week on the broadband research agenda proposed by the Obama Administration. We left off with the example of “multihoming” as a locus of multidisciplinary research.

Do We Need a National Broadband Research Agenda?

The question of multihoming research may shed some light in the broader question of the value of a National Broadband Research Agenda (NBRA.)

In terms of technical research on the Internet (rather than “broadband”, a much more vague term), multihoming and mobility are best seen as features of the Next Generation Internet rather than as a standalone research topic.

Given that we’ve already had more than 40 years of technical research on multihoming, it’s hard to make a case for more funding. If we do any funding for multihoming research at all, perhaps it makes sense to examine the reasons why it hasn’t already been adopted.

But the NBRA is meant to be a multi-disciplinary program that examines technical issues in their social and economic context, so it has different priorities. The workshop suggests that multihoming might have effects on competition, usability, and broadband pricing, topics that are outside the bounds of purely technical research; these are listed above as RQ3-5.

Effects on competition, underserved communities, and pricing are economic issues that don’t require close collaboration between economists and technologists. But as they don’t become relevant unless multihoming is a near-term reality, an argument can be made for multidisciplinary research with loose coupling.

While economists don’t require collaboration to perform analysis of specific technical features, there could well be some benefit in leading economists (and sociologists, in some cases) to pay attention to possible side effects of emerging technologies before they become evident.

Predicting Market Take-up of Academic Research

But coordinating such research depends on a raft of assumptions about the relationship of government-funded technology research to the marketplace realities of interest to economists and other social scientists. If NTIA (or the NSF, or any other government program) supports multihoming research, such research may or may not bear fruit in the academic setting.

Assuming it does create excitement among academics, it may or may not become a marketplace reality. In practical terms, the wide-scale use of multihoming would require a new software layer between TCP and IP throughout the Internet. TCP would be changed to use a new form of addressing, a host address, while IP would continue to use traditional addresses.

This new “multihoming layer” would perform the calculations to split traffic among IP paths in some conditions, and to switch from one LTE carrier to another in other ones. This has been done in experimental form by Multipath TCP (MTPCP) but more work remains to be done for multihoming to become a fully general capability. UDP, for example, will also need a multipath form.

TCP virtual circuit creation would need to be modified so that both endpoints would be aware of all IP paths, and they would need to be synchronized when end user systems switch from one LTE carrier to another.

Because the timing of these events is unpredictable, it’s difficult to know if and when economic research should begin. Consequently, it’s hard to say that economic research should be closely coupled with technical research. Certainly economists should be broadly aware of emerging technologies, but only when the emergence is taking place in markets.

Preventing Adoption of New Technologies

A more intriguing implementation of a unified research agenda would reverse the timing of economic analysis and technical research. Instead of freely supporting a wide-ranging set of technologies, what would happen if we used economic and methods to prune the list of technologies in advance to those that appear to have social benefits?

Would it be sensible to curtail federal support for research on multihoming until economists deliver the verdict that it’s likely to increase innovation, competition, and social values such as digital inclusion? Is such an analysis likely to be worthwhile before the technology becomes a marketplace reality? Do we want economic research to have veto power over technical research?

Fund Research on its Merits

I raise this issue not because I support it, but because it seems to be implied by the notion of interdisciplinary coordination of the research budget. At a minimum, directing limited federal dollars toward technical topics with supposed high social value has the effect of the directing them away from topic with less social favor. As the political is social and vice versa, it effectively politicizes technical research. This leads technology research in a dubious direction.

In other words, a social policy researcher may very well to inclined to reject a proposal to research a method of making networks better, faster, and cheaper on the grounds that such a technology would aggravate the digital divide between urban and rural America.

I would maintain that the technical research should go forward without prejudice to the research budget for means of mitigating the divide. The digital divide deserves attention regardless of technical means for better networking, and probably always will.

So I would suggest that researchers should have access to research topics in their own and other fields, but research should not generally be coordinated. Hence, Internet and broadband research programs in technology, economics, and the social sciences should be evaluated and funded on their own distinct merits.