How hard is it to obtain interface information for commercial wireless systems?

Abstract:

Experimenters are now striving to develop, apply, and refine mesh networking for wireless data. Cheap, common unlicensed 802.11 "Wi-Fi" equipment forms important testbeds used by small innovators to create the mesh. Unfortunately, the user-driven development of dynamic meshing is currently slowed or foreclosed. While the sum of all deployed Wi-Fi devices has no single owner and was not centrally deployed, it is nonetheless a network and access to it is now constrained by secrecy among manufacturers in the concentrated network card chipset industry. The authors are academics and engineers currently involved in developing mesh networks. We contacted all major manufacturers of Wi-Fi chipsets in the US (2000-04) and requested interface documentation. We had little success and found unsupportable rationales for secrecy. We contend that constellations of private part 15 equipment should be considered as an "ownerless" whole network where interfaces should be compelled using a procedure similar to Sec. 68.110. More broadly, as radios like these become increasingly defined in software, this presents a regulatory crisis: as the basis for fixing spectrum allocation rules was formerly hardware, the increasing configurability of radios may seem to create new rationales for interface secrecy. We find few benefits to interface secrecy, and argue that the benefits of user-driven innovation (like this example of dynamic mesh networking) outweigh them. Finally, the empirical ground of Wi-Fi allows us to reassess the appropriate role of regulation and its past distinctions between manufacturer and user, hardware and software, wired and wireless.