The algorithms underlying DQ technology have been investigated over the years but virtually all the research has addressed their use in wireless networks. A starting point to learning about the efficacy of DQ is to start with the LinkedIn entry of Dean Luis Alonso of Catalonia Polytechnic University in Barcelona, https://www.linkedin.com/in/luis-alonso-73996826/. He describes how as a PhD student his dissertation used as a foundation the seminal paper on DQ published in 1993; he is now Dean of Telecommunications and Aerospace Engineering. Over the years he has directed research into DQ and by now he and his students, and his students’ students, have produced scores of papers, many of which demonstrate the superiority of DQ over existing protocols. In addition the basic DQ algorithm is the basis of the MAC used by IEEE 802.14 - Cable Modems. Unfortunately the cable industry could not wait for the 802.14 to complete their work and instead adopted DOCSIS.
So why is DQ not better known? The main reason is that virtually all wireless activity takes place at the “edge” of the Internet, i.e., WiFi, cell towers, RFID, etc., and most of this traffic must then traverse the “backbone” of the Internet. Major features of DQ, i.e., support of fixed-bandwidth traffic and freedom from congestion, are lost when a packet enters the backbone. So at present nothing would be gained by converting wireless to DQ. An exception would be 5G, DQ would improve the parts of the 5G infrastructure that could be considered as part of the backbone.
I contend that DQ should be first introduced into the non-wireless part of the Internet. But where? The carrying capacity of optical fibre has substantially increased to the point where consideration could be given to using the DQ technology in a manner that would enable an optical-fibre circuit, now used almost solely to carry packets between routers/Ethernet switches, to be treated as a “distributed switch”. This would enable Internet traffic to simply bypass the routers. Not en toto but certainly enough to have a not insubstantial impact on performance and economics.
To some it seems counter-intuitive to use fibre to broadcast data to all recipients rather than sending a specific packet to a specific destination. But the decision should be based on economics and performance, not intuition.
My estimate is that, using the current version of DQ, up to 50% of Internet traffic could go from source to destination router-free; the benefits of 90% utilization of raw bandwidth, instead of 60%, plus the reduced capital and operating costs of the displaced routers would be considerable. This estimate is not arbitrary, it is based on DQ’s ability to support fixed-bandwidth services, intermixed with packet services, thus simplifying the delivery streaming video. Aside from its superiority in delivering on-demand video, the increased capacity of optical-fibre circuits means that at some time a DQ network could be treated as the equivalent of typical server storage “track” of a popular website. A single track could hold the initial pages of hundreds or even thousands of sites. Millions of viewers could then simultaneously “hit” on that site, with no congestion. Subsequent requests for further pages are efficiently processed in the normal manner. Speeds in the hundreds of gigabits/s on fibre are now being developed so one can envision entire videos being delivered in this manner.
The applications described above require only the algorithm and hardware that already exist, or will exist due to normal development. A first product, probably one that would address the backhaul problem for both WiFI and cell operations, could probably be introduced in 24 months generating a considerable amount of Intellectual Property in the process. A DQ distributed switch can support a variety of standards but would initially support the standard Ethernet environment and thus can be described as an enhancement to Internet switching.
DQ technology in a sense combines the MAC/Data Link and PHY layers in providing a superior service for a subset of the Internet. However, DQ research does suggest a path that can be taken in developing a form of synchronous switching that would combine the Data Link/MAC and PHY layers, i.e., make the optical-fibre infrastructure “packet friendly”. This could represent, after Bell’s circuit-switching in the 1870s and packet-switching in the 1970s, a third form of switching, one that would satisfy both asynchronous and synchronous requirements.
An investigation could be carried out in 3-6 months by two or three persons that could justify a serious consideration of DQ switching.
More details on implementing DQ available in “DQ Switching - An Enhancement to Internet Switching” will be available on this website soon. Personal info available on LinkedIn.
Professor Campbell died on August 15th, 2019, but several groups continue his work. Contact information for them will be added soon, as well as more papers and a simulation. His daughter, Kellas, set up this site to replace his old one and is moving content over.