TECHNICAL
Summary This document is an updated version from an initial version (circa 2023), and to date, the “industry” is even further from any agreement/consensus. We will continue to sow the seeds of confusion and to use an often-quoted phrase of mine “this is where the bunfight starts.”
c. What is the measurement unit and (more importantly) where are the measuring points? 2. Who owns the collation of measuring point information? 3. If there is no agreed end-to-end mechanism for doing so, why do so many commentators describe (and quantify) Internet traffic with such gusto?” 4. As mentioned within the introduction to this paper, my interest was originally piqued following many discussions within the “fair share” debate. My question to the protagonists is quite simple “prove it i.e., demonstrate the amount of traffic being used by Big Tech over the Telco infra” – this must be an engineering piece of analysis, otherwise it is an exercise in futility. 5. Speaking on behalf of an engineering approach (rather than the loose commercial model promoted by many) that a definition and metric for Internet traffic should be promoted and supported by all, and I am tempted to suggest we plagiarise the circuit switching approach using RFC 9522 as the starting point. This is where the IETF may be able to help? I further observe that there is some confusion from within the current EU Regulatory body, concerning the terminology of the Internet and Telcos being “intertwined” – this is reported upon and considers the ramifications of a recent consultation. The politics of the review are of no interest to me, but without an agreement of a metric, such a review is meaningless. I am further suggesting we name any such metric a “Poisson. ”
The challenge, as I see it is that each network within the model illustrated in Figure 6 is invariably, under the control of a unique management domain. Whilst systems can measure traffic (again in bit/s, or multiples thereof?) within a network management domain, how do we measure traffic that flows across multiple management domains? To continue, at what period of time and duration, do we measure and who “owns” the measurement? My thought process is that such a model could be a starting point to address the above, but will need all interested parties to particpate – logically this should progess under an IETF workgroup, or maybe there is an existing workgroup for any such an activity? What unit do we use – I currently observe multiple instances of the use of bit/s (again, or multiples thereof). Are these examples indicating the x bit/s flow across the Internet? The only way to confirm this is by measurement. Closing observations I have yet to find that elusive definition of Internet traffic and similarly to earlier comments, if I am wrong, I will gladly promote that approach. It may be that this paper is a starting point.
Further work is necessary, if only to dispel, mistruths.
www.firstmilenetworks.co.uk
In summary, and from this research: 1. Continuing with the flow model
introduced earlier, traffic must enter/ exit a network and hence: a. What is the standard (and agreed) definition of Internet traffic? b. Where are these measurement points from within the “Internet ?
Volume 47 No.4 DECEMBER 2025
91
Made with FlippingBook - Online magazine maker