I am, allegedly, an expert in telecoms. This is not possible, as the telecoms industry does not really exist. So what does exist?
It is nearly impossible to be an expert in an imaginary thing. The exception is a part of art and literature called “pataphysics”, which is a pun on metaphysics. It is the jokingly invented science of imaginary solutions, ones that we have merely conjured up in our heads.
I wish to put to you a possibly disturbing pataphysical idea. The industry called “telecoms” does not truly exist any more, if it ever did. It is a bedtime fantasy story we use to comfort ourselves that we understand the world.
We urgently need a new and better story if we are to engage with reality rather than fiction. To do this we need to address two distinct, but related, naming problems: the “tele-” and the “-coms”.
The problem with the “tele” bit is that it implies that the goal of our industry is anchored in overcoming distance. It is self-evident that telecoms networks span our globe, and also extend above and beyond it. (How far down is a devilish question I don’t know the answer to…)
That means distance is obviously taken as a given at one level. Yet defying distance is not our core purpose. We can readily and cheaply send information on postcards or DVDs over the whole planet at low cost. So it cannot be what defines us. Rather, our value must lie somewhere else.
So what is it? Well, what networks uniquely do is to collapse time. No matter how high you set the union rate, FedEx won’t move your parcel of packets around the world in milliseconds. We are the information industry that exists to alleviate the tedious pain of waiting. We shorten delay, rather than sever distance.
That means the value we offer is solely in the ability to expedite information. When it arrives from afar with sufficient freshness, that is success. If it is fouled with excess fungus of the future, it is useless for computational consumption.
It is a bit like networks contaminate packets with “chronotoxin” as they pass through all the elements, and across all the links. What we do is to limit the damage in transit. It is impossible to avoid some chronotoxic impairment, since the speed of light sets a minimum, but every application can tolerate a certain amount.
If this were the only issue, then we could quickly rename our industry. Rather than “telecoms” we could be “chronocoms”. (For the pedants, “achronocoms”, or maybe “anti-” or “non-“, or perhaps even a “distempocoms”. I don’t lay claim any classical linguistic understanding or accuracy.)
However, there is an even more profound and serious issue lying underneath. We don’t even do “-coms” anymore.
The true telecoms industry is essentially a bunch of circuits that transmit data with fixed timing. That business model is receding fast. Just as the cloud statistically shares computing resources, we now statistically share network resources using packet data.
The acts that routers perform when directing and scheduling packets are inherently computational ones. The end-to-end copying of packetised information may superficially look like a circuit service. It just isn’t.
What packet networks do is to compute the identity function. They are specialised stretchy computers with a side-effect of copying the information from one place to another, albeit imperfectly and tardily. “The network is the computer” was nearly right: it is a computer. We just have to also give some credit to the rest of the computers involved in the distributed computation.
This means the routers in packet networks are not really part of the telecoms industry. Sure, they use telecoms transmission facilities as an input to make that “information relocation” work. No doubt telecoms analysts examine their features and finances. They may even be owned by telco shareholders. But…
…they are more like a cloud data centre than a true telecoms facilities. We are misled by the fact that these routers are “out there” in the field, sitting in the same racks as “real” telecoms transmission. These routers are computers that are really the “field agents” of a distributed computing infrastructure.
When your browser talks to a Web server, the routers on the way are taking active part in the distributed computation of your application. They are not merely passive accessories to a consenting act of data transmission between end nodes. They “compute the performance” of you application. So a more honest description of what we do is “identiperformancecomputing” rather than “communications”.
Now, even I have to admit that “achronoidentiperformancecomputing” is only better than “telecoms” if you are playing an amusing word game with friends. It’s not going to fit into the column heading of the Financial Times above AT&T, BT or NTT. Yet “chronocomputing” is indeed what we are all about, not “telecoms” any more.
These two category errors, based on “tele-” and “-coms”, result in real-world consequences. For instance, the whole “net neutrality” debate is a branch of pataphysics. It is an impossible-to-enact imaginary solution to a fantasised problem, cause by a fundamental mislabelling of the service on offer.
The attempt to force a computing service into a telecoms circuit paradigm has resulted in a large volume of academic and regulatory nonsense. That is because those involved falsely believe they are debating telecoms regulation, which cannot be true, since telecoms is effectively gone.
This unavoidable present reality requires us to engage in some fresh thinking. We need to deal with the new industry of “chronocomputing”. Otherwise we’re stuck with pataphysics for answers.
For the latest fresh thinking on telecommunications, please sign up for the free Geddes newsletter.
Read It Now
