Back in the 70s, and maybe the 80s, on New Years Day and other such events, it was sometimes impossible to make a phone call, you would get the NC intercept (No Circuits), meaning that somewhere there was a bottleneck, where the number of calling lines had reached capacity. Sometimes you couldn’t even get a dial tone.
Is there, today, an attainable limit to the number of people in the world who can be talking on remote devices to each other at the same moment? A denial of service threshold?
Yes. Of course.
Speaking just to the US:
Local overloads still routinely occur on both landline and mobile/cellular systems. The typical cause is a disaster or panic of some sort. Large assemblies of people in unusual places will often overwhelm cellular coverage even if the telcos bring in temporary mini-towers. Think outdoor concerts, parades, and such.
By contrast, venues like stadiums generally have extra provisioning built in to accomodate 50 or 70K people in a confined space. Even then, if suddenly everybody at an NFL or MLB game tried to place a call it ain’t gonna work.
What *has *changed from the Olden Tymes is the relative provisioning of local versus long distance capacity. It used to be that on those special holidays the demand for long-distance skyrocketed versus the usual load. Back in the 1960s some people only splurged on long distance calls a very few times a year. But all those people did it on Mother’s Day and on Christmas. Resulting in overloading the supply of long distance circuits.
Nowadays the base volume of long distance traffic is vastly higher compared to the base volume of local traffic. So the provisioning is greater too. As well, not so many people have to call on Mother’s Day specifically because they probably called yesterday or the day before anyway.
So it becomes much more difficult for end users to overwhelm that part of the system.
As to the rest of the world …
You’re a seasoned world traveler. As you well know some areas have pretty decrepit infrastructure prone to all the same problems. I’ve never tried to place a long-distance call from a small town 150km from Bangalore to another small town 150km from Chennai. But I bet it’s a fraught experience. Especially on holidays.
When the earthquake and then tsunami occurred in Fukushima my wife was in Germany giving a paper while I had remained in Tokyo.
While I couldn’t get through to her via cell or land lines, the 3G data link allowed me to Skype her.
Eventually that got overwhelmed as well but recovered sooner.
That contains the core point of the modern world. Ther is vastly more IP traffic than voice. Almost all the bandwidth in the world is in the form of data, and only a small amount is dedicated voice. All are digital in modern systems, but the underlying protocols are different. It used to be that voice and data shared the same infrastructure, typically using ATM, and IP traffic was reticulated over the same ATM system as the voice links. Nowadays ATM is waning and the shift is to run IP traffic directly (not making use of the circuit switching nature of ATM) and further, to even use the IP link for voice. Quality of service issues are the main problem. Basic IP doesn’t have the same level of QoS as ATM, and additional protocols to try to bridge the gap are used. But Voice over IP (VoIP) is now a pretty much an established thing, and many people use it instead of a conventional home phone.
So currently there is vastly more bandwidth carrying IP traffic in the world, and so long as you can get a voice channel (ie via Skype, or VoIP) you will not see the same level of congestion as on basic voice links. But there are bottlenecks. Mobile data is still limited within each cell - not everyone in a cell can use data at once. And many parts of the globe are still running with much older infrastructure, including analog trunk lines and mechanical exchanges (or worse).
Right now, it would not surprise me if everyone in the US could make a Skype or VoIP call from home to the home of another person in Europe without the infrastructure overloading. (Apart from perhaps the core servers that control call initiation.)