Glad someone else appreciates circuit switched technologies though :) . It really does shine in any application that isn't burst-y data transfers.
if you’re using IPv6 (nobody does) that increases another 20 bytes.
You’d typically have a receive buffer in the 1ms range though , although in theory you’d get away with 500us or even 375us on paper.
I believe (not really my area, in video we deal with timings starting in the tens in the tens of microseconds, and quantisation in the milliseconds) the rule of thumb is the total system latency between microphone and speaker have to be all within 5ms to avoid problems
Using FPGAs you can get to very low latency.
Also at some point the latency is a non-issue.
Speed of light even in fiber optics is significantly faster than the speed put sound.
And compared to TDM which is the comparison here the only thing added in packet networks latency wise is buffers, for network switches that can be nearly instant so it is only on the send and receive side.
So pretty much if you want to drop dark fiber between endpoints you can get the exact same benefits of TDM on packet switched networks without the disadvantages.
The people asking for TDM are ignoring the fact the literally several industries and some that care greatly about audio quality and latency have prettier much determined that packet switched networks are fine.
Now, on many calls, the first few seconds are spent on "Do you hear me? Hello? Yes I hear you.". (At work, we call it the Microsoft Teams Random Number Generator - every call gives you one bit with around 50/50 chance (aka one bit)).
by far the clearest phone calls i've ever been on which were not landline were 2001 era Sprint CDMA, probably because the vocoder bitrate was higher than in later configurations (like LTE), or the MOS scores were held to a higher standard then.
Phones of old could be high quality, and they were always extremely low latency. But usually they were scratchy or muffled or noisy or otherwise compressed. A Teams call on a low-latency low-loss line with good microphones on each side is a far better experience than any long distance call from the 90's. It's just that normally the data connection is not so good.
There's an unverified mention of a fiber cut, which might result in fewer circuits being available. That usually returns an "all circuits busy" signal.
None of the explanations will make sense until they stop coming from unqualified political appointees.
A TDM system can't have infinite throughput, can it? Either you operate with fixed slot windows, as legacy systems usually do, in which case excess data has to wait and buffers fill up, or else you operate with variable time slots, and at some point the windows get too narrow to transmit usable information.
It doesn't make sense to me to claim you "cannot overload" a TDM system. Bandwidth is finite.
Busy signals are literally a capacity limit being hit.
Ah, the glorious 3kHz frequency range of old.
Your feelings here have no technical basis. TDM is pre-provisioned, bandwith-reserved, latency-capped network that was probably hard for packet switched network 25 years ago. But with all that has gone in -- switched ethernet, qos, oam -- there is absolutely no reason to go back. Heck, even whatsapp calls across 12-hour time zones are quite crisp and free.
Good to remember that a WhatsApp call is being routed over a much more diverse, heterogeneous set of networks than a Western Electric 500 phone was. 32-bit stereo raw audio is roughly 2.82 Mbps, which is a lot better than what an older phone did. On any link between two endpoints with 20 Mbps or more symmetric, you can do this trivially. WhatsApp can give you good performance between two crowded venues on completely different parts of the world. Just a very different experience.
Thank you! learnt about Henry Dreyfuss. It's sad the way he died, but can understand.
What kills realtime is buffer bloat
basically, for realtime stuff, you want to prefer packets being dropped rather than buffered.
this is an excellent resource on the subject: https://www.bufferbloat.net/projects/bloat/wiki/Introduction...
this would beat the laws of physics. realistically, it is more like "fixed latency, and very little jitter".
I don't think anyone seriously think that 5g and VoIP voice quality is bad - but latency is absolutely a problem, especially in group calls.
I have everything from dedicated inter building fibre cables, through cores, wavelengths, ethernet services, IP circuits (still have one international MPLS and a VPLS network eoth 4 exit nodes), to normal internet with varying levels of contention.
Even the fibres aren’t private - they are patched through other people’s equipment rooms, even on short runs if a few miles, and even on the circuits that don’t, they run through ducts over other people’s land.
Packet switching has variable latency. This can be controlled with ptp and qos, and in reality we’re talking sub microsecond differences if you need it, switch latency is measured in nanoseconds And that’s way below the timing ministry is worried about (which is in the tens of microsecond level). TDM will give you guarenteed latency though, you get a slot every X time units regardless, jitter isn’t 2us, it’s zero.
I don’t believe any application needs sub microsecond jitter, but there may well be some scientific applications that do, but at that level you’re running on dedicated cores or maybe wavelengths, between specific equipment.
It won’t surprise me ifs y places are services provided by other companies - either ip or ethernet, to connect to larger sites. I suspect some will even be connected via public networks.
Major sites may have dedicated fibres between them but whether that’s just a wavelength on someone else’s core, or even a dark fibre but just one core on someone else’s cable, is anyone’s guess. I’d be surprised if they could justify the cost of a highly resilient network dedicated to themselves rather than renting capacity.
What I’m certain is they don’t own their own fibres to every end point.
Some US airports (~90) are still using paper flight strips while other jurisdictions long ago moved on: https://enotrans.org/article/electronic-flight-strips-brief-...
If you can afford to allocate statically the space for all the variables that your program will ever use, you will certainly increase a lot both the performance and the reliability of your program.
Alas, for many tasks this would result in a requirement for more memory than you can afford, especially today.
To use more efficiently the limited amount of memory that you have, you need to use dynamic memory allocation.
The same happens with your communication infrastructure. Using it with packet switching reduces the reliability and the performance of individual communications, but it uses the capacity much more efficiently, i.e. the aggregated throughput of all communications passing through a given infrastructure is greatly increased.
TDM was more stable because nobody was hopping on meetings while walking their dog on a noisy street and then taking the elevator back home. The expectations have shifted.
Modern consumer links (even over WiFi) are usually adequate enough to do reliable, high-quality voice calls over the public internet. We basically choose not to; most VOIP apps are optimized to work decently in a noisy car driving through a forest when on speakerphone, instead of working very well in a quiet room on headphones and crapping out everywhere else. This is a choice we make; there are apps out there that make different choices. Once upon a time, I was very unhappy with that choice, then I grew up and have experienced what the consequences of making different choices really are.
With that said, classical Bluetooth audio urgently needs to die. There's no reason for it to be as bad as it is, and it is truly abhorrent. Better protocol design is definitely possible; most wireless (2.4GHz, gaming) headphones have much better quality and much lower latency. BLE audio will fix most of Bluetooth's warts, and it is slowly, so very slowly appearing on the market.
Of particular note is this section: Risks of FTI Outages for Existing Operational Services As of December 2009, Harris reported that there are a total of 20,982 services operating on the FTI network. While FAA has transitioned all 1,808 IP services supporting FAA flight plans to the new FTI fiber optic network, additional risks will be introduced when the Agency begins transitioning other FAA services, such as En Route Automation Modernization (ERAM) and NextGen platforms, which also require IP technology to operate. Moreover, there are 19,174 existing services that may be transitioned to the new FTI fiber optic network. According to FAA, transitioning these remaining services is important because FTI now connects older FAA systems that provide safety-critical voice and surveillance radar information by utilizing Time Division Multiplex (TDM) technology, instead of IP services.