How much power does BLE draw?
We periodically pose a question to our devs on slack, then post the lightly edited conversation here.
How many BLE advertisements would you say are advertised globally per day and how much power is consumed by all that broadcasting?
dean Seems like this question would require estimating quite a bit about the number of devices are on the planet that have bluetooth turned on.
I assume zeke will be able to help us out with estimating how frequently devices advertise.
20ms is the fastest?
between 20ms and 10s depending on how much battery you want to use
i mean, the number can be pretty high, for one device in a day…
So that’s 4.3 million advertisements in one day for one device, max
And, if constantly broadcasting at the slowest rate, 8,640 in one day
i wonder if it compares to number of hashes per day…
That’s an interesting comparison
What do we know about iPhones in typical usage?
is that somewhere we could start? Are they closer to 20ms or 10s, on average?
did someone have an estimate for number of IoT devices out there? (I wonder what percentage have BLE)
The recommended advertising pattern and advertising intervals are: First, advertise at 20 ms intervals for at least 30 seconds. If not discovered after 30 seconds, you may change to one of the following longer intervals: 152.5 ms, 211.25 ms, 318.75 ms, 417.5 ms, 546.25 ms, 760 ms, 852.5 ms, 1022.5 ms, 1285 ms
But the iPhones (and every other smartphone) are advertising themselves.
the bitcoin network hashrate is approximately 177 EH/s
aka 177,000,000,000,000,000,000 hashes per second
Does Bitcoin have relation to BLE advertisements?
I’d say, on the global average, there are 3 BLE devices per capita. Someone in a wealthy country might have many more. For example I have my phone, watch, headphones, thermostat, door locks, laptop, and I am sure several other things which send out BLE advertisements. There are lots of BLE advertisers that are not owned personally, like iBeacons.
8B people * 3 BLE devices = 24B devices
10 advertisements/device/second * 3600 s/hr * 24 hr/day = 864k advertisements/device/day
24B devices = 20,000,000,000,000,000 advertisements/day
so about 20 peta-advertisements per day
there are about 10 thousand times more bitcoin hashes per second than there are BLE advertisements per day.
ok, so how much energy does the bluetooth advertising industry use annually?
BLE transmit power is average 10 mW?
It’s about 10 mA @ 3.3V, right? How long does it last?
advertisements are 31 bytes max and average bitrate is 1Mbps
BLE is like max, 15mA but depends on output power
Does the packet length extension extend any advertisements?
Also, is there any overhead other than the transmission of the bytes at the bitrate?
Yea. It’s up to 255 bytes
I assume there is some overhead, but maybe it’s a wash with the fact that the signal is modulated?
255 Bytes * 8 bits/Byte / 10^6 bit/s = 2 millisecond
Does that sound reasonable? That is assuming the largest possible advertisement.
For scaling purposes, let’s just call it 15 mA @ 3.3V for 1 millisecond
15e-3 A * 3.3V * 1e-3 sec = 50 microjoules/advertisement
So, for the whole system:
50e-6 joules/advertisement * 20e15 advertisements/day = 1e12 joules/day
There’s 2.8e-7 kWh per Joule
So that’s about 300 MWh/day
Put in terms of power, there are 24 hours in a day. So, it takes about
10 MW of continuous power to supply all the BLE advertisements in the world. That is about 1% of the capacity of a large
1 GW gas, coal, or nuclear power plant or 1 large
10 MW wind turbine.
My guess is that the 3 BLE devices per global capita is low.
If it’s 30 device per capita, then it’s 10% of a single large power plant or 10 large wind turbines!
Hmm, techjury.com claims there are 46 billion “IoT” devices in the world, which is just 6 devices per capita.
I wonder how we should consider something like Airpods, which is actually 3 different devices, each of which broadcast. There are tens of millions Airpods sold (and how many lost) and there are a lot of copycats out there now too.
I suppose Airpods broadcasting for 3 devices is a rounding error on the difference between 10s and 20ms frequency.
Lets see, looks like Danny estimated every 100 ms as the average for his calculation, seems reasonable
How many of the ad packets are 255 bytes vs the 37 bytes in 4.2 and lower?
that’s another place where we could be off by an order of magnitude
Totally. But if the framing is “does Bluetooth advertising meaningfully contribute to global energy consumption?” I’d say it a clear “no.” “
Looks like we’ve arrived at a reasonable answer: Somewhere between 10 and 100 peta-advertisements per day, globally.
That’s a big number, but it’s not a ton of power, and it’s a much smaller number than just about anything bitcoin related