> “It's just a vast amount of sort of thrown away compute and recycling is a terrible option for most of these smartphones.”
Oddly enough, Google withdrawing support for older, still working phones is also a terrible option.
Once you're done with your smartphone, it either ends up in a drawer, on the growing second-hand market, or perhaps in a recycle bin. However, it's a computer and, when combined with others like it, can offer real processing power. Computer scientists at the University of California, San Diego, working in collaboration with …
> “It's just a vast amount of sort of thrown away compute and recycling is a terrible option for most of these smartphones.”
Which is why Google is nagging me about the battery in my 6a and offering discounts on a new phone.
Building phones that can last would be a better solution than figuring out ways to second-life the phone after screen or battery stops working.
I've never had a phone that didn't last --with the exception of my first Samsung, which I took in to get the screen repaired and the battery replaced and the clown in the repair shop botched the job so badly I ended up just buying a new phone.
My fault for going to Bonobo Bob's Plumbing and Phone Repair I guess. Wanting payment in bananas should've been a dead giveaway, I expect.
The two or three phones languishing in the bottom desk drawer are perfectly servicable except that they're no longer being given system software updates and, as such, are sort of dangerous to use given the porosity of security.
Following the trend set by other sectors, it can’t be long before vendors stop selling phones and lease them. Thus permitting them to “sell” the same device multiple times.
This approach would make it in the vendors interest to make their devices “repairable” ie. Easier to take apart and reassemble for its second life usage.
Here in the States back in the "Ma Bell" days, up till the Carterphone decision1 around 1970, the phone company owned anything connected to its network. It you wanted an answering machine or any other widget attached to the network you had to lease it from the company. It wasn't really until the earlyish 1980s, after the breakup of the Bell System into regional "Baby Bells" to foster competition,2 that consumer owned equipment was legitimately available.
When I was in commecial radio in the 1970s and we wanted a phone patch, to put phone calls on the air, radio station owners being of the tightwad variety,3,the engineer would kludge up a bridge circuit and hope that it didn't bork the phone line.
1 Wikipedia: Carterphone decision
2 That worked out great, didn't it?
3 At least the majority of the owners for whom I worked. Goniff doesn't even come close to describing a couple of them.
Exactly, users don't want to throw their phones after 4 years, but if google don't provide security updates of current software, how could they be reused as a storage cloud?
Its becomes like a public FTP server with anonymous uploads of old?
Absolute greenwashing silly project without continuous software support available.
Exactly, users don't want to throw their phones after 4 years, but if google don't provide security updates of current software, how could they be reused as a storage cloud?
From the article: "For this to work, the devices first need to be flashed with a Linux operating system suitable for the job"
There are many tasks that don't require a lot of compute, or at least less compute than a modern phone can provide. Think stuff like DNS, NTP, some kind of monitoring daemon, etc etc.
Absolute greenwashing silly project without continuous software support available.
You think making use of stuff that would otherwise end up in landfill is 'greenwashing'?
I think it's greenwashing for a different but probably similar reason to the original poster. If I want to do that with old phones I've got access to, will I be able to? Will Google, who wrote the Android code on them, find a way to let me reuse that hardware, or does that only apply to theirs? But fine, I can't blame Google for Samsung not unlocking something or providing required code*. So if I've got a bunch of old Pixels, will they give me this? Or is this perhaps only going to be available for one user and one device, a device that incidentally still receives updates so they didn't need to be recycled anyway? Greenwashing is frequently done by doing something that would be useful but only to the extent necessary to get some articles published, not to the extent that it makes any difference.
* Actually, I can blame Google. Let's say Samsung refuses to let me unlock the bootloader. That's not a problem, as I don't need to. The Linux kernel is already there, complete with driver code for the contained hardware. I just need to delete the Android service stack above it, and Google software running as root has the power to do that, as do I if I have successfully rooted it. Once done, Google can provide the software that runs on top of that kernel. So, will they?
> Google, who wrote the Android code on them
Regardless of support considerations Android wouldn’t be a good fit for compute clusters. A good deal of a mobile OS is designed to support sensors, IO and a UI screen.
Things that aren’t needed here. Plain old Linux is better.
It’d be the same for iOS though Apple has instead talked up having robots that can strip an iPhone apart for materials recycling.
The point of Google having written the code is that they should have the access and ability, and we know they do, to remove the Android parts, strip it down to the bare Linux kernel on which Android used to be running, and replace it with the version they've built here, even on hardware they didn't manufacture. I agree that you wouldn't want to keep the rest of Android while using it this way.
You think making use of stuff that would otherwise end up in landfill is 'greenwashing'?
Not OP, but kinda yeah. To do this you have to strip out the main board and acquire a custom-made power and Ethernet adaptor (which presumably then needs some fiddly soldering and tinkering - how many mobile boards come with Ethernet breakout pins?! I suppose it’s being attached via the USB controller?).
So the reuse requires some virgin manufacturing. And then we’re sending the screen, battery, etc to recycling.
As others have said, better that Google supports phones for longer, or provides an unlocked boot loader for the last update so the phones themselves can continue to be used.
This is not so much reuse as recycling-with-one-component-being-diverted-for-reuse.
They do mention the main board makes up a disproportionate amount of the embodied carbon, so that’s something. I question what the actual Watts/Operation comes out at. We all know that old-generation hardware is less energy efficient, which is tolerable for periodic/client usage but not as good for always-on applications like servers. It’s why old supercomputers get retired rather than donated or handed down to lesser institutions - the cost of power (and the room to put them in) is too high. By the time they’re being retired you can often get the same performance from half as many racks or less.
On a small scale this doesn’t matter since the whole cluster can probably be powered from a couple of solar panels and turned down at night.
The point they make is valid - there’s a tonne of low-computer functionality needed and a tonne of mobile devices with decent compute getting churned out of use.
But the emphasis ought to be on life extension for handsets, which is partially limited by battery replaceability but also by software updates and locked bootloaders - which is 100% in the control of Google.
I agree, but for clarity, Google supports their devices for 7 years. Officially they say 5 years, but the previous Pixel I had went for 7 years...my Pixel 7 Pro is due to be EOL in October 2027...but I expect they will likely go beyond that...it might not be two years more, but there will be some security updates beyond that date...in between Pixel phones, I ran Graphene to plug the gap. In the world of Android phones, Google is the least awful of all of them because at least you can install another OS once support ends.
For me personally, I'm not really that arsed about long term support from a phone manufacturer, I'd much rather they allowed other operating systems to be installed. Allow the software to be my choice. Lack of competition in the software is space is what leads to crappy support life cycles...if we had the option to immediately ditch the default OS, the manufacturers might try harder to keep us on their operating systems...they do want the data after all. If we can swap, they lose a lot of value in the device they sold because they no longer get the data they want.
“Early experiments show that even a moderately-sized cluster of 20 phones is capable of supporting peak submission rates for a 75+ student class.”
I'm impressed. Not.
We were handling larger class-sizes than that with a PDP11 50 years ago. Usually multiple classes on one box.
Yes, I know, modern computing is obese, and AI-slop isn't helping any, but c'mon ...
Back in the day, folk round here were expected to bring their own silicon to work on in the classroom. Typical students, when the course was over they'd just dump everything they couldn't be bothered to take home for the long vac, so the whole of the country is now littered with crumbling circles. Some of the nerds wanted to take theirs home so the jocks would shove those on top, "lintelling" as they called it.
Our CS department had a PDP-11/45 running Unix, and later, a VAX 780 (also Unix).
All our terminals were character-only displays (no X11).
I didn't like having to wait 15~20 minutes for my troff jobs (school papers) to be processed and come off the laser printer, but mainframe batch job turnaround time was around 3 hours.
Our CS department had a PDP-11/45 running Unix, and later, a VAX 780 (also Unix).
All our terminals were character-only displays (no X11).
My first proper job, lo, these many years ago, was as a CAD engineer on some ancient 4-terminal minicomputer graphical system. [1]
I discovered that it featured programmable macros which could then be assigned to a spare button on the big pen-operated touchpad. This confused both my manager and the minicomputer ops manager - both of whom (allegedly) had computer science degrees. [2]
So I wrote a whole series of macros and used them to double my productivity. Much to my colleagues annoyance. [3]
I left there after about 6 months. Initiative and ability were very much not appreciated in a junior staff member.
[1] Can't remember the name. The site also used a Vax but the two were not, IFAIK, related. And calling it 4-terminal was generous - if you had more than two terminals going the performance was *terrible* - knowing who I worked for they'd probably bought the minimum-spec model.
[2] And I'd dropped out of polytechnic after one year of an HND. But I had what they lacked - ability, curiosity and not expending all my time trying to climb the greasy pole.
[3] Who mostly seemed to be younger relatives of managers there and coasting along with minimum effort. To be fair, my brother also worked there but had nothing to do with me working there other than to tell me a job was available.
I suppose it depends what they do when a student submits something in both cases. Since they weren't clear about what they're doing on the phones, let's start with what you were doing. Was that just copying the submission for later grading or something more? For example, if your submission script was just storing the submission for grading after verifying that the files matched what you expected, but their submission script runs a few thousand test cases and a plagiarism check against all the other submissions and a bunch of gathered possibilities, it makes sense that theirs would be more compute-hungry.
Cool my hardware? Why would I want to do that?
I've got a 256K Core memory array from an IBM S/360 Model 50 in storage (from an H50, has 262,144 "bytes" (words))[0]. It is not mounted in a frame. It weighs about 500 pounds, is roughly the size of a three drawer file cabinet, and it runs about 3,500 watts, mostly to keep the core within its most efficient operating temperature.
Yes, memory was often intentionally heated back then ...
[0] It's not from the PDP11s I mentioned above, but only because those '11s probably[1] had prototype Core-to-DRAM conversions ...
[1] I honestly don't remember, but we installed a couple dozen or so of those conversions across the two campuses, so probably.
Of course it is a cooling question. You can clock those CPUs as far as the temperature-efficiency allows, to the limit of throtteling. Standard in consumer tech for 20+ years to clock the CPU as fast as the temperature allows. Whatever computer you use to post here, there is a 99.99% chance it uses this tech. CPUs, especially notable for Laptops, can have TPD of 35 watt, but the Laptop manufacturer limits it to 15 watt to keep the temperature in a sane range without needing a Noctua NH-D15 G2 LBC attached to to bottom.
And thus you feed in as much power as you can, which is for such CPUs within the 5 to 15 watt range per CPU or whole system (RAM + flash etc add heat as well).
They did answer your question. With no cooling, the chips will rapidly throttle and consume less power. With some cooling, maybe possible with passive heatsinks if there's a lot of space between boards and the room itself is cool, less throttling. With active cooling, no throttling, meaning more performance, more power consumed, and even more power consumed by the active cooling. Therefore, how you cool it makes a difference to how much power is consumed.
But fine, let's use some numbers. The Google Tensor G2 processor has a TDP of 7 W. Add some extra for the RAM and storage chips takes us to 10 W. So the boards themselves could consume 20 kW altogether, about as much as 6 of those Gigabyte pizza boxes assuming the power supplies in those are redundant. There are 12 cores per phone, so 24k in total for the Pixel cluster, and 40 nodes per box and 8 cores per node, so 1920 cores (but much faster ones) in total for the Gigabyte cluster. There is 12 GB of RAM per phone, so 24 TB across the cluster, and 32 GB per node, so 7.7 TB (but more available per process) there.
While the reuse is nice, it is too expensive, and requires too much workload and the extra outbreak-USB PCB.
In the real world, to make this viable, this could be used in places where cheap people with time + knowledge are available, i.e. areas of the world which are considered poorer. But will google send them their former phones for free, including unlocked to get a custom Linux on them? Unlikely...
Does seem a lot of work to do what people have been doing with Pis for years. But, unless you can really integrate things onto a single board, you're going to have a fairly inefficient system. In addition, much of the value is in the battery and screen rather than the chips themselves.
Now some way of quickly and easily recyling the chips onto a new board that could fit in a rack might be more interesting and, frankly, useful.
> much of the value is in the battery and screen rather than the chips themselves
You mean, AT THE MOMENT, people are more willing to give you money for the battery and screen, because those are the bits that they can usefully reuse or strip for parts/chemicals - to be reused.
If clustering 'phone boards took off wholesale - or even "just" stripping the chips and reusing those - then those parts of the old 'phones (could) become where "much of the value" is found.
> a lot of work to do what people have been doing with Pis for years
"Only" because the R'Pi and similar are available uncased and with easy to access pins. Preferably, modify the 'phones to allow the same at end of life[1] but failing that if you can get enough 'phones of the same model (like Google has here) a jig to attach connections to the relevant places quickly and a carrier board design to hold (partial) clusters are both viable: as clusters of headerless Zeroes have shown.
> some way of quickly and easily recyling the chips onto a new board that could fit in a rack
As the chips in 'phones are tightly integrated does it really make sense to strip the off individually, respool them then pick'n'place onto a new server board, rather than have a carrier for the existing entire PCBs? You'll still have to extract the PCBs from the casings to strip them (*maybe* you could do so with a little more prejudice).
[1] fat chance getting the manufacturers to agree, BUT if it is clear that the required mods are for afterlife use, NOT intended as a way for Joe Bloggs to repair and keep using his 'phone, they will be more agreeable: Joe still has to buy a new one and the manufacturer charges for using there screwing-together knowledge in reverse (e.g. don't pry at the phone case, just drop it in IPA and the hot glue will release - oops, forget that, it is a confidential commercial secret!)
You mean, AT THE MOMENT, people are more willing to give you money for the battery and screen, because those are the bits that they can usefully reuse or strip for parts/chemicals - to be reused.
Er, no. This is also in the bill of materials. Chips are cheap chips for a reason.
> This is also in the bill of materials
No.
Pricing in the BOM is the cost of those materials at the time of manufacturer; how much you value what *you* pay in order to make the thing new.
Which includes a lithium battery that has its entire charge cycle life ahead of it. A pristine screen that has been lavished with lint-free cloths, bubblewrap and careful handling. And includes how much value you ascribe to getting parts, in a timely fashion, to make a cutting-edge, latest model[1] 'phone that you can a premium price for.
We are discussing the value that someone else then puts on AFTER the 'phone is otherwise deemed landfill (fairly or not).
The lithium battery no longer holds a useful charge. The screen's glass is permanently scratched, no consumer is paying full whack for it anymore.
> Chips are cheap chips for a reason
>>> easily recyling the chips onto a new board
Which is why stripping, respooling etc to reuse the chips as parts, as you suggested, is not worth it.
BUT those individual cheap chips have had lots of lovely value added AFTER they appeared on your BOM: the chips are now neatly arranged on a PCB, correctly connected to their support devices...
(curses, missed the edit window)
The screen's glass is permanently scratched, no consumer is paying full whack for it anymore, and (with a new enough 'phone) the PCB can feed the UI down the USB-C, so in the cluster the screen is an undesirable power draw.
Who's a clever boy, then? You are, yes you are!
You hadn't noticed the thriving economy in replacement screens or the questions on forums that start "I dropped my phone and the screen is broken/badly scratched, what is a good replacement nowadays?".
I know it happens and I've dropped my phone from a moving bike several times – it helps to have a good case, which is going to be cheaper than a new screen. Just goes to show that not all people trash their screens, thus rendering the rest value of it when they change phones worthless.
You'll still have to extract the PCBs from the casings to strip them (*maybe* you could do so with a little more prejudice)
Easier and cheaper to just crush the whole thing into tiny, tiny pieces and then drop the fragments into a nice hot smelter to extract the metals..
If that is all you want to do, totally ignoring any concept of reusing them as working parts.
But the whole point of TFA and this discussion is to look at extracting more value from the 'phones *before* they are simply rendered for chemicals.
High-end mobile chips are somewhat valuable, especially when they're being discarded so are cheaper than if you had to try to buy them from their manufacturer. Since they're talking about devices old enough that people want to replace them, the battery's value as a battery is mostly behind it and, since you don't need it for the compute part, it can be sent for lithium recovery. The screen is trickier to reuse anywhere else, but it's true that a lot of money went into making it and probably won't be coming back out.
Agree. Serious research, convenient PR for Google.
That’s indeed a lot of (very skilled specialized) labor to generate lowish savings. One thing that causes me pause is that having 2000 near identical phones donated en masse by the vendor may not be representative of challenges with more heterogenous supplies. Another is the challenge of dealing with a cluster with components that will fail at different times but whose replacements may again differ. Not insurmountable but adds more complexity.
Still, well worth looking into. Maybe an optimal near-automated configuration method can be figured out and maybe some of the strip and reassembly work can be done in poor countries (where quite dangerous home smelting methods are used to recycle electronics). And maybe this research just crosses off this type of solution as unsuitable: a big part of engineering is figuring out what doesn’t work.
The collection part at least isn’t too inherently problematic: some manufacturers have trade ins when you buy a new one.
That idea remembers me of the end of the first Internet bubble when I used in our fast shrinking company free laptops as headless Linux servers for web development/staging servers.
My biggest issue with all that old mobile hardware is that much of it is closed. Apple is the worst offender as many Android devices can made to boot alternative software.
We are at a point where even 10-15 year old hardware can serve a lot of use cases (mostly media consumption) but with the closed Apple environment and the forced WebKit for browsers that becomes a security risk.
We need legislation that EOL hardware needs to be opened.
That would boost the second hand market and I do not think it will hurt sales of new devices that much.
Apple laptops that use X86 processors (ie pre M1 CPU's) run Llnux very well. Apple M* series is a work in progress but so far none of them have been EOL'd.
My 2018 MBP runs Rocky Linux nicely. It will be EOL'd with MacOS 27 later this year. I bought an M2/Apple adapter and tried it. It worked.
This post has been deleted by its author
(for a dubious use of practical) might be to ship the phones mainboard† to Ukraine as I imagine the gps and cellular functions along with the camera(s) would be ideal for embedding in an autonomous weapon or reconnaissance system (drones.)
† I notice that neither Kastner nor the author shrivelled up and died of shame referring to motherboard so perhaps we're done with that precious nonsense.
First, getting the obvious out of the way: The cpus, gpus, and memories on these things are still good to go. If they are hooked to a server cluster or still in your hand, it is irrelevant. Power management is also probably there, on par with best practices. Cooling would be also quickly solved (since the originals had no active powered cooling whatsoever, being phones). But the old batteries are still a liability. Just as they were about to become spicy pillows in your hands, nobody wants that in a server environment.
Now, on the other side of the equation: I could see Colleges and Institutions giving discounts for students that donated specific models of old phones still operational, batteries excluded. Or the population in general (even in 3rd world countries) donating their old phones for STEM causes, whatever they are.
A worldwide NGO just to retrieve and process phones in an industrial scale, while sending the batteries to recycling in Countries with that capacity.
It solves two or three worldwide problems with phones. Cheap compute, e-waste, and logistics for retrieving batteries for recycling tackled in a single lump.
It was the statement that the batteries weren’t safe to use in a data centre that caught my attention - if that’s the case, how go Google think they’re safer in people’s pockets? Or are users of old phones no longer of any value…
I think it's partially that they intended to run the CPU at high load for a long time, which would make a lot of heat. Android has software intended to limit that, and the phones were built around the assumption of that software, so running something hotter than it would normally go next to a battery that has already seen a lot of wear might eventually cause something to fail, and if one of them catches fire then having a bunch more batteries right next to it is not ideal. Neither happens normally in your pocket. It's likely neither was going to happen in the server rack either, but since the batteries weren't useful, taking them out made that risk zero instead of low.
Nintendo Switch 2 gets a specific EU version due to that. The hard enforcement day is 18th February 2027, and even Apple changed their designs accordingly.
That's a good move by the EU, but well overdue by at least FIFTEEN YEARS, so applause should be appropriately muted.
I had a look at the iFixit guide to replacing the battery of my Pixel 9 Pro (pretty much the same as the 8 and 10 variants) and it is horrifying how difficult Google intentionally made the task of battery replacement. It is pretty clear that (a) Google don't give a shit about the environment, and (b) this article shows the company engaging in its usual mendacious greenwashing.
Google estimates wishes that the average person upgrades their phone every four years or so
I've had precisely two phones in the last 12 years. New one last February, my wife then got my old one.
So, as a household, we fall *well* below the Google-wished-for [1] average. Just like we do in cars [2], laptops [3] etc etc.
[1] Or more accurately, Google wants us to replace our phones with new Pixel phones every year. Just to make sure that no other VAR gets a look at our information as well as them.
[2] I average 8 years on my car replacements. My wife is running at about and average of 25 years..
[3] I change mine when the Mac-Killer Cat manages to introduce a red wine/fruit juice/coffee Mac inundation event. So far, three times in 6 years. Current Mac sits up off the table on a metal stand [4] and the lid is *always* closed if I'm not using it. And coffee goes no-where near it.
[4] One of her sneakier laptop kills was knocking over a glass of white wine several feet away from the laptop. Said white wine reliably ran under the Mac, soaking into the thermal vents in the curve near the bottom of the base. End result, one cooked MBP motherboard. So now it no longer sites directly on the table. Previous two were glass of orange juice knocked onto the open keyboard and a half-cup of coffee knocked off the side onto the (also open) keyboard of the replacement. She's gone suspiciously quiet in the laptop slaughtering stakes since the white wine incident.
I recalled reading this story last year. Looks like any number of people are sorting through cellphone "trash" to build computing clusters. A team at the University of Tartu’s Institute of Computer Science has shown that obsolete mobile phones can be wired together to do the sort of heavy data processing normally reserved for expensive server farms while keeping thousands of handsets out of landfills.
Led by associate professor of pervasive computing Huber Flores, the engineers stripped the batteries from four discarded Google Nexus devices, fitted them with 3-D-printed holders, and powered them from an external source. The cluster cost about €8 (US$9.20) per phone, yet comfortably outperformed popular single-board computers such as the Raspberry Pi for image-analysis and website-hosting chores.
But as the article in The Register mentioned, you do need to extract the batteries from the cellphones and use an external power supply. No one wants a lot of potentially dangerous cellphone batteries all piled together.