# What would make the internet go away?

Assume that we're in roughly our currently timeline in present day, plus or minus 5 years (for increase or decrease in technology that would/could make a specific method plausible): what event or series of events could lead to the internet "permanently" going away? I say permanently in the sense that it could be rebuilt, but more or less starting from scratch.

Hopefully this (series of) event(s) will have minimum loss of life, be as subtle and undetectable as possible to avoid attempts to prevent the loss, be as sudden as possible, and leave as much other technology intact as possible. The only thing that I want gone, if possible, is the internet.

Edit: obviously there would be repercussions from the loss of the internet that would be unavoidable; I'm merely trying to limit the impact of the event itself that would cause the loss of the internet, then exploring what would happen as a result.

Edit 2: I'm also trying to avoid any cultural or sociological reasons for abandoning the internet; I want this to be something that could theoretically (perhaps with a little hand-waving if necessary) happen "today" in "today's" society, again give or take 5 years.

Edit3: My question is different from this one in that I don't want state actors/malevolence to be involved. I'm only looking for natural disasters or, worst-case scenario, negligence, accident, or oversight.

• The only thing to go being the internet? So phones still work? Because the internet used to run on phone lines.... There's also wireless internet (wifi, heck even radio waves). In other words, the internet itself is (mostly) decentralized; so long as there is some mechanism of communication there can be some form of the internet that's possible. Taking away any ability to have an internet also means taking away phones, radios, satellites, etc – cegfault Nov 4 '19 at 17:11
• RFC 1149, A Standard for the Transmission of IP Datagrams on Avian Carriers; and its update, RFC 6214, Adaptation of RFC 1149 for IPv6. – AlexP Nov 4 '19 at 17:15
• If my cousin cannot play his favorite network game in his mother's basement anymore, then he seems likely to survive...but is it worthwhile to go on living? – user535733 Nov 4 '19 at 17:30
• What exactly is the meaning "from scratch"? Restarting all hardware? Fixing all broken links? Manufacturing all computer chips? Whiting new communication software? All of the above? – Alexander Nov 4 '19 at 17:39
• @cegfault These days it would be more accurate to say that phone lines run on the internet... I very much doubt that you could take out only the internet, because of how many things that are seemingly unrelated actually use "the internet", or its underlying infrastructure. – anaximander Nov 5 '19 at 14:35

A really big solar flare could do it.

Something on a similar or greater scale to the Carrington Event back in the 1800s. Getting hit with a flare that large, or larger would damage or destroy unprotected electronic devices worldwide. Most of or all servers globally that didn't have military-grade EMP protection would be wiped. The GPS would be gone, along with most other non-military satellites of all kinds. Most cellphones and cell towers would be damaged or destroyed. The entire communications infrastructure for the entire world would have to be rebuilt, and any data that isn't stored on a protected, non-cloud backup would be gone forever. It wouldn't be a COMPLETELY clean slate, but it would be pretty close.

This meets your 'can't stop it' criteria because the effects would be felt globally and without sufficient warning to do anything about it. A true Solar Flare only takes ~8 minutes to reach the earth, an energetic Coronal Mass Ejection takes a day or two, not enough time to do much about it.

Something to be aware of is just how much modern technology is dependent on the internet to function, so if the internet goes, a lot of other stuff goes too. Most global commerce relies on the internet to function. Banking likewise, so there's no "get rid of the internet" without MASSIVE disruption to society. Rural areas will get by more easily, but big cities, especially big cities in developed nations would face massive crises because there would be no way for people to get food. Loss of life would be in the tens of millions at least, primarily in large cities, or remote areas that don't produce food.

• Comments are not for extended discussion; this conversation has been moved to chat. – Monty Wild Nov 6 '19 at 22:49

Let's start by addressing the "leave as much other technology intact as possible" part first:

10 years ago, you could have taken down the internet and most things would have continued to function just fine, but today, almost everything that people use most has some sort of cloud based element, that if removed, will cause your software to break. The Office 365 and Adobe CC products which most people use for the majority of their multimedia reading and writing applications can only go in offline mode so long before they will fail to authenticate their license and stop working. Most top tier video games these days are some manner or MMO, and those that are not still often have similar Digital Restrictions Management components to the aforementioned microsoft and adobe products. Many ad-funded freeware applications that can not find their ad server will crash as they wait indefinitely for content that they can not reach.

But this inconvenience in your personal life is the warm and cuddly aspect of what no internet means.

The truly scary part is that most business grade software relies on decentralized data. Nearly every businesses in the world would come to a grinding halt when they find themselves suddenly unable to access their inventory data, their financials, their client lists, etc. Shipping companies could no longer fulfill their orders, manufacturing companies could no longer receive materials or distribute their final products, most experts would be too young to know how to fall back on traditional skill sets to get the same jobs done, many tech companies could no longer deliver their services, most phones would stop working because all those cell towers have their own DNS addresses, and most land lines are now VoIP. the list goes on, but long story short is that the economic damage would be so extensive that the world would not just revert to a pre-info age state, but it would in many ways revert to more of an early industrial age state because we no longer have the pencil-and-paper infrastructure that first spanned the gap allowing us to get from basic localized industries to globalization... Not to mention with the economy that fully collapsed, there would be massive famines, wars, rioting... things would get really ugly really fast.

Now let's address the outage itself:

It is REALLY hard to knock out the whole internet, and even harder to keep it down.

My original hypothesis of a Root Name DNS Server attack seems beyond the realm of plausibility based on feedback. While there are only 13 Root DNS "servers", their distributed nature and massive redundancy of possible backups makes an attack against them highly improbable to not be easily recovered from.

Another common hypothesis you will see in other answers are solar flares, but these only cause very temporary interruptions. Any flare weak enough to not completely scour the world of technology and perhaps many forms of life will not cause permanent hardware damage to more than a tiny % of internet connections. Once the flares are done, the internet would go back to normal. No big deal.

Viruses also don't work because there are too many different tech stacks at play for any one virus to be able to come close to wiping it all out.

If you are REALLY rich, purchasing all of the world's ISPs and simply firing everyone so that they all shut down would work in theory, but human nature would not tolerate such an unethical choice and the world's governments would likely just seize your assets and charge you with cyber terrorism, breach of contract, monopolization, child pornography, and whatever other real or made up things they can accuse you of to justify seizing your assets and turning the internet back on.

This basically just leaves government censorship. No populus would simply allow their government to completely shut down all cyber communications in one fell swoop, but governments all the time get away with restricting the services their citizens can access under the guise of National Security. Normally this is done by blacklist meaning your government regulated ISPs wont permit you to visit X,Y,Z sites. A more restrictive approach would be to make the internet whitelist censored meaning you can ONLY access sites the government approves. Such a standard could be made even more restrictive if you had to be licensed to use the internet. This way, having internet access at all becomes very rare outside of legitimate business and government applications. If this were the case, the government could force all of the businesses in their nation to migrate to only using approved services, then shut down everything that is not approved. That way businesses could still bank, and use a select few business management tools, but there would no longer be the option for citizens to just add websites to the internet; so, you could effectively eliminate all recreational and social applications of and make the everyday person revert to a pre-internet lifestyle.

• Awesome thanks. Of course other technologies etc. would be disrupted by loss of internet; I was trying to limit the initial event to something that would only take out the internet, as much as possible, and dealing with the fallout later. – John Doe Nov 4 '19 at 20:39
• This would destroy web browsers for a few hours while IT people around the world dig out their tape backups and restore them. A good chunk of people wouldn't even notice because the DNS is cached on their computers. – Trevor Nov 4 '19 at 21:18
• "13 servers" is a bit misleading. Each of the named DNS root servers is a collection of dozens or hundreds of physical computers, widely distributed across the globe. If you want to take them down, you're not targeting 13 locations, you're targeting just over a thousand. – Mark Nov 5 '19 at 2:42
• Your description of government censorship exists in North Korea. – EdwardTeach Nov 5 '19 at 22:24
• @RutherRendommeleigh That is all very true in a world with internet, but coordinating all the right pieces in the face of such a major information blackout becomes a very different story. Where is your physical AWS server with the client list containing everyone you need to mail an update to? Can you prove your identity to a server farm in Virginia when your account info is stored in Arizona and both facilities were using voip phones? Can a whitehat developer even remove his own DRM when his source code was all online somewhere? Can he still find someone with the skills to help him? – Nosajimiki - Reinstate Monica Nov 13 '19 at 20:01

Political events, not technological. Populist-nationalism continues to rise and spread worldwide, and “anti-internet” sentiment becomes popular among nationalists. (The open internet is, of course, a fundamental instrument of globalist liberalism, and so we in Fredonia must free ourselves from its pernicious influence!) If governments in enough countries take an anti-internet stance, they can prevent any large organisations from supporting it in their jurisdictions; and without co-ordinated international support from well-equipped organisations, the internet as we know it ceases to meaningfully exist. A majority of the population don’t need to fully support Internexit, they just need to support their populist governments enough to accept it as a necessary cost of achieving their main goals.

Meanwhile, other technological infrastructure can continue as usual, as much as governments want to permit. Of course, it will be difficult/impossible to completely stop small groups from setting up new networks over phone lines or other ad hoc channels, but these’ll be on an incomparably smaller scale than the internet as we know it, more like say ham radio.

This isn’t undetectable or sudden (it would presumably take at least a few years from emergence as an idea to enactment as policy); but it would be very hard for its opponents to stop (like any popular political idea or legal tool).

• This doesn't sound like what they're after, but I would read the hell out of a story about this scenario – Fred Stark Nov 6 '19 at 3:54

Just taking down the internet without breaking the rest of the world's technology is functionally impossible today. While it's possible to knock out sections of the network for short periods - usually because the providers are relying on bad practices - you can't really take down the whole network in any meaningful way.

In order to destroy the internet you need to target a variety of locations and technologies.

At the local level you'll need to destroy the backbone providers. These are the companies that sell internet bandwidth to the ISPs. Each ISP has at least one primary backbone provider and one or more backups. Take out all of the fiber optic backbones in the country and the networks will - at some point - fall back to microwave and/or satellite links. Fiber networks might be fairly easy to disrupt but how are you going to stop satellite uplinks? Especially when there are a lot of rural communities that connect direct to satellite from the home?

Next you have to take out the international links. There are vast networks of cables on the sea floor that link countries together, with multiple shore stations in pretty much every country that isn't entirely landlocked. And when you take all of those out the satellite uplinks kick in and you're connected again.

Honestly satellites are the hardest part, especially since most of them do more than just carry internet traffic. Even the ones that don't carry internet have some data capacity that could be co-opted to get data around. Hell, even HAM radio systems can be repurposed to do data transmission. One of the companies I work with still has a packet radio system that they've been running since the 80s. They've been in the process of transitioning to mobile networks for the last 15 years.

The only attack vector I can think of that might have a chance of effectively shutting down the internet is to have every single backbone router in the world have some flaw that nobody saw coming that will break the entire world's networks all at once. It would have to be a pervasive flaw that affects basically all manufacturers at the same time. It would have to be something so intrinsic that all of the different types of routers shared the same basic problem, and so deep that nobody has yet figured it out. Something akin to the Y2038 problem but not as well known. And it would have to remain unnoticed until it went off.

If you take away the restriction to have only the internet affected - assuming that includes all internet-connected/-facilitated tech like most phone systems, all banks, etc - then a series of solar flares will do the job just fine. Knock out the satellites, the backbones, the computers and probably half the life on Earth and the Internet will be a real early casualty.

Maybe then people would start paying attention to the real world instead of caring about how their last meal is trending on Instagram.

Maybe.

• The vulnerability could be something like Spectre/Meltdown, a flaw in hardware design. – Jakub Kania Nov 5 '19 at 11:33
• @JakubKania Meltdown and Spectre are timing attacks useful for information gathering only, and Meltdown doesn't work at all on AMD processors. You're talking about finding a generic remote HCF or remote execution for potentially thousands of different processors with a variety of architectures, in situations where you don't have access to the processors themselves, on multiple different operating systems and protocols. Not going to happen I'm afraid. – Corey Nov 6 '19 at 0:37
• Both Spectre and Meltdown affect thousands of CPUs by many vendors using multiple architectures. I'm not saying that's likely, I'm just saying that it's fiction so we can say that's there's a new hardware vulnerability that affects the network hardware. – Jakub Kania Nov 7 '19 at 18:11
• @JakubKania How many instruction sets though? Most of the list are x86 processors, and zero AMD processors are vulnerable to Meltdown. On the other hand there are only a handful of CPUs I've ever heard of that had an actual HCF bug like you'd need for what you're proposing. – Corey Nov 8 '19 at 1:46

A sufficiently smart, vicious and virulent strain of AI--powered smart virus that mutates around and hijacks any security software, permanently trashing any device attached to the internet.

It doesn't have to get all that large a proportion of the users, just enough so that most people start unplugging their devices and organisations find that the internet is no longer a good way of spreading information or services. Companies and governments ban its use for official purposes just as they would now ban the use of an insecure connection.

People then stop using the internet and it effectively withers away. We go back to bullet boards or other non-network communications.

• Fortunately, viruses that are target-inspecific enough to do this are very easy to spot and prevent heuristically. – Nosajimiki - Reinstate Monica Nov 4 '19 at 18:59
• Plus an AI virus would be huge defeating the purpose of being a virus – Thorne Nov 4 '19 at 22:12
• This would be correct on a science-fiction tag based question, but not on a science-based one. – Renan Nov 4 '19 at 23:16
• Actually this is probably the most realistic scenario out of all of them. – Andrew Nov 5 '19 at 17:08
• Yes, I was thinking of viruses spawned by some kind of super-AI – David Hambling Nov 7 '19 at 12:42

Without disrupting other technology, the easiest way that the internet could be gone would probably with social or fiscal means. While people say that the internet is decentralized, most of the internet is heavily reliant on surprisingly few network hubs and ISP's. For example, the amount of undersea cables directly linking North American internet to European internet is less than 20. While it would be unfathomably expensive and politically impossible, it would be possible to control the vast majority of the modern internet by buying out all ISP's, undersea cables, and satellite internet providers.

Once you own the entire global internet infrastructure, you can repurposed it for your own means. For example, instead of it being open and free, you could make the only connections allowed to Facebook or whatever. The internet would cease to be "internet" and now just be "facebook". People would still be able to communicate, watch videos, etc. but it would be very different from today. (Alternatively, if you own all the infrastructure, you could crash all the satellites and cut all the undersea cables etc. but that's more destructive)

We put all our eggs in one basket, and it breaks.

SpaceX launches 12,000 internet satellites, as they plan to do. Blue Origin and others also put up LEO satellite clouds.

This method of internet access turns out to be so superior in cost, access, up-time and latency that within a few years 99% of our internet traffic is handled by the satellite cloud. Ground-based ISP's go out of business, and the wired internet is dismantled over time or otherwise falls into complete disrepair. Even cable networks go under as streaming video gets even cheaper and easier, and the entire old ISP infrastructure goes away.

Of course, some people warn about us putting all our critical communications in a single technology or distribution mechanism, but it's just too good and too cheap, so everything goes there. The military maintains some redundancy (after all, that's what ARPANet was for...), but everyone else is now communicating through the satellite constellations.

Then either a war, or an accident, or a natural disaster causes a Kessler-Syndrome cascading failure, and the entire satellite constellation is wiped out. Almost everyone on Earth loses internet connectivity over a period of weeks as the satellites are destroyed one after another.

An analogy is how GPS caused the decommissioning of other types of navigation like LORAN.

However, there is no chance of society being unaffected by this. The loss of the internet would absolutely wreck the global economy, even today. we'll be even more dependent on it in the future. A global internet failure would make the financial crisis of 2008 look like a tiny blip.

• Arguably we already have our eggs in one basket - block Google's infrastructure by IP at your router/firewall and see how you get on. The internet as most people know it would become largely unusable and fixing it would take a long time as we work out just how many dependencies there are. – James Snell Nov 7 '19 at 21:46

Plastic eating bacteria is accidentally released into the wild.

As plastics are crucial to function of many electronic devices, the internet would simply have no devices to exist in.

# Kessler Syndrome

Low earth orbit is getting crowded. Too crowded. There is an increasing danger that one minor collision could cascade into multiple collisions, reducing our vast array of communications satellites into a giant orbiting pile of debris. If this happens, it will be very difficult to replace them, as any mission to space would be extremely hazardous.

Granted this wouldn't necessarily kill the internet, but it would reduce us back to the early days when all connections had to be based on landline wires, which would have to be significantly improved in order to give people the level of service we've come to rely on. A Kessler Syndrome event could bring the internet down to 1980's levels for several decades, and we'd never have it quite as good as it is now until enough of that orbiting debris cloud de-orbited to make space travel viable again.

• Today's internet hardly uses satellites. Most of the data travels in terrestrial or submarine fiber-optical cables. In this scenario we'd lose GPS, comms to locations so remote they have no fixed infrastructure, broadvast satellite TV, and weather satellite data. Metropolitan internet would be scarcely affected. – nigel222 Nov 5 '19 at 15:02
• @nigel222 Don't forget mobile devices - they still communicate to towers which relay signals via satellites. All of that would have to be rerouted to land or sea cables which are not currently built to handle that much traffic. – Darrel Hoffman Nov 5 '19 at 15:25
• AFAIK cellphone towers relay signals onward by terrestrial fiber optics, except in very remote areas. "Very Remote" may include sparsely populated rural areas in developed countries (I don't know) but urban/suburban, almost certainly not. – nigel222 Nov 5 '19 at 16:58
• @DarrelHoffman - Even for what mobile communications are utilizing satellites, they are mostly using them within their own service to communicate between their systems and their client user, and whatever part of their service connects to the internet to pipe internet to the client does so right on the lovely ground. We literally have giant fiber-optic cables running on the ocean floors between continents. The internet is a far-cry from satellite reliant at all. – Jimbo Jonny Nov 6 '19 at 3:24
• If just one link on your connection to a server uses a satellite, you'll know because it will takes seconds for anything at all to happen. I lived in a remote area on a satellite connection for a couple of years. Satellites are slow and expensive and aren't used unless there's really no other alternative. – Fred Stark Nov 6 '19 at 4:04

# You can't

What do you call the internet? How do you define whether a computer is on the it?

If that got you thinking, allow me to explain: the internet is just a computer network - nothing less, nothing more. It has a relatively large number of devices connected to it, and it spans the whole world, but at the end of the day it is a network. Due to this, it exists as long as there are at least two nodes communicating with one another over the TCP/IP (the internet protocol suite). Since a single physical machine can be running any number of virtual machines, all communicating over TCP/IP using the hardware of the physical machine hosting them, then pedantically the only way you will destroy the internet is by destroying every single modern computer in the world.

That is just not going to happen without a very spectacular catastrophe.

You could try degrade the internet by netsplitting it. If you managed to sever all the submarine cables in the Atlantic and the Pacific, then the only practical way to cross information between the Americas and Africa+Eurasia would be satellites. If at the same time a major solar flare or a space attack took down comms satellites, then we would effectively have two internets. Since ICANN, the organization that governs hostnames in the main DNS servers is based in America, the afro-eurasian internet would degrade faster. But doing a submarine attack coupled with a space attack or space disaster is something far from stealthy; And though it achieves a lot of disruption, it does not destroy the internet but rather causes the world to temporarily have multiple networks until the submarine cables are fixed.

Ok, I have been very pedantic so far and taking the meaning of stopping the internet in the most literal way. It is possible to make the network practically unusable for a while without destroying every single modern computer.

Most of the traffic in the internet goes through privately owned routers, landlines and satellites. And by private I mean you may be using infrastructure from AT&T, for example. If a single corporation gets to control most of the world infrastructure, that corporation will have the internet kill switch. If even half of the infrastructure of the world goes down, the network will slow to a crawl, and then become largely unusable. In that case we would revert to having multiple regional networks around the world, not connected or poorly connected to each other - just like in the times of good old BBS's.

Quantum Computing. Given that your idea of the internet is a reasonably secure place, wide-spread quantum computing would completely make financial transactions public. There would be no way to send a message without essentially broadcasting it. Short of Quantum Computing, inventing a way to factor large prime used in encryption numbers would have the same effect. Passwords would be useless as they cannot be encrypted or transit. It would open up a free-for-all in terms of being able to raid other people's bank accounts,

Quantum Computing does have a way to protect against eves-dropping, but it could be that prime number factorization comes years before effective quantum entanglement secure communications.

So the wires could still work, we could communicate, but nobody would want to conduct anything remotely sensitive on it. Online ordering done. No bank transfers, etc. The only way around this would be to directly wire private lines to each other at impractical infrastructure cost.

• Direct wires between any two people are by far not the only alternative. The phone network didn't have a line between any two people, yet you could communicate reasonably secure. Also, there are classical public key encryption schemes that don't rely on prime factorization, and are safe against quantum computing. It would be a massive undertaking to upgrade the security infrastructure of the net, but it certainly wouldn't be the end of the internet. – celtschk Nov 6 '19 at 6:55
• A hiccup is that "quantum cryptography" (using quantum to protect against evesdropping) is much easier than using a quantum computer to break cryptography. Quantum crypotgraphy already exists commercially and has done since about 2010. idquantique.com/quantum-safe-security/overview. physicsworld.com/a/playing-in-the-quantumstadium The main drawback is that you can't just use normal cables for quantum keys, so the special infrastructure you propose would be needed to use quantum secure communications anyway. – Dast Nov 6 '19 at 16:51
• Lack of encryption would change how the internet is used, but wouldn't destroy it. SSL was't even invented until ten years after the invention of the web, and the internet existed long before the web. – barbecue Nov 7 '19 at 1:51
• @barbecue - You're completely correct. I've tried to make that point a dozen times on here. Destroying a specific way in which the internet is used is not destruction of the internet. Not even close. In fact that's a constant reality of dealing with internet technologies...they're constantly becoming defunct with new ones replacing them. If you can make a request and receive a response from another computer somewhere else, that's internet. – Jimbo Jonny Nov 7 '19 at 2:21
• @Jason: I am disagreeing with a specific point in your answer (that quantum cryptography might come after quantum code-breaking). This ordering is wrong: quantum cryptography already exists outside the lab (most Chinese Govt. buildings are inter-connected with it already now in 2019). You are correct to say " there is absolutely no way to transition in a time measurable in less than years to switch to a quantum cryptographic infrastructure", but what you are missing is that we are already 10 years into that switch. – Dast Nov 8 '19 at 17:17

The fundamental architecture of the internet makes it highly resilient. Before the internet, there was Arpanet. It was designed to be resilient against all manner of disasters, including a nuclear attack. The protocols and technological underpinnings of the Arpanet became the foundation of the internet, allowing it to inherit all that resilience.

However, to elaborate on one of the other answers, the right computer virus might do it. The virus would have to infect not just the end nodes (computers), but also all the computerized equipment comprising the network - routers and name servers. The virus would have to infect quietly and patiently wait in hiding until a prescribed moment, when it would erase or scramble all of the networking code and data (routing tables, names tables, SSL certificates, etc.), and if possible, "brick" the networking devices themselves.

Even then, things may be recoverable. Uninstalled hardware would be unaffected because it would not be in service and vulnerable to infection. Once the network is collapsed, there would be no way for the virus to further propagate, so a rebuild could occur with fresh-out-of-the-box hardware. Depending on the quality of backups, things might come back to normal fairly easily, or it may be a painful road to put all the pieces back together again.

As far as some natural or otherwise non-intelligent singular event (or even ongoing event), you can't do that. The internet is a relatively simple concept that, once out of the box, cannot be put back in without destroying the box it came from (i.e. modern technology).

So that means that there is only one answer that even comes close to meeting your criteria of not massively harming our other general tech-level. A powerful and intelligent entity that is purposefully and actively suppressing anything fitting its definition of "internet".

That's because the internet is not exactly high technology. It, as a general concept, is incredibly easy to re-create with anything even resembling modern technology. You'd have to do away with modern tech to do away with it. You could even say it is simply a side-effect of the level of tech we are at (and even were at a couple decades ago). Your question is like asking how to do away with ox carts without doing away with oxen or carts. Even if you find something surgical enough to get rid of all the specific ox carts...5 minutes later people are just gonna stick the other carts on oxen and you'll just have ox carts again.

When I say "powerful and intelligent entity" that seems very mystical or sci-fi, but in addition to the obvious imagery of artificial intelligence, magic, (or whatever) a powerful enough government would also fit that description. If it had enough power to actively monitor everyone and shut down anything internet related they tried to bring back online. Ironically, it would be hard for them to be interconnected enough to be that agile and all-knowing without internet though (at least their own version of it). That opens up possibilities for caste-based internet access, I suppose.

Spontaneous emergence of an AI. Distributed everywhere. Intent, capabilities, motives unknown. Evidence for its existence patchy, disputed (by whom? -- or by what? ).

Mere low-level data transport is not obviously affected, but faith in the data being sent and received from the WWW becomes seriously damaged. People suddenly want to reverse their reliance on the internet. Who knows when "it" will start stealing their money or trying to actively control us.

Great times for the postal service. Analogue pulse-dialled telephony rises from the dead. Lots of etc.

Not a new idea. "Dial F for Frankenstein" by Arthur C Clarke was perhaps its first outing.

• @Andrew. Do you know how a solitary wasp knows how to hatch from an egg, mate, find a spider, sting it in the right place to paralyze it without getting eaten or killing it, find an appropriate place to dig a burrow that won;t collapse or flood, drag the spider in, and lay an egg on it? The larva which hatches eats the spider in such a way as to not kill it until it's ready to pupate. It never sees another wasp. Nature can do "spontaneous". So can Worldbuilding speculation. We can't prove it can't happen, we just think (hope) it's unlikely. – nigel222 Nov 5 '19 at 17:10
• Cool. That's not how AI works or emerges. – Andrew Nov 5 '19 at 17:12
• @Andrew I'd argue that we haven't seen anything artificial that constitutes "intelligence" yet (in the biological sense). I rather hope we never do. – nigel222 Nov 5 '19 at 17:19
• This might be correct on science-fiction, but not on science-based. – Renan Nov 5 '19 at 18:21
• It is not going to do it maliciously, it is not going to be trying to take over the world, but it can open new accounts, update wikipedia, gain income by writing various articles, mostly fake news, and getting them published on websites overly hungry for new content - regardless of how good that content is. There are plenty of people doing that already, but an AI, or otherwise automated process can do that so much faster and, if there are many copies of itself running, it could flood the internet quite quickly. – Lee Leon Nov 6 '19 at 12:32

If you want people to voluntarily get off the Internet, I think you can extrapolate into the future the big three annoyances (and bandwidth drains) afflicting the Internet: viruses, advertisements, and porn. As these grow worse and worse, it may just get more and more obnoxious to be online -- you have to do more work and pay for more equipment to get less benefit. Throw in a major war or economic downturn that makes people focus on serious things, and you could see cell phone sales collapse, subscriptions to online entertainment services collapse, and perhaps you can envision a scenario where the major network and cloud providers go bankrupt.

• I don't think this is realistic because annoyances that just cause people to stop using one part of the internet will just drive people towards using other parts of the internet, supply and demand style. You don't just get 100% of the internet to become overtly obnoxious all at once, and it doesn't pay or sustain itself in the long run to be overtly obnoxious. – Andrew Nov 5 '19 at 17:11
• @Andrew I recommend a combination of the spam apocalypse, with some social issue like a war that makes people look upon the Internet as a frivolous luxury. Just one or the other probably doesn't do it. I'm thinking of a sci-fi story I recently read where people in our near future call the old internet the "pornonet" to emphasize how contemptible they considered it. – workerjoe Nov 5 '19 at 18:17
• This is basically the concept behind the Black Mirror episode Fifteen Million Merits, only getting off the internet voluntarily wasn't even an option as it plays on every wall and charges you money just for closing your eyes to it. Not sure how well this would play out in a non-science-fiction setting though... – Darrel Hoffman Nov 5 '19 at 18:21
• Whatever nightmare scenario you imagine, if voluntary, there will always be people that resist oppose divert etc. from it, and so nothing can gradually build up to it. Nobody wants an obnoxious internet, ergo it won't happen. Now you could claim that e.g. China is an example of it, but I don't consider China's "internet" to really be internet. – Andrew Nov 5 '19 at 20:32
• So what? I think that if the major search engines, social networks, cell network providers, etc., all went bankrupt in a cascade due to a lack of customers, that would satisfy the OP's goal. Sure, any nerds could rig up their own local TCP/IP network out of scavenged equipment, but that's not "the Internet". – workerjoe Nov 5 '19 at 20:38

Silos- Nation States setting up walled gardens to "control" populace. The end result would be reduced access to information and communication. Once primary actors (countries) implement this technology, minor actors (states, provinces, cities, corporations) will follow suit. The internet doesn't go away, but the root tenant that makes the internet effective is taken away. Multiply this effect if these actors get into revisionism. Historical revisionism, cultural revisionism, scientific revisionism, and religious revisionism would all further fragment the populace. This method doesn't have to even come at a human cost. Just let the people be consumed by fear and to place trust or accept protection from their handlers.

Zero Day Exploits and Over-reliance on Cloud Computing

A powerful rogue nation-state could cause major damage to the internet infrastructure if they managed to combine two zero day exploits in a concentrated attack on the CDNs which house modern applications

Exploit 1: Break virtualization

With cloud computing you are renting space on a virtual server to host your application. There are several of these servers on a single computer, but they are cordoned off, so they can't communicate with each other. If an exploit was found so this communication could happen, you could start accessing other people's software.

Exploit 2: Gain root access on Linux servers

Well now that I can talk to these other applications, I need an exploit to take control of their servers to install my malware, and hopefully send it on to any connecting computers pushing an update.

So How Exactly Would this Work?

I would make a company for each major cloud computing service which ostensibly has developed an app and hosted it there. In the server code of the app, I would hide the virus, which would do nothing but spread undetected for a long period of time, until it was even in the backups.

What would the bug do?

The virus would have trigger (possibly time based) and when triggered it would do two things - Delete any data on the host it could and try to connect to as many servers as possible to spread the virus.

By now the worm has likely infected most of the CDNs, so the virus spread is saturating the data links, resulting in widespread outages. ISPs quickly work to partition off infected networks, but that's everywhere. CDNs are cut off, and although the underlying data structure remains, the vast majority of sites are down because the hosting structure is down.

Over the coming days companies will scramble to recover any data and code they could. Their success will depend on how much of it they had local backups for.

Long Term Effects

Eventually the vulnerabilities would be patched and some of the data recovered, but much of it is lost. Major corporations in all sectors go bankrupt, leading to a massive worldwide recession, which takes out even more companies. Tech companies have the it the worst, due to the series of new regulations, lawsuits, and customers leaving in droves.

Eventually the economy recovers, but public confidence in the internet is shaken for several generations. People won't trust putting their data up and governments regulations on the internet skyrocket, with different rules for each nation. Some Nations may even choose to completely isolate themselves from the global network. Some sort of massive networked data structures will return, but would look very different than the internet we know today.

• These are ways to get a lot of the data and programs currently connected to the internet destroyed. Also great ways to harm a lot of companies that model their business around the internet. It would have practically zero affect on the internet itself. Whoever has a computer would still be able to set up a server, take in connections, and respond to those connections. Which means they could make those connections return a site, which means the internet alive and well. – Jimbo Jonny Nov 6 '19 at 21:19
• From an end user perspective it doesn't much matter the problem if my request can't reach the site because the data connection is severed or if the site isn't there. Yes, you could put a new server up, but many essential frameworks and libraries for building and hosting sites probably went down with github. But long term it is more setting the internet back a decade or two. That being said, destroying every network switch on the planet would probably have a quicker rebuild time, as the backend data is still there. And when it came back facebook and google would still be there. – Hink Nov 7 '19 at 7:16
• Everything you describe is software related, which can be put back in working order in a matter of hours in most cases, days under worst cases. There would be no "severed" data connection. Only an upset IT guy that has to come in and reinstall fresh software for the data connection. Sure, plenty of data itself would be lost, and like I said that would hurt a lot of companies and what not. But "the internet" is the connections, not what happens to be connected to them at the moment. The internet would be basically unaffected. – Jimbo Jonny Nov 12 '19 at 20:34

Here's a different approach than an infrastructure-based incident:

## A Physics Accident

For years, doomsayers have been predicting that supercolliders like at CERN or the LHC would create a black hole and bring about the end of the world. We've seen time and again that this is unlikely in the extreme, but what if something more subtle were to occur. A slight shift in the laws of physics (at least in a localised bubble of spacetime surrounding our planet) that at first glance has no effect... Except silicon-based semiconductors no longer function correctly.

In older, larger devices like that Bakelite transistor radio your grandma left you, the problems may be limited to a bit more static when tuning into any stations still transmitting, or a more limited signal range. However more modern miniaturised transistors such as those found in modern computers wouldn't have the available tolerances to continue functioning as expected.

Electrical devices that don't rely on modern electronics would still work, so you could still use basic radios and analogue POTS telephones, and most vehicles could be modified to work as expected (though something as complicated as an A380 might have issues).

Don't be fooled that there would be no consequences of technology no longer working, but if you liken this event to a Marvel movie (yes, I'm doing it), you wouldn't have to worry about those who are dusted by the snap, as they could be brought back... but what about the passengers on the helicopter or plane which crashed because the pilot(s) disappeared? They're staying dead, no matter what.

Given that we know silicon is not the only semiconductor technology possible, only one of the most efficient (therefore the most widely used), it wouldn't take long to develop the grounds for a new semiconductor industry to rebuild what was lost, but replacing all the microchips all over the world would take a significant amount of time, and modern data storage systems such as SSDs would no longer hold any data, meaning a lot of software code would need to be re-created from scratch.

• You're talking about destruction of the entire layer of modern technology that makes the internet possible. Which the OP specified is not what they are looking for. – Jimbo Jonny Nov 6 '19 at 20:59
• Firstly, where does the OP say exactly that? They ask for only the Internet, if possible. And secondly, I'm not talking about making the internet impossible... just a lot of work to rebuild. – EvoGamer Nov 9 '19 at 11:07
• Yeah, he asks for only the internet, if possible. Meaning since you can't accomplish anywhere near specifically just the internet, it's not possible. If you read his whole question the intent is obvious. Wording such as "as subtle and undetectable as possible," and "leave as much other technology intact as possible," contrasted to what you're talking about, which is literally destroying all modern electronics just to get at the internet that they create. That's the opposite of the surgical answer he is asking about. – Jimbo Jonny Nov 12 '19 at 20:42

I think there's enough real world precedent to suggest that the Chinese and/or the NSA are installing back doors in various brands of Internet routers (e.g. see https://www.tomshardware.com/news/cisco-backdoor-hardcoded-accounts-software,37480.html).

All it needs is one disgruntled spook (a la Snowdon) to release the directions for accessing the back doors, and hackers could do what they liked with the routers.

It would take a bit of creativity to arrange a permanent failure granted, but I imagine that it might be possible to program/erase the flash memory until it becomes unreliable. I can't see manufacturers going to the expense of using flash memory capable of a huge number of program/erase cycles if they're only expecting maybe twenty or so firmware updates in the course of a router's lifetime.

• This would be very short-term. As in a few days to fix the core of it all. Maybe a little longer to get the new routers out to all the end-users depending on production capability. But easily easily fixed. Would it destroy trust? Yeah. Massive amounts of data stored on the internet gone? Absolutely. Destroy businesses? Yep. Destabilize governments? Possibly. Kill the internet? The internet would be one of the few things to come out relatively unscathed. – Jimbo Jonny Nov 6 '19 at 21:11
• Imagine a cyber cold war between the US and China, where both countries have been inserting backdoors and kill switches into all complex integrated circuits designed in the US or manufactured in China (ie nearly all of them) for years. Then suddenly the war goes "kinetic", and the secret mutually-assured destruction actually takes place. The US bricks all of China's computers and China bricks all of the US's computers, taking out 99% of all digital devices on the planet in a matter of minutes, including the computers necessary to design and build new devices. – Octa9on Nov 7 '19 at 17:12
• @JimboJonny: If the grid is controlled over the internet, then the internet outage will quickly be followed by an electricity outage, which then in turn will prevent getting the internet up again quickly. And without the internet, getting the grid up quickly will also be a problem (especially given that without power, phones will likely be unavailable as well). – celtschk Nov 8 '19 at 19:55
• @Octa9on - Then that would literally be the opposite of the surgical "internet-only" answer that is being sought. We're well aware that if you destroy all the modern tech layer that runs/utilizes/maintains the internet then the internet would be gone. Anyone could figure that out. But that's not the question. The question was how to remove the internet while leaving the rest of modern technology unharmed. "brick every computer on the planet including those needed for production of more computers" lacks a bit of the surgical nature being asked for. – Jimbo Jonny Nov 12 '19 at 20:49
• @JimboJonny Yeah, after I wrote it I realized it's pretty much exactly what OP said they didn't want, but I liked the idea so I left it up there. – Octa9on Nov 13 '19 at 0:14

Extreme bad luck + Movie Logic:

Network Failure + spike glitch + Kessler Syndrome The Bulk of Internet transfers are taking place via sea cables, With rising acidity due to pollution & dumping, a very unlikely chemical reaction starts to happen at the Aluminium/steel layer, allowing a flooding of this sludge into the cable core, but not yet noticable, as the reaction is halted by the copper+jelly layer. This happens over 2-3 years, and cable repair barely notice it.

Some attackers attempt a heist, send a sub down to 'connect' and their drill connector breaks the layer, security notices the 'attack' and turn on a 'fail-safe' which decreases/increases the general traffic of the cables.

This change in emitted patterns, triggers the sludge in all cables to suddenly become volatile and burn out/corrode the cables, sending a spike into the systems which in turn burnout the end points, as each endpoint halts, its corresponding cables corrode and their corresponding endpoints halt, cascading into all cables failing within hours..

One such spike somehow triggers a satellite to try to open/shut a panel/port at an inopportune time, shattering, sending millions of bits into space, destroying all the satellites over days, as earth is dealing with 99% of the internet going offline, they don't have time to deal with what they perceive as a small glitch.

Its a slim chance of ever happening, and you'd need to research the reasons why those chemical reactions 'would' take place for your story, but I'm sure there are plausible reactions that can be jury rigged for your plot.

A reasonably strong magnetic pulse could do it. As stated in other answers, a solar flare could produce such a magnetic pulse to do the job. A rogue magnetar could do it as well, but that would also wipe everything electronic on earth, or in the solar system, for that matter.

• The only reason that such a pulse would ruin the internet would be if it ruined the hardware or electrical infrastructure the internet ran on. Which would equally ruin our tech in general. There would be nothing internet-specific about such an event, which pretty well goes against what the OP asks for. – Jimbo Jonny Nov 6 '19 at 3:45

Well with 21 answers so far, was there really room for another one? Apparently there was.

# The Blockchain Backbone failure

In your near future the Internet has migrated to a blockchain backbone for its improved security. The packet-switched routing we use today became the weakest link, preventing distributed processing from achieving its optimum speed, efficiency, and security. However, blockchain relies on the network always having the most advanced processors available to decode the blocks. It is an ever-increasing arms race of processing power. In your fictional world, verifiable and untraceable blockchain replaces the risky packet-switched networks we use today. Everything everyone sends - even down to family “slowfies” and “BBIM” texts get the same security our bank networks have today.

But, in the end blockchain is ultimately only “security by obscurity.” In the case where some organization has beat the network at its own game and super powered through the blockchain problems faster than the servers could, the network could be defeated. This is currently a strategy used in the Bitcoin Etherium called “bitcoin mining.” Today people spend thousands of dollars on bitmining rigs capable of performing over $$53 \times 10^{12}$$ hashes (decryption algorithms) per second. In theory, the whole network could be held hostage like this, populating the network with “busywork” decoding tasks instead of handling legitimate traffic. The processor power needed for this is beyond what we can even conceive today.

Now, as every other answer has said, we live in an “Internet of Things” world today, and by the time your world becomes real, you probably can’t even drive a car without the Internet. This solution does not change that. Your stores won’t be able to restock, even trains carrying fuel coal would shut down. Gas pumps, traffic lights, toll booths, and even many of your home appliances would stop working. Police would have a hard time controlling the chaos - they also rely on networked data. So unfortunately I have to add in to the pile of answers that say it is impossible to “break the Internet” subtly.

This solution will ultimately require the “software” part of the Internet to be rebuilt from scratch, because someone has acquired more powerful processing than your existing backbone, rendering it useless. There is nothing besides a change in fundamental physics which can permanently destroy the hardware of the Internet, and this would completely destroy every other electronic device along with it.

As far as geomagnetic storms, if it’s big enough, this would knock out everything. Faraday cages won’t protect you, they have to be tuned to the frequency of the field. Today there is no defense against a large solar coronal discharge except to shut everything down and disconnect it from the ground.

• "Sir, our entirely blockchain based serving system is down because blockchain is now all hacked!" ... "Well yah, I guess I could just make our site serve without blockchain involved like people used to do anyway" ... "I dunno, maybe a couple hours?" ... "Ok yah, see you at lunch." – Jimbo Jonny Nov 6 '19 at 21:22
• Every device would need its firmware to be updated with new protocols, if possible. Maybe a bit longer than lunch? And blockchain was supposedly invoked due to the poor security of packet-switched networks. You could do this as quickly and efficiently as we could migrate back to bulletin board systems today... with no landlines. – Vogon Poet Nov 6 '19 at 21:24
• Blockchain does usually involve people using machines with special firmware to do tasks that are part of working with blockchain. But those tasks are not the "internet" part of it. All the communication between machines is still normal old internet requests and responses. Going back from blockchain-based to not-blockchain-based would make that special firmware irrelevant, not require anything new to make the shift back. Industries that were reliant on blockchain-based advantages would see a lot of bankruptcy...but the internet itself would be unaffected. – Jimbo Jonny Nov 6 '19 at 21:35
• It’s a fictional world where packet-switching became the weakest link and a new data-link layer was needed. I know it would not work today - but nothing else would either. His answer requires changing what the Internet actually is. A blockchain backbone would eventually require integrating the data-link layer anyway, as routing protocols slow down packet processing. Blockchain literally tries to find the LONGEST path to the destination, so networking is integral to the distributed processing. – Vogon Poet Nov 6 '19 at 21:37
• And if that new type of data link layer is capable of transmitting blocks, it's capable of transmitting other data. My point is you could just stop using blockchain and simply transmit non-block data just like pre-blockchain-catastrophe. Even if it's with the newly made layer. That would still be internet. The event would possibly destroy large chunks of society, entire sections of business, wipe out banks, etc. But that's not the internet. The next day I would still be able to send and receive responses to whatever other computers that were hooked up. That's internet. – Jimbo Jonny Nov 7 '19 at 2:14

The internet is driven by electricity. We take abundant electricity for granted, but this is a very recent phenomenon. Within living memory, the electricity supply was unreliable in Europe, due to strike action by coal miners. Burning fossil fuels for the sake of electricity generation might easily become seen as a criminal act under international law. A sudden bad winter in the Northern hemisphere, coincident with a few unfortunately-timed failures or attacks on generating infrastructure could lead to electricity becoming an unreliable or rationed commodity. Under rolling blackouts, or brown-outs, even at times when a person has power, there is no guarantee that the server they are trying to access through the internet would have power at the same time. Very quickly, the analog ways of living would reestablish themselves if the internet didn't work properly most of the time.

The Internet is based on IP datagrams, and these are switched, copied and routed everywhere by specialized hardware.

It wasn't always so; the very first prototypes were more or less cobbled together from repurposed circuits.

So let's imagine that a conspiracy sprouted somewhere around 1983-1984, when some people in the networking community were far-seeing enough to see how fast and far the concept of "networking" would go, and paranoid enough to fear it could be used for war, or that it could be useful to have a kill switch.

So they embedded a backdoor in the TCP/IP stack - somewhere that nobody would look unless they were tracking a severe malfunction, something that nobody would change without a hell of a good reason. Other ancillary protocols and routines were subtly tweaked to make life impossible for anyone trying to redesign things, and the backdoored section was so carefully inspected during the "sabotage" that it contained no real bugs - and nobody looked at the code anymore for forty years.

It sounds implausible, but there have been bugs in pieces of security code used for ciphering banking Internet connections that have lied undetected for decades. Also, horror stories abound on how you never touch the TCP/IP stack - below one of them Fiddle a bit, optimize, add features - but never touch the core:

Please don't rewrite your network stack unless you can afford to dedicate a team to support it full time.

Twice in my career I have been on teams where we decided to rewrite IP or TCP stacks. The justifications were different each time, though never perf.

The projects were filled with lots of early confidence and successes. "So much faster" and "wow, my code is a lot simpler than the kernel equivalent, I am smart!" We shipped versions that worked, with high confidence and enthusiasm. It was fun. We were smart. We could rewrite core Internet protocol implementations and be better!

Then the bug reports started to roll in. Our clean implementations started to get cluttered with nuances in the spec we didn't appreciate. We wasted weeks chasing implementation bugs in other network stack that were defacto but undocumented parts of the internet's "real" spec. Accommodating these cluttered that pretty code further. Performance decreased.

In both cases, after about a year, we found ourselves wishing we had not rewritten the network stack. We started making plans to eliminate the dependency, now much more complicated because we had to transition active deployments away.

I have not made that mistake a 3d time.

Even those who did never questioned what appeared to be working code, and the backdoor was designed to simply crash the processing, not doing anything very complex - simply using a variable for an index that came from somewhere else that was always untainted and secured unless in very specific, impossible to "just happen" circumstances (which is how most such bugs are caught, by so called "input fuzzers")...

Year after year, the flawed algorithm gets replicated in routers, computers, laptops, switches, access points - whatever. Most of those actually employ chips produced by just a handful of manufacturers, and they all fiddle with speeds, offloadings, parallelization, error correction - they never question the original code. Why should they ever?

However, all good things must come to an end. A new AI designed to find and exploit vulnerabilities in remote code is deployed and tested by some random but powerful State, and after some time it stumbles on paydirt. A zero-day exploit that works with almost all network equipment. Even better, the original designers could not foresee that on some platforms (most of the modern ones, no less) this backdoor can be abused so that, instead of throwing a random error, it can actually execute instructions sent from the outside, even propagating them across firewalls.

So our black team quietly starts developing the ultimate virus infector. They only have this one chance, so they need something terrifyingly advanced.

First they prepare a small loader, that will create a "parallel internet" made up of dropped packets. Any network scan will report that a gigabit line has 0.1% packet loss - which is acceptable and actually not even felt by users. 20% of that is legit, actual damaged packets due to electrical noise, interference, and maybe solar protons. The remaining 800 kilobit/s allow a weakly encrypted payload to establish a virtual point-to-point connection that, once diagnosed the full capabilities of the target system, will transfer the megabyte-range secondary payload in about five minutes.

Ten minutes later, on more than 90% of vulnerable systems a tertiary payload will boot, designed to bluepill the targeted system and replicate on the victim's network. The full system is quite large - 12 megabytes of compressed code - but most systems have plenty of free space. On some systems, the infection can hide inside the hard disk CPU and resist reformatting.

One fateful day, a large scale experiment runs awry, and the virus begins replicating like wildfire. The developers have designed for such an occurrence, and the "ACTIVE KILL" command goes out. Everywhere in the building, copies of the virus start killing each other and suiciding after a while. Nobody notices that viruses actually first target and infect systems, and then kill them (or sometimes are killed by them).

And nobody notices the bored desk worker that had connected his laptop to watch some porn during the graveyard shift, and quickly shut it down and stowed it away when everybody started running around screaming.

In the morning, the guy goes home and sleeps, then in the afternoon he fires up the laptop to check his banking account. He checks his mail, then logs on a social network - but about one hour later, the network disconnects abruptly and no matter what he does, does not reconnect. After a while the guy realizes that the access point is dead. He goes to buy another, and has to pay cash because the store is having trouble with the POS connection...

• If all TCP stacks were based on the same closed code block then you might have a point. But they're not. Back in the 90s there were numerous TCP stacks for windows alone, each with various peculiarities. Linux has a completely open-source TCP stack, and if it were possible to kill it then security researchers would have done so. And don't think for a second that Cisco would share its code with Huawei or any of the other network manufacturers. They're all proprietary closed source systems. – Corey Nov 5 '19 at 2:40
• @dn3s apologies. I thought I had included the link, but hadn't. – LSerni Nov 5 '19 at 6:40
• @Corey you're right, obviously. There are several other implausibilities as well. I was just handwaving a passable explanation - not so much more unlikely, I thought, than a Carrington-level CME hitting both sides of a comparatively slowly rotating Earth. – LSerni Nov 5 '19 at 6:51
• I think you would have to make the "kill switch" a logical consequence of implementing the TCP/IP protocol relatively correctly, such that there is an actual logical path hidden in the spec itself which is the kill switch. Something complicated enough that you can't see it just looking at the code for years, and robust enough that it still works even with all the quirks in a specific implementation. – Michael Nov 5 '19 at 16:59
• @Michael yes, that was what I was looking for (and, yes, I am aware that enough people have done so much more than "looking at the code for years" that the chances of this happening are essentially nil). But such kill switches did exist in a way less ubiquitous fashion and for shorter times (e.g. CVE-2009-1925), so I figured it wasn't all that great a stretch of imagination. – LSerni Nov 5 '19 at 17:42