Imagine all the roads into and out of your neighborhood are closed. You can still access stuff within your neighborhood. All the stuff outside of your neighborhood also still exists. The route between you and the rest of the stuff is broken, making that stuff temporarily inaccessible.
Now imagine outside of the neighborhood is a store you want to visit. If that store caught on fire and burned down, you wouldn’t be able to shop there anymore, and all the products inside are gone. There are probably plenty of other stores with the same products though, you just have to go elsewhere.
The two things that make the Internet the Internet:
1. A common set of communication protocols – IP/IPv6, TCP/UDP, HTTP, et al, as well as routing protocols such as BGP.
2. A system of unique addressing of network elements and domain names, all administered by various parts of ICANN
You can run a data network on protocols that aren’t IP/IPv6, but it won’t talk to anything else on the internet. You can also do your own IP addressing and domain name scheme – hell, you can create DNS for [google.com](http://google.com) on your own isolated network – but it’s not going to interact with the rest of the Internet.
So, as long as the protocol is stable, and those registries – which are highly distributed to ensure their stability – exist, the internet will never “disappear” in an existential sense.
It’s not fragile from a network perspective. It’s actually super robust.
It’s fragile from a consolidation and regulation stand point.
The more consolidation the less robust and more single points of failure.
For Web, the biggest danger is consolidation and walled gardens as data then becomes available on a pay to play basis.
Cloudflare, crowdstrike, and akami represent real dangers….when these go down do to their enormous reach, huge swaths of the web goes dark.
On the hosting side the danger is AWS, Google, and Microsoft
In general it’s pretty resilient. “The internet” is really just all do the computers in the world connected to each other, but a lot of what you think of as the internet exists in a data center ran by either Microsoft, Amazon, Google, or Oracle. For the most part the buildings are built to withstand a lower level natural disaster. If they lose power, they have backup generators. If an earthquake opened up and the building fell into the center of the earth, most of the data that has already been replicated to another building in another city would start back up and stay running, albeit maybe at a slower speed.
However, we saw on Friday how a software bug can affect the internet in a way that even a natural disaster can’t touch. Microsoft hosts about 25% of the world’s web traffic and Windows is on about 70% of all computers around the world. Crowdstrike, a security company that markets to enterprise businesses, pushed out a security update that had a bug that unintentionally tanked any Windows computer that used Crowdstrike and was able to receive that update. As a result millions of windows computers went down, many websites and applications hosted in a Microsoft data center went offline and airlines, banks, hospitals, grocery stores that are reliant on accessing the internet to handle their business operations went down.
It was fixed in about a day, but imagine if this was performed by a bad group of people, strategically and intentionally over the course of a few months.
The internet is not fragile it is pretty resilient. You have big companies like Google, AWS, azure and other well known data centers have global redundancy. It would take a global catastrophe to wipe out the internet. A common misconception of this recent crowdstrike issue is that it was Microsoft’s fault. The reality CS pushed out an update that took millions of devices offline. If you really think about it monopolized software will likely be the cause for major future disruptions not really internet.
Latest Answers