We all migrate to smaller websites try not to post outside drawing attention just to hide from the “Ai” crawlers. The internet seems dead except for the few pockets we each know existed away from the clankers
We all migrate to smaller websites try not to post outside drawing attention just to hide from the “Ai” crawlers. The internet seems dead except for the few pockets we each know existed away from the clankers
I have a testing website. I have never gave the address to absolutely anyone, ever. It’s not linked with anything. It’s just a silly html site living in a domain.
It’s still being ping and probed to death by bad actors. No necessarily AI scrappers. But it’s dozens or hundreds of http petitions a day for random places all over the world.
There’s no black forest. It’s all light up and under constant attack, every tree is already on fire.
That’s because it’s numerically possible to sweep through the entire IPv4 address range fairly trivially, especially if you do it in parallel with some kind of botnet, proverbially jiggling the digital door handles of every server in the world to see if any of them happen to be unlocked.
One wonders if switching to purely IPv6 will forestall this somewhat, as the number space is multiple orders of magnitude larger. That’s only security through obscurity, though, and it’s certain the bots will still find you eventually. Plus, if you have a doman name the attackers already know where you are — they can just look up your DNS record, which is what DNS records are for.
I like seeing them try and then thinking “begone thot! There is no entry for you”
In fact, I might make a honeypot that issues exactly that
Servers which are meant to be secure usually are configured to not react to pings and do not give out failure responses to unauthenticated requests. This should be viable for a authenticated only walled garden type website op is suggesting, no?
I have suggested a couple of times now that ActivityPub should implement an encryption layer for user authentication of requests and pings. It already has a system for instances vauching for each other. The situation is that users of “walled garden” instances in ActivityPub lack means of interfacing with public facing instances that doesnt leave the network open for scraping. I believe a pivot towards default registered users only content service built on encrypted handshakes, with the ability for servers to opt-in to serving content to unregistered users would make the whole network much more robust and less dependent on third party contingencies like CloudFlare.
Then again, maybe I should just be looking for a different network, I’m sure there are services in the blockchain/cryptosphere that take that approach, I just would rather participate in a network built on commons rather than financialization at it’s core. Where is the protocol doing both hardened network and distributed volunteer instances?
There are several things you could do in that regard, I’m sure. Configure your services to listen only on weird ports, disable ICMP pings, jigger your scripts to return timeouts instead of error messages… Many of which might make your own life difficult, as well.
All of these are also completely counterproductive if you want your hosted service, whatever it is, to be accessible to others. Or maybe not, if you don’t. The point is, the bots don’t have to find every single web service and site with 100% accuracy. The hackers only have to get lucky once and stumble their way into e.g. someone’s unsecured web host where they can push more malware, or a pile of files they can encrypt and demand a ransom, or personal information they can steal, or content they can scrape with their dumb AI, or whatever. But they can keep on trying until the sun burns out basically for free, and you have to stay lucky and under the radar forever.
In my case just to name an example I kind of need my site to be accessible to the public at large if I want to, er, actually make any sales.
But an IP can have multiple websites and even not return anything on plain IP access. How do crawlers find out about domains and unlinked subdomains? Do they even?
@kossa @dual_sport_dork If you’re using HTTPS, which is by and large the norm nowadays, then every domain is going to be trivially discoverable via certificate transparency logs: https://social.cryptography.dog/@ansuz/115592837662781553
thinking about this, wouldn’t the best way to hide a modern websie be something along getting a wildcard domain cert (can be done with LE with DNS challenge), cnaming the wildcard to the root domain and then hosting the website on a random subdomain string ? am I missing something
I do something something like this using wildcard certs with Let’s Encrypt. Except I go one step further because my ISP blocks incoming data on common ports so I end up using an uncommon port as well.
I’m not hosting anything important and I don’t need to always access to it, it’s mostly just for fun for myself.
Accessing my site ends up looking like
https://randomsubdomain.registered-domain-name.com:4444/My logs only ever show my own activity. I’m sure there are downsides to using uncommon ports but I mitigate that by adjusting my personal life to not caring about being connected to my stuff at all times.
I get to have my little hobby in my own corner of the internet without the worry of bots or AI.
Every SSL certificate is publicly logged(you can see these logs e. g. under crt.sh) and you might be able to read DNS records to find new (sub)domains. The modern internet is too focused on being discoverable and transparent to make hiding an entire service(domain + servers) feasible. But things like example.com/dhusvsuahavag8wjwhsusiajaosbsh are entirely unfindable as long as they are not linked to
Random subdomain on wildcard certificate, IP written in the host file to mitigate DNS records, only given by word-to-mouth 😅.
Nobody said the uncrawled dark forest would be comfortable.
I love your “multiple orders of magnitude”. I don’t think you appreciate or realise how much larger ipv6 address space is :)
I wasn’t going to type that many commas for the sake of brevity, but it’s 340,282,366,920,938,463,463,374,607,431,768,211,456 possible addresses. I.e. 2128. So yes, I do.
I consider 96 orders (in binary, anyway) as “multiple.” Wouldn’t you?
No need to be defensive. I’m not insulting, I just find it funny :) usually people call that “dozens”. But dozens of orders of magnitude really doesn’t give the sense of scale.
You could have 8 billion in habitants in every 10^24 stars in the universe and everyone could still have 42k addresses.
I have a DDNS setup. Pretty random site name. Nonetheless, it’s been found and constantly probed. Lots of stuff from Russia, China, a few countries in Africa, and India. A smattering of others, but those are the constant IPs that are probing or attempting logins.
It it had a DNS its obviously lit up like a Christmas tree. If its ipv4 its on shodan. The only way to hide is to have no DNS and an ipv6.
crt.sh and certificate transparency
Do you know how they find it? Is it just random input of address over and over?
Almost certainly. There are only 4,294,967,296 possible IPv4 addresses, i.e. 4.3ish billion, which sounds like a lot but in computer terms really isn’t. You can scan them in parallel, and if you’re an advanced script kiddie you could even exclude ranges that you know belong to unexciting organizations like Google and Microsoft, which are probably not worth spending your time messing with.
If you had a botnet of 8,000 or so devices and employed a probably unrealistically generous timeout of 15 seconds, i.e. four attempts per minute per device, you could scan the entire IPv4 range in just a hair over 93 days and that’s before excluding any known pointless address blocks. If you only spent a second on each ping you could do it in about six days.
For the sake of argument, cybercriminals are already operating botnets with upwards of 100,000 compromised machines doing their bidding. That bidding could well be (and probably is) probing random web servers for vulnerabilities. The largest confirmed botnet was the 911 S5 which contained about 19 million devices.
That’s amazing and scary at the same time. Thanks for putting it into perspective!
If it’s https it’s discoverable by hostname.
https://0xffsec.com/handbook/information-gathering/subdomain-enumeration/#certificate-transparency
I don’t know exactly how they do it, but probing every ipv4 address isn’t that hard
But there can be multiple websites behind one IP address?! They would not show when onhy accessing the IP. They would need to know about the domains somehow.