I’ve been hearing a lot lately about a possible mass internet outage for something like 72 hours sometime in the near future. Something about switching out large relay nodes relating to some sort of law enforcement activity of some kind. If this turns out to be true, is there anything in place to prevent mass node disqualification due to something like this?
1 Like
Id like to know where your reading these internet outage rumors…
4 Likes
Tbh, I’ve heard the same thing from DC people. Often related to temporary shutdown of BGPs (https://en.wikipedia.org/wiki/Border_Gateway_Protocol).
Good point around disqualification -> but luckily DQ is still a design doc and wouldn’t yet kick in for this scenario. Something to think about in the future though. Ideally we start to build in more SNOs to store pieces on mesh networks, etc to create further resilience.
1 Like
Sounds like what I’ve been hearing. Good to know, thanks!
1 Like
I heard the same things in the news of my country. To be honest, I suggest the biomass tardigrade technology for resiliency in the future, but I think that’s the long-term project of Tardigrade; recreate the organism and give him the ability to reproduce for resiliancy. Then the SNO could just drink it in a goat coffee crema.
2 Likes
Any links? First time I have heard about this
This is definitely tin foil hat territory.
3 Likes
I don’t have any links or anything but you can google it if you want. Lots of conspiracy related stuff around it but it’s probably just infrastructure upgrades or something… that’s assuming we actually even see any interruptions in the first place. I was just curious what this would mean for SNO’s if something like this did actually happen. Anything beyond that I think is probably beyond the scope of the forum.
Don’t forget to ground your tin foil hats or it actually acts as an antenna!
I have high doubts there will be any major interruption. Maybe a reduction in capacity if some big change is happening, but not a complete loss of connectivity. That would break so many things. Didn’t see anything on a quick google search or on reddit. Can’t see any reference to changes on BGP.
Ok, then let’s just say DNS servers went down… that’s happened before. Many people use dynamic DNS services for their SNO’s and that could effect large geographical areas all at once. By the way, the title does say it was just a rumor, but I figured it might be a good point to bring up. I’d hate to loose an established node with good reputation and a lot of data over something like this. Apparently as of right now it’s not a problem though so I guess we’re good. You know, I know many wealthy people who had high doubts the markets would crash and the whole world would basically take an extended vacation but you know… things happen. Better to be prepared ahead of time.
1 Like
For dealing with DNS problems the satellite should cache the results longer than TTL, that is, when the TTL expires, the satellite should try to resolve the name again, but if that fails, just try using the previous (expired) result. A lot of SNOs use Dynamic DNS services and if no-ip or similar goes down a lot of nodes would be offline at the same time. And TTL on thses is short so that the downtime after an IP change is minimal.
3 Likes
First of all most of the things outline here are downright impossible.
BGP and DNS are two completely different and independent protocols. One deals with path vector route table distribution and the other with resolving IP addresses to names, endpoints, etc. Any DNS issue that may arise is easily solvable with introducing a caching DNS resolver in your home, so that a bump in TTLs can be done in a consistent way.
Now, the only important development happening in the world of T3 BGP inet providers, which may result in the mentioned levels of disruption is called MANRS
However, even then, it is simply impossible for there to be a “mass internet outage”, because BGP is not a hub and spoke technology. The most that could happen is that the internet of a country goes out, because, say, they have too few BGP autonomous system uplinks towards the rest of the world and the better known comms, entertainment, etc services out there.
The other thing that could happen, is limited to countries which have a crappy government-led Internet Exchange infrastructure (IX, https://en.wikipedia.org/wiki/Internet_exchange_point). It is through that infrastructure, if its crappy, where BGP gets reduced from a conceptual path vector routing protocol to a simple hub and spoke-type protocol, because T3 ISP use IXs as points for terminating their infra close to one another and interconnecting.|
One such country is the US, for which Ive heard nothing but police state nightmares when it comes to their IX infrastructure and particularly how private businesses cannot open up IXs at will and have to coordinate with the government, which over the years has led to a fairly small number of IXs in the US.
In that case however, the worst that could happen is for a country-wide mass outage to occur.
Furthermore, most cloud providers run their own ASs, and regionally-redundant services hosted on the cloud won’t really suffer from and IX-level outage, because client traffic wont be leaving the cloud-provider AS even for backed parts of said service that might be limite to a single region.
2 Likes
and even if shit does hit the fan, you always have p2p apps for communication, which youll never be prevented from using in your local area with others.
1 Like
How would one even shut down the internet… sure it could be done locally to a certain degree…
but really the whole concept of the internet is based upon it being able to resist the loss of huge clusters of the network…
sure your ISP can shut down your internet, but if any ISP shuts down your internet for 72hr, then they either really suck or is knocked out by a natural disaster…
then there is the whole… stay at home people deal… covid19 if any isp globally took down their network for 72hr, they would get a shitstorm from all sides.
and then on top is the whole bad business thing… i mean it wouldn’t take many hours to power up a new ISP, if you had the required hardware…
so yeah your statement makes my bullshit’o’meter go into the red…
1 Like
I’m not sure what your source is [share it please] but I can assure you wherever you’ve gotten it from, that it is wrong…
reddit.com/r/sysadmin has nothing about it, neither does The Register, or any other credible news source…
Plus, the internet that we know does not use relay nodes…
However, services like TOR, IPFS, Psiphon, misc. darknet, etc. do run on relays…
Source: It is my job to keep up the latest on network threats and keeping companies connected in the modern world.
3 Likes
I woudnt consider it a mass internet outage though.