Hacker News new | ask | show | jobs
by arcticbull 1222 days ago
> The tech has already been proven—carriers just need to be mandated to add it because there's no business reason to do so.

People say that but in the decades since it was thought up mesh networks have never actually managed to work, not at any scale. Routing efficiently is extremely complex especially when all the nodes keep moving around. I concede it may be better than nothing in an emergency but I strongly suspect it's just not very useful in general, which is probably why carriers haven't mandated it. After all if it worked you wouldn't need carriers at all, yet here they are...

2 comments

A route on a mesh network is only as big as the smallest link. Of course as a general substitute for connectivity it holds little promise. Like any technology it has its pluses and minuses.

In situations without access to a fat, centralized link or situations where resilience is more important than bandwidth or latency, mesh networks are a great choice. Emergencies are such situations but remote areas or low power needs are other situations suited for mesh networks.

I've played around with a few different "simple" mesh techs in my time, some of the really cheap/easily available stuff can be really useful. I've seen examples of people sending entire kilobytes-per-second of data over a kilometer with the right antennas and line-of-sight with cheap, sub-£100 kit.

Things like LoRa which I've not had a chance to play with yet look extremely promising too for long range, low bandwidth.

If there was good reason for it, and sufficiently easy to set-up kit, I don't see why more people wouldn't help build out networks by setting up routing nodes in their homes; everyone installs a WiFi router because their ISP tells them to, but what about if people's only choice to get internet was to add a femtocell or similar (and it was legal and cost effective)? People would do it.

The real hold up, I'd say, is not so much effectiveness, but need. Nobody _needs_ to usurp the role of their carrier and build out a community network, so it doesn't happen.

If for example, using that network became so costly, say for example, like electricity recently has in the UK and Europe, that for many, building out their own infrastructure with solar and batteries became cost effective, I could see it happening here too; with frequency band licensing etc likely being the main barrier.

i really don't want to be this person, but Starlink is a mesh network, i have a crap connection to it, and i still routinely get over 200mbit download and 30-50mbit uploads. and a ping 1/3rd of my at&t connection (minimums around 19ms, which i haven't seen since i lived on the west coast)

> After all if it worked you wouldn't need carriers at all, yet here they are...

It'd work fine in more dense areas. Probably not the higher band 5G stuff, but lower bands and 4G. There's no carrier incentive to do this, that's why OP said "mandated".

I'm not sure I'd call it a mesh. If you have both full control over all the nodes and monitoring from a central place, the problem is significantly simpler. Mesh typically implies independent/unpredictable actors and self reorganisation / self healing.

Centrally pushing new routes based on quality/load data is not it.

> Starlink is a mesh network

Are the inter-satellite links actually active yet? Last I heard Starlink satellites all still needed to be in range of a ground station (explaining the lack of ocean coverage for their marine service)

my internet exit point varies. Sometimes it's in TX, but the last time i checked i was exiting in Bellevue, WA. as opposed to any other provider around me which always exits from an F5 in Dallas-Fort Worth, TX.

I'm not sure this proves it, but i can't think of many reasons for that to be the case.

Currently, AFAIK starlink is still Point-to-point, bouncing once to a satellite and back down to fixed ground stations. They intend to add inter-satellite links, but I'm not sure that's the case yet.
ISL is active