As buzzwordy as it sounds I think between AR/VR/metaverse there will be some infra challenges that extend beyond "just needs a fatter pipe".
A bit like games need complicated netcode to compensate for latency.
Ask HN: What are the next internet infra problems?
121–130 of 230 posts
Re: Ask HN: What are the next internet infra problems?
#122Three things spring to mind:
1. IPv4 will persist, possibly forever. There's really no compelling reason to migrate to IPv6 other than address space and we've had decades at this point of getting around this problem with various flavours of NAT.
2. Ossification. We've taken the quite reasonable step of discarding any packets or traffic we don't understand from a POV of minimizing threats. For example, there were cases of bypassing security using packet fragmentation. But this makes it increasingly difficult to extend the protocols (eg reliable connectionless messaging aka a reliable UDP).
3. We don't really have a good solution for roaming. If you switch hotspot and get a new external IP it'll typically break your connections. A lot of work has been done to workaround this (eg carrier-grade NAT for mobile IPs) but identifying an endpoint with (address,port) (or just (address) for IPv6) is less than ideal.
Re: Ask HN: What are the next internet infra problems?
#123There’s still no good way for me to write an open source web application and have its users bear the cost of running it. This is a major regression from open source desktop software, and IMO is the reason open source web applications haven’t taken off more.
This is unfortunate because a couple of attempts get _so close_.
You've got public docker repositories, terraform, aws service catalog (or cloudformation), all of which could fill this gap directly or via their own services if they had a slightly more beginner-accessible workflow...
Re: Ask HN: What are the next internet infra problems?
#124There’s still no good way for me to write an open source web application and have its users bear the cost of running it. This is a major regression from open source desktop software, and IMO is the reason open source web applications haven’t taken off more.
I think we need something like micro payments. Take lichess for example it's the biggest open source chess website completely ad free and runs by donations. If you divide the total monthly costs to run the servers by the number of games and you get a tiny fraction of a dollar cent. I don't remember the exact number, but it has been in the order of 1k chess games equals 1$. If you could charge 0.1 cent per game, then…
"How to pay" always seemed like the easy half of the problem to solve, "how to manage what should be paid" seems much harder as nobody wants to deal with sorting through fraction of a penny payment approvals. Is 0.01 cent per game acceptable? 0.05? If you accept at 0.01 and it goes to 0.05 do you have to re-accept? If everyone decides sets 0.1 cents as an auto-approve boundary does that mean every site is now going to try to charge 0.0999 cents? Does some centralized entity try to set these rates instead? How does the system protect against the equivalent of collect call scams? Does it protect against that in a way that doesn't limit actually using a service very quickly?
It's like the permissions problem (really easy to prompt, really annoying to do so, really really bad to just assume yes or no all the time) except worse.
In regards to distributed serving using the client endpoints it has a strong tendency to be more work, less reliable, and not as scalable (see peertube). What has seemed to work is offloading as much of the functionality for that user as you possibly can to that user's device. E.g. want to play a game against the computer? Run the chess engine WASM bundle on your device instead of the server (lichess does this).
Re: Ask HN: What are the next internet infra problems?
#125How can we get the layperson to run a "homeserver" to host all their data locally and have a strong pki infra. 30 years ago, people would've said the same things about routers, so I think it's possible with the right ui/incentives
Routers work because ISPs require them. ISPs however aren’t fond of supporting the home server use case. I wish for a world where battery tech wasn’t so limited. Imagine if everyone could just run a full-fledged server 24/7 on their phone, as a simple app, with a reasonable data plan.
> Routers work because ISPs require them. ISPs however aren’t fond of supporting the home server use case.
This is the obvious and simplest solution. A built-in self-hosting platform right in the router, extensible with an external USB drive if the user needs it. But ISP's are notoriously terrible at everything and certainly can't be trusted with something like this.
Re: Ask HN: What are the next internet infra problems?
#126How can we get the layperson to run a "homeserver" to host all their data locally and have a strong pki infra. 30 years ago, people would've said the same things about routers, so I think it's possible with the right ui/incentives
Probably some ISPs will sell NAS or similar device and, with a monthly fee, you would have external access via a custom domain name (like john_doe.verizon.com, for example). A lot of non-techy people already have NAS, external hard drives, and things like that. I don't know how ISPs haven't done already this.
That's a DOA product. ISP's are notoriously terrible at everything; especially firmware. They can't be trusted with something like this.
Re: Ask HN: What are the next internet infra problems?
#127There’s still no good way for me to write an open source web application and have its users bear the cost of running it. This is a major regression from open source desktop software, and IMO is the reason open source web applications haven’t taken off more.
What's wrong with subscription fees?
Re: Ask HN: What are the next internet infra problems?
#128There’s still no good way for me to write an open source web application and have its users bear the cost of running it. This is a major regression from open source desktop software, and IMO is the reason open source web applications haven’t taken off more.
There's actually a lot of open source apps that have this property in a particular category: web3 apps. Users (or other intermediaries, not necessarily the app dev) pay the cost of transactions, and thereby pay the cost of the data storage / computation layer. HN doesn't seem to like web3 very much yet, but one of the most positive innovations it brings is giving open-source apps a direct business model (instead of t…
web3 apps do have high transaction costs, but that's not really reducing the costs that the web3 app would have otherwise paid to have a local database. So it costs money, but the developer doesn't really earn anything because of it.
Re: Ask HN: What are the next internet infra problems?
#129It would be nice to have a cable alternative to IEEE 802.15.4/6LoWPAN. IP seems to be the future for home automation. The problem with wireless is that you end up with devices you have to change batteries and reliability issues in high density housing. Laying a twisted pair ethernet cable to each temperature sensor is overkill and too expensive.
single-pair ethernet?
Re: Ask HN: What are the next internet infra problems?
#130Provable identity. Yes, you can do Oauth via the google or the facebook but sooner or later we need something that isn't tied to getting all your user interaction data ... Digital notary. So a third person (digitally) signing a transaction or other document exchange.