Most active commenters
  • seestem(4)

←back to thread

The Internet Archive is back online

(arstechnica.com)
339 points Wingy | 12 comments | 16 Oct 24 11:09 UTC | HN request time: 0.001s | source | bottom
Show context
seestem ◴[16 Oct 24 12:10 UTC] No.41858155[source]▶
It would be better if the Internet Archive was decentralized without a central point of failure, maybe run on something like bittorent.
replies(8): >>41858170 #>>41858243 #>>41858272 #>>41858279 #>>41858590 #>>41859340 #>>41860089 #>>41861043 #
Cheer2171 ◴[16 Oct 24 12:26 UTC] No.41858279[source]▶
You say this as if it is an original idea. Of course the IA is working on this and have been for over 6 years. There already is a DWeb version. They have been advancing DWeb infrastructure. The IA hosts all kinds of DWeb developer events.

But it is over 50 petabytes and the IA gets a huge amount of traffic through the regular web that they need to serve quickly and efficiently to their users.

Guess what has happened over 6 years of decentralization of 50 TB? People only seed what they want or care about and there aren't enough seeders to host. They set all this up and nobody volunteers. You're a DWeb advocate and you haven't been seeding. That's a recipe for disaster if they rely on the goodness of volunteer seeders. The IA's mission is broader. DWeb will ever only compliment the IAs mission.

https://blog.archive.org/2021/02/18/behind-the-scenes-of-the...

https://www.bleepingcomputer.com/news/technology/archiveorg-...

replies(8): >>41858315 #>>41858402 #>>41858496 #>>41858621 #>>41858851 #>>41859300 #>>41860711 #>>41861723 #
1. seestem ◴[16 Oct 24 12:30 UTC] No.41858315[source]▶
It is not my original idea, but it is an obvious idea to anyone who knows how the internet works, I just added it to get the discussion going.
replies(1): >>41858355 #
2. Cheer2171 ◴[16 Oct 24 12:35 UTC] No.41858355[source]▶
It already exists. IA has had a DWeb for 6 years. Nobody seeds it.
replies(2): >>41858418 #>>41861742 #
3. seestem ◴[16 Oct 24 12:43 UTC] No.41858418[source]▶
Incentivising seeding is hard. Maybe cryptocurrencies can be useful here, but I understand not everyone likes them especially here on HN. In retrospect the ideal setup would have been if archiving was included into the core HTTP protocol.
replies(4): >>41858442 #>>41858806 #>>41858994 #>>41859181 #
4. 6footgeek ◴[16 Oct 24 13:04 UTC] No.41858627{4}[source]▶
Actually I think their wild ideas are contributing to this discussion.. I think your apparent hurt at this suggestion is a bit silly tbh.

Hear me out. I hadn't heard of the prior work you are raging about. But I am interested in decentralised tech - so, by this person mentioning their idea it got me interested to read further. Then you entered and went a bit madlad on them.. I'm not sure why when you could have just said "good question, here is everything we know about the effort to do exactly this" and then maybe that hacker would have been inspired to go read and maybe contribute. Just my 2 cents ;-)

5. ◴[16 Oct 24 13:15 UTC] No.41858718{4}[source]▶
6. Cthulhu_ ◴[16 Oct 24 13:26 UTC] No.41858806{3}[source]▶
Cryptocurrencies implies that people would pay / get paid for it... just pay the directly IA then, or your own servers. Cryptocurrencies imply someone's skimming off a lot for their own pockets.
replies(1): >>41858904 #
7. seestem ◴[16 Oct 24 13:36 UTC] No.41858904{4}[source]▶
I meant more like in bitcoin how the miners get paid for mining, or how validators are rewarded in proof of stake blockchains. This techniques can be used to incentivised seeders.
8. usr1106 ◴[16 Oct 24 13:46 UTC] No.41858994{3}[source]▶
Maybe someone can invent a proof of seeding protocol? So that would bring some good to the public instead of just burning energy. Don't ask me how it would work...
replies(1): >>41859531 #
9. ◴[16 Oct 24 13:46 UTC] No.41858996{4}[source]▶
10. greenie_beans ◴[16 Oct 24 14:06 UTC] No.41859181{3}[source]▶
i'm doubtful that whatever crypto incentive that is offered will make up the cost for me to DIY this in my home. which is why crypto miners scale, making crypto a centralized system. i don't care what the latest white paper says, it is still controlled by few people and not decentralized. in the same way that the ussr replicated the centralization of american capitalism, crypto replicates the centralization of trust while marketed as something else.
11. nikisweeting ◴[16 Oct 24 14:36 UTC] No.41859531{4}[source]▶
Storj, Filecoin, etc. fill this gap but it's still really hard to earn enough to justify the effort at small scales.
12. creer ◴[16 Oct 24 17:47 UTC] No.41861742[source]▶
If nobody seeds it (or continue development) then it's dead. Inspiration and perhaps code for the next effort sure. But not "exist" that makes a difference.