Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Decentralisation: the next big step for the world wide web

theguardian.com

Post reply on HN
Print view View original

171–180 of 218 posts

Re: Decentralisation: the next big step for the world wide web

#171

Post by imtringued »

skybrian wrote: post #31
sheeshkebab wrote: post #27

With phones doubling performance every 1-2 years, and desktops/laptops largely stagnating, it looks like in a few years, a server rack will fit into your smartphone. And most enterprise software (and almost no individual user) barely needs more than a rack of current gen servers. So yeah, decentralization will be upon us soon enough.

This seems implausible. Moore's law is over and phones are limited by size and power. There's no reason to believe progress will continue until a phone is the same speed as a typical desktop machine, let alone a server rack.

Especially if you consider how much of a lead apple had over ARM and the custom designs of SoC vendors like qualcom then you will realize that the easy growth is already gone. 2 years for a 50% increase in performance just to catch up with the strongest competitor is very far away from doubling performance every year.

Re: Decentralisation: the next big step for the world wide web

#172

Post by imtringued »

lolive wrote: post #158
vezycash wrote: post #149

IMO, there are two key barriers to a decentralized internet (other factors are minor in comparison.) (1) IPv4 (2) Bandwidth limits IPv6 makes NAT unnecessary. With IP scarcity gone, IP addresses might become permanent like phone numbers. ISPs are currently making money off fixed IP addresses. Market forces would change that eventually.

The need for a search engine seems, imho, to be the major argument versus (full) decentralization. Another formulation for that: there is a (valid) need for (some kind of) centralization. (and from there, you all go back to full centralization, first because of search, then because of convenience, at last because of laziness).

I wonder if it is even possible to decentralize search. An authority needs to create an index and letting the authority query that index is far faster than downloading gigabytes of data on every single machine that wants to search the decentralized internet.

Re: Decentralisation: the next big step for the world wide web

#173

Post by buboard »

drasticmeasures wrote: post #165

This kind of decentralisation -- taking existing platforms controlled by megacorps and making them P2P... ...feels, to me, like a huge mistake. How would one eliminate hate speech and toxic content from it? Or illegal content? Or anything you put there and need removed to keep living your life freely? The technologists developing this tech hand-wave these concerns away citing "freedom of speech" -- but one's freedom…

That is an important point but kind of orthogonal to P2P. A popular P2P website, say DTube could delegate its moderation to some group of "censors", and have some of its listings removed. Besides, unmoderated spam/trolling-infested sites are not going to become very popular anyway. Now you surely can't completely remove content that has been distributed but you can condemn it to be forgotten. 'Organic' forgetting may actually be easier in distributed web because sharing of content is generally temporary - just like you can no longer find older movies seeded in torrent sites.

Re: Decentralisation: the next big step for the world wide web

#174

Post by lucozade »

zzzcpan wrote: post #166
drasticmeasures wrote: post #165

This kind of decentralisation -- taking existing platforms controlled by megacorps and making them P2P... ...feels, to me, like a huge mistake. How would one eliminate hate speech and toxic content from it? Or illegal content? Or anything you put there and need removed to keep living your life freely? The technologists developing this tech hand-wave these concerns away citing "freedom of speech" -- but one's freedom…

> How would one eliminate You simply make your own choices and don't follow/subscribe/view all that illegal, toxic, hate content. You know, the same way you do today by not visiting all those illegal, toxic, hate websites. They still exists though for those who don't share your views on policing content for other people.

As you say, you can ignore this content today. And yet the authorities still feel obliged to make access to child pornography etc illegal. There is no reason to believe that changing the platform would change this stance.

Just because it can't be dealt with by threatening the odd CEO or two, doesn't mean that it won't be dealt with some other way. Now, it may be the case that governments react to a change like this and just accept that they can't control terrorism, child porn etc etc. But it would be astonishingly naive to assume that that's the most likely outcome.

Re: Decentralisation: the next big step for the world wide web

#175

Post by wslh »

fghtr wrote: post #160
amelius wrote: post #147

Earlier quoted context omitted.

There is a semi-centralized solution: make search run on a handful of central machines, and check if the results of (e.g.) two of them match.

This problem is essentially solved by YaCy: http://yacy.net

I don't think you both understood what I said. It does not matter if you can find a decentralized solution to a problem because at the end the user will access it through an UI/UX/App that, in you example, can choose how to rank the search results beyond what the protocol dictates.

In the past people used the simple "mail" command to read emails but now they choose GMail or others because the UI/UX (or any other reason) is better or they like it more. The SMTP (federated) protocol is hidden.

Re: Decentralisation: the next big step for the world wide web

#176

Post by LinuxBender »

hazz99 wrote: post #133
pankajdoharey wrote: post #110

Hasnt half of the decentralisation problem been already solved by the Tor network which is encrypted and already decentralised? And they still regularly identify and stop tor servers. What makes them so sure that the decentralised nodes on a bitcoin network cannot be physically identified and stopped ? A truly decentralised network cannot be built on borrowed network. A truly decentralised interruption free, governme…

Important note: Tor does not encrypt your traffic Tor simply hides where your web requests originate from - it's up to you to to visit HTTPS sites and encrypt your communications. Also, Tor is quite decentralised but the existence of directory authorities undermines this, since presents a centralised component.

Sorry you are getting downvoted. This is very much correct and folks simply put a lot of faith in the proxy transport as the ends to a means. One vulnerability / bug (Tor has had many) can weaken that link. Tor is rarely installed correctly or in a secure manor. (forcing all packets through it and dropping anything that leaks from the browser, for starters)

Re: Decentralisation: the next big step for the world wide web

#177

Post by briatx »

hazz99 wrote: post #132
briatx wrote: post #109

Repeat after me: The Internet is already decentralized. The web is already decentralized. Blockchains do not have a monopoly on decentralization. People who assert this are trying to redefine the term to mean some kind of extreme P2P model that fits their narrative.

Repeat after me: It's effectively centralised. It's effectively centralised. Almost all our traffic goes through Google, Amazon and Facebook. It's extremely centralised. Blockchain a don't claim to have a monpoloy - they're just the most recent thing that's repopularised decentralisation. If Amazon servers go down, so does a significant portion of the internet. That's centralisation at work! Blockchain tech doesn't c…

> Repeat after me: It's effectively centralised. It's effectively centralised.

So is bitcoin. Only 3 or 4 companies own the majority of mining.

Companies on the internet are centralized, but not the internet itself.

Re: Decentralisation: the next big step for the world wide web

#178

Post by wslh »

buboard wrote: post #169
mr_toad wrote: post #67

Earlier quoted context omitted.

> It's the only box that can stay online with a stable connection 24/7. My router stays online 24/7. It already has a web server built in. I could hack it to make it serve a public website. But there’s absolutely no way I’m going to do that. The security and maintenance requirements are just too much of a PITA. It’s much easier, more secure and more reliable (and likely cheaper once you figure in depreciation and opp…

> The security and maintenance requirements thats why you need one of the sharing protocols, like IPFS that make security everyone's responsibility, not just yours. I don't get why you think cloud solution is so much better. Glorified CDNs are a clumsy intermediate solution until internet connections get fast enough for everyone, that running a sharing node will have negligible impact. E.g. no cloud provider can comp…

> thats why you need one of the sharing protocols, like IPFS that make security everyone's responsibility, not just yours.

Everyone responsability = No one responsability.

If your data is lost by IPFS you don't have anyone to sue.

Re: Decentralisation: the next big step for the world wide web

#179

Post by the_clarence »

heavenlyblue wrote: post #155
codeka wrote: post #124

Earlier quoted context omitted.

None of that really matters if Mars is on the other side of sun from Earth. You'd need relay satellites to direct the signal around the sun, and even at the speed of light that's going to take tens of minutes, one way. As the parents suggest, even at Earth-Moon distances, we need to completely rethink things.

So who cares? Just adapt the TCP timeouts (resend timeouts) to 10 minutes. We’ll end up with the same, TCP-based system, except it’s going to be somewhat different, as time scales are not invariant for us humans.

I just wanted to play counter strike with the mars people

Re: Decentralisation: the next big step for the world wide web

#180

Post by AllegedAlec »

gunnihinn wrote: post #150
gunnihinn wrote: post #148

Earlier quoted context omitted.

I suppose those continue working fine in very local systems (contained in a ball of a few light-seconds radius) whose components move at speeds where relativistic effects can be discarded. Drop those constraints and you also need to drop even system-local globality because of relativity.

You _could_ introduce the One True Lamport Clock, and as long as you're in its light cone you can get global syncronization, but that comes at the cost of having to learn a lot about patience.

That would only work up until the point at which we start to travel at portions of c.
Post reply on HN