In a surveillance state with secrecy orders. Also aiming for IPO with all the positive effects that follows with Wall St’s influence. The Google analogy might be better than the writer thought.
All that said, this is a pretty awesome offering that will do way more good than bad.
In the US, a handful of companies control the backbone acting pretty similar. A few more are ISP’s of most of us. They’re all shitty. So, it feels more like a small club benefiting itself as much as possibly is getting a new member. Well, plus that company that connects networks for companies using its datacenters as middlemen. Can’t recall its name.
On inside of a country, net hasn’t really been decentralized for a long time for many of us.
As someone actively working to to re-decentralize routing (by building turnkey meshnet solutions for communities), this is a brand of defeatism. We have the technology to re-decentralize; we just need the will.
It’s an objective statement of whats been going on with commercial networks for decades, not defeatism. I just like bringing reality into the mix so anyone reading aiming to change things knows where they’re starting. And Im glad you’re working on actual decentralization.
It’s an objective statement of whats been going on with commercial networks for decades, not defeatism.
Fair enough. I just don’t think it’s a good reason to cede more ground, especially not to CloudFlare, who have demonstrated some… rather uninspiring practices of late.
Can you give details? If you’re taslking about their decision to write their own Wireguard implementation and do it outside the Wireguard project, I think that’s defensible.
I agree with your main point, though. What decentralization project(s) are you working on?
CloudFlare’s goal is to improve the performance and reliability of their customer’s sites. The sites that sit behind CF’s CDN and firewall are the ones footing the bill.
Routing mobile traffic to their data centers gives their customers a slight, but measurable, advantage over all other sites on the internet. It certainly gives their CDN an edge on Akamai or Fastly, in the same way that 1.1.1.1 gave them an edge on all the other DNS providers.
Besides, for any sites that are CloudFlare clients (which is an awful lot of them, especially long-tail little sites that qualify for the free plan), they’re carrying the traffic either way. Using a VPN to get it into their data center vs just carrying raw TCP doesn’t meaningfully change how much bandwidth they’re using.
I don’t understand the upside from the user’s perspective, though.
Most of the net is already on TLS, the parts that people mostly use at least, so the data confidentiality part is not really a reason why people would use a VPN.
Privacy-wise, Cloudflare has so many PoPs that they are effectively covering only the local isp part of it in a VPN. Sure it might hide an end-user ip address, but then still give away at least the country of origin if not more.
Performance-wise, it’s pretty nebulous and hard to quantify, it’s not like the DNS case where switching to 1.1.1.1 added easy latency wins.
Yes, because when they IPO everyone’s privacy is screwed. eBay v Newmark set the standard that by law they have to maximize shareholder profits, so that means it would be wasteful to not capitalize on the user data they have access to.
So I expect within a few years for this to become a mass data collection service selling to Google and whoever else.
Let us give all our traffic to an American for profit company. What could possibly go wrong?
In a surveillance state with secrecy orders. Also aiming for IPO with all the positive effects that follows with Wall St’s influence. The Google analogy might be better than the writer thought.
All that said, this is a pretty awesome offering that will do way more good than bad.
[Comment removed by author]
Let’s take a nice decentralized network and put a centralized overlay on top. What could go wrong?
In the US, a handful of companies control the backbone acting pretty similar. A few more are ISP’s of most of us. They’re all shitty. So, it feels more like a small club benefiting itself as much as possibly is getting a new member. Well, plus that company that connects networks for companies using its datacenters as middlemen. Can’t recall its name.
On inside of a country, net hasn’t really been decentralized for a long time for many of us.
As someone actively working to to re-decentralize routing (by building turnkey meshnet solutions for communities), this is a brand of defeatism. We have the technology to re-decentralize; we just need the will.
It’s an objective statement of whats been going on with commercial networks for decades, not defeatism. I just like bringing reality into the mix so anyone reading aiming to change things knows where they’re starting. And Im glad you’re working on actual decentralization.
Fair enough. I just don’t think it’s a good reason to cede more ground, especially not to CloudFlare, who have demonstrated some… rather uninspiring practices of late.
Can you give details? If you’re taslking about their decision to write their own Wireguard implementation and do it outside the Wireguard project, I think that’s defensible.
I agree with your main point, though. What decentralization project(s) are you working on?
Maybe I missed something. What they been doing?
Too good to be true?
Not really.
CloudFlare’s goal is to improve the performance and reliability of their customer’s sites. The sites that sit behind CF’s CDN and firewall are the ones footing the bill.
Routing mobile traffic to their data centers gives their customers a slight, but measurable, advantage over all other sites on the internet. It certainly gives their CDN an edge on Akamai or Fastly, in the same way that 1.1.1.1 gave them an edge on all the other DNS providers.
Besides, for any sites that are CloudFlare clients (which is an awful lot of them, especially long-tail little sites that qualify for the free plan), they’re carrying the traffic either way. Using a VPN to get it into their data center vs just carrying raw TCP doesn’t meaningfully change how much bandwidth they’re using.
I don’t understand the upside from the user’s perspective, though.
Most of the net is already on TLS, the parts that people mostly use at least, so the data confidentiality part is not really a reason why people would use a VPN.
Privacy-wise, Cloudflare has so many PoPs that they are effectively covering only the local isp part of it in a VPN. Sure it might hide an end-user ip address, but then still give away at least the country of origin if not more.
Performance-wise, it’s pretty nebulous and hard to quantify, it’s not like the DNS case where switching to 1.1.1.1 added easy latency wins.
Yes, because when they IPO everyone’s privacy is screwed. eBay v Newmark set the standard that by law they have to maximize shareholder profits, so that means it would be wasteful to not capitalize on the user data they have access to.
So I expect within a few years for this to become a mass data collection service selling to Google and whoever else.
Sounds great! How much faster is it?