We used Verisign for mitigation of a 44Gbps volumetric attack and it worked very well. We also evaluated Neustar, but Verisign's infrastructure seemed to be more robust.
The problem is that you are basically mitm:ed all the time.
That to me sounds very much like MITM, although it is not a MITM attack since the entity controlling the domain opted into it, so basically it is voluntary MITM.
Using a VPS like EC2 is a different story since the decryption happens within the layer that you control. Of course you need to make sure that you choose a vendor for that layer that you trust, but on EC2 the traffic that amazon sees is encrypted with keys they don't have and decrypted with keys stored on a layer that I control. Amazon could read out the memory of my EC2 to get the keys but their business depends on not doing so, so in this case either I have a vendor that always will decrypt and read traffic (Cloudflare), or a vendor whose business depends on hypothetically being able to but not doing it. There is a clear difference to me.
That is the same for most CDN's (including CloudFront and all the other major offerings), so I'm not trying to single out Cloudflare.
This is why having a threat model is so important: it keeps you from wasting effort on things which sound like security but aren’t actually changing anything meaningful.
There has also been times where cloudflare (when setup improperly as I mentioned in the previous comment) has misrepresented the security of a connection, as shown by https://www.theregister.co.uk/2016/07/14/cloudflare_investig...
Still, large proxy-based CDNs do have the ability to completely bypass all the same-origin protections in the browser. Even if they are angels and don't abuse this trust for identity theft and surveillance, it makes them a juicy target for bad actors, state sponsored and otherwise.
Great way for a state actor to intercept your traffic. little bit of volumetric dos and the target themselves responds by tunning through your partner(s).
What's the logic behind this? It's still a single point of failure and relying on a corporation. If the daily stormer or 8chan tried to use them, they would probably kicked off as well.
Going from that to 'undesired political speech will be censored' requires more of a slippery cliff than a slippery slope.
What is this "direct link" you speak of? Did the shooters plan/recruit/organize their attacks on 8chan?
These include the Christchurch shootings, the Poway synagogue shooting and the El Paso Walmart shootin.
The Christchurch shooter shared his Facebook stream to 8chan before the shooting started, and it was spread from there.
The Poway shooter blamed/thanked 8chan for his views.
You are expanding a lot of effort defending 8chan here. Perhaps consider that it might not be worth defending.
I lived in a socialist country and you did not. Perhaps consider that you might not know where these current trends are pointing to.
Beyond that, given the announcement there, it stands to reason they were convinced to do it there.
Legally, a "direct link" is irrelevant, you can rarely find a "direct link" between two of anything. What matters legally is whether 8chan was a "proximate cause" in creating the mass shootings. Whether one thing is the "proximate cause" of another is often pretty difficult to discern.
However, as a helpful guide towards determining proximate cause, lawyers ask whether one thing was the "but for" cause of another, i.e., would the mass shootings occur "but for" 8Chan? Put another way, if 8Chan did not exist, would these shootings occur?
Unfortunately, we do not have an alternative reality to play out events without 8Chan, so we cannot know for certain, but we can use evidence (e.g., 8Chan chats, how the shooter interacted with 8Chan and others on the service, etc) to try to simulate that alternative reality. All of this analysis also needs to consider related issues like freedom of speech on public forums and any commercial interests.
I'm not saying 8Chan is guilty or innocent, just that the existence (or lack thereof) of a "direct link" is pretty meaningless.
It would be “censorship” if they actively antagonized any attempt to spread the information, such as by lawsuit or DMCA notice. They are just refusing to participate.
And given that the “information” is definitively known to be child pornography and violent white supremacy propaganda presented as news, I would personally say refusing to participate is the only responsible action.
But it's clear that it matters just what's being censored. Surely you wouldn't say the same trite clever-sounding hackerspeak if we're talking about censorship of threats, assault and child pornography, would you?
Cloudflare simply has the luxury of choosing which politically disagreeable parties they do not want to associate with because they are insignificant customers.
Pretending that this is not due to differences in politics and moral judgment is semantic smoke and mirrors.
Anyway, the point is that they are not a neutral carrier/providers. Unlike banks or telecoms which are required by regulation to accept any legal business. CF styles itself as neutral infrastructure, until they decide they are not.
The risk of getting deplatformed due to someone's moral judgment is quite real, even for an entity such as Wikipedia. For example they were blocked in the UK because the Virgin Killer album cover landed it on a block list used by major ISP.
But again, this is just a tangent. The core argument is that it is best not to rely on providers that have the freedom to make political/moral decisions who they deal with because that freedom makes them susceptible to moral denial of service attacks. You are one moral outrage away from being deplatformed.
Additionally, helping to block Wikipedia because China says so is much easier to excuse than blocking 4chan - they would just be complying with local regulations after all.
[1] https://www.cloudflare.com/press-releases/2015/cloudflare-an...
Is this a slippery slope argument.
Because there is a world in difference from discontinuing a few extremists customers, to discontinuing service for something akin to Wikipedia.
I'm not sure every slight compromise of principals is a slippery slope. It seems to me that CF generally aims at being neutral.
You also don't get to claim it supports 2Tbps if you've only weathered 44Gbps.
The issue with on-demand BGP mitigation is that an attacker can do short attacks on and off over a long period of time. Each time the mitigation kicks in, BGP propagation takes at least ~1 minute and will cause some downtime. Proper protection is always-on without requiring redirection.