> Free customer base—Free customers are an important part of our business. These customers sign up for our service through our self-serve portal and are typically individual developers, early stage startups, hobbyists, and other users. Our free customers create scale, serve as efficient brand marketing, and help us attract developers, customers, and potential employees. These free customers expose us to diverse traffic, threats, and problems, often allowing us to see potential security, performance, and reliability issues at the earliest stage. This knowledge allows us to improve our products and deliver more effective solutions to our paid customers. In addition, the added scale and diversity of this traffic makes us valuable to a diverse set of global ISPs, improving the breadth and economic terms of our interconnections, bandwidth costs, and co-location expenses. Finally, the enthusiastic engagement of our free customer base represents a “virtual quality assurance” function that allows us to maintain a high rate of product innovation, while ensuring products are extensively tested in real world environments before they are deployed to enterprise customers.
(page 80)
Which just continues to reinforce what I said at the end of [2], about their business model instilling confidence in leveraging them at a free-tier level.
It's basically real world chaos monkeys that are stress testing their systems.
Combined with signal intelligence, those chaos monkeys are actively improving their systems.
Then, as a final bonus, absorbing the DDoS attacks aimed at those chaos monkeys (and their paid tier) provides a lot of globally diverse and distributed inbound traffic to them, favorably offsetting the outbound traffic they're serving from from their network. Improving their position during interconnect negotiations.
There are quite natural synergies that emerge as a product of their particular choices in structuring their business. Which is far more rare of a thing than you'd expect.
And we all know public companies never change 'key elements' of their business models... (hint: they do)
1) checks what's your current IP (an easy way is to do a GET request to http://icanhazip.com/, available thanks to Red Hat engineer Major Hayden)
2) checks what was the last IP you sent to Cloudflare (saved in a text file on your disk or something like data)
3) if they are different, send a PUT request to the Cloudflare API with the new IP: https://api.cloudflare.com/#dns-records-for-a-zone-update-dn...
4) Then save that new IP in the text file, so you can check next time
Finally, you edit your crontab to run that script every few hours or so.
What's the reason for having such a low lease time on public IPs?
Additionally they're less than friendly to users who strive for anonymity via forced captcha which further empowers google's massive machine learning/data classification efforts.
Cloudflare are big and competition is something they clearly lack.
The lack of competition probably shows the difficulty of the task. A global CDN is complicated and expensive (physical infrastructure, all the interconnect agreements) and a big technical challenge.
Cloudflare was rather innovative, with their authorative and recursive DNS, peering agreements with many ISPs including the ones that others didn't bother about, free SSL (Which they offered with Comodo even before Let's Encrypt), etc.
However, they give me a small sample as an inducement to buy; presumably to increase sales beyond a baseline, in quantities and at prices sufficient to cover the cost of not only the free food they give away, but the labor to do so.
Cloudflare can provide a free product that's WORTH paying for, if their paid product is so much better, and its sales increased by so much, that the free giveaway can be considered a marketing cost that pays for itself.
No, they don't compromise between them, nor do they reflect one more than the other: the basic law of supply and demand is that markets clear at the quantity and price where the marginal cost to produce equals the marginal value to customers which also equals the sale price.
Common examples that violate each one in term: Facebook, Used cars, Veblen goods, Utilities, Healthcare.
A market that has one large player that completely dominates certain segment(s) is not likely to be a healthy one that is providing value at near cost, why would it be? The incentive is always to charge as much as possible.
I don't think there is actually an equilibrium price defined as some kind of compromise between supply and demand curves absent those circumstances, either; there's an actual price at any given time, but it's not an equilibrium (particularly, without rationality—which incorporates perfect information—you don't have any expectation of an equilibrium even if all the other factors are present, at any price.)
Once you are discussing equilibrium—as the post the grandparent addressed did—you are implicitly discussing at least some of the ideal circumstances you’d like to negate.
Additionally, irrational does not mean unpredictable, after all advertising is extremely effective.
It may or may not be worth it, but it isn't a panacea.
It gives you a basically unlimited CDN, SSL, HTTP/2, DDos protection and fast DNS servers with almost 0 effort... for free ...
Thus increasing the stability and speed of many web properties that otherwise wouldn't pay for it.
I'm afraid this could easily change once the company is public and is under the pressures that come along with it, which is a shame.
(I also try to get them paid business when I can by nudging companies to use them)