When major online services suddenly went dark earlier this week due to an outrage of Cloud-fare, it exposes how deeply inter-twined so much of the web is with a single infrastructure provider. The outrage itself stemmed from a largely technical but crucial aspect of the company’s infrastructure, its bot management feature.
Cloudflare’s systems support a wide swath of internet traffic, and a breakdown on its network reverberated globally. But just how big is Cloudflare’s footprint on the internet, and what does “running on Cloudflare” really mean for websites, enterprises, and users?
Cloudflare’s own public data shows that its global network comprises hundreds of data centres across more than 100 countries, allowing it to process traffic at the ‘edge” — that means close to users rather than routing all requests back to origin severs.
The company serves on average 81 million HTTP requests per second second across its network.
Independent usage-tracking sites report that Cloudflare is used by roughly 20.4% of all websites worldwide as a reverse proxy. A reverse proxy is like a middleman between a client and a server that triages client requests. With a fifth of all websites using Cloudflare, the infrastructure provider is at the top of the reverse proxy category. Among websites that employ some reverse proxy service or content-delivery network (CDN), Cloudflare’s penetration is even more pronounced, with over 81% of those sites.
Publicly available lists of companies and domains relying on Cloudflare include names like LinkedIn, X, Vimeo, PayPal, Shopify, Chat GPT, and Discord. This demonstrates that its user base spans high-traffic enterprise platforms as well as smaller web properties. In India, names like Titan, Air India and HDFC rely on on the infrastructure provider.
The company notes that its service reaches more than 200 cities globally, facilitating coverage near user’s locations rather than relying on centralized routing. The dominance of Cloudflare in this space carries both power and risk. The scale of its adoption means that when service disruption occurs, the impact can be broad, affection not just niche sites but also widely used applications and services.
The outrage this week illustrates how dependent the platforms are on Cloudflare’s network. The usage statstics suggest that “running on Cloudflare” is not an exclusive characteristic reserved for giant, high-traffic firms. The large subset of web properties using Cloudflare spans many tiers of traffic volume, meaning that the company’s engineering decisions and failure modes have ripple effects across the entire spectrum of the web, from niche blogs to global enterprises.
Research in internet infrastructure and dependency further underscores the fragility inherent in centralization. Academic studies have documented that a relatively small number of third-party CDNs, DNS, and hosting services provide critical infrastructure for a large share of popular domains globally.
These trends create a dynamic where a failure or misconfiguration at a key provider like Cloudflare may propogate widely.
While “running on Cloudflare” does not literally mean every web request anywhere relies on its systems, the coverage is significant enough to make the company a backbone for large swathes of online services.
The sheer number of websites under its reverse proxy, the global scale of its data centers, and the presence of major enterprise platforms among its customers all contribute to a scenario where the company’s stability and engineering robustness are crucial to the functioning of the modern internet.
