As part of their ongoing ‘birthday week’ announcements, Cloudflare announced a new object store service, R2, that has garnered attention for several reasons.
- Adding unstructured storage to their portfolio will nicely augment Cloudflare’s existing platform. It has the opportunity to make the CDN more compelling and give Cloudflare Workers (a serverless platform hosted on its edge network) more functionality.
- They are implementing the service using the S3 API, meaning that AWS customers can easily migrate files to Cloudflare.
- They’re pricing the base storage rates competitively with cloud providers, with the additional kicker that
- Cloudflare won’t charge any egress fees on the data served from R2. (This is the biggest element of the announcement.)
Corey Quinn breaks down what no egress fees mean for customers, and it’s worth taking a look at his analysis because the potential savings are considerable.
One final point: Now let’s remember that the internet is 1-to-many. If 1 million people download that 1GB this month, my cost with @cloudflare R2 this way rounds up to 13¢. With @awscloud S3 it’s $59,247.52.
THAT is why people are losing their minds over this.
— Corey Quinn (@QuinnyPig) September 29, 2021
While this is a clear win for customer economics, I’m quite curious what it means for Cloudflare’s broader financials.
Cloudflare’s CDN heritage colors the expectations of any additional services they launch. First, there is an expectation for CDN-level delivery speeds. Second, there are no explicitly defined regions at this time, as Cloudflare is designed to be a global delivery network. (And on top of these expectations of global reach at low latencies, Cloudflare has built this service with eleven-9’s of reliability.) Running this service is going to cost $$$.
While Cloudflare’s R2 announcement references their plans to algorithmically manage storage across its network to optimize regional availability and frequency of access, the general implication of storing and serving data objects at extremely low latency across a global network is that this will need infrastructure. And infrastructure is capital intensive.
Right now R2 is at the waitlist stage and is not publicly available, but this pricing-driven, head on attack is clearly designed to siphon cloud workloads at volume (in addition to other S3-compatible services, like MinIO, Scality, and Storj). Presumably the company’s existing infrastructure has the capacity to deal with the launch of R2, but if they get the customer uptake they presumably want, it seems like capital expenditure to support the service is inevitable.
In short, Cloudflare is deliberately attacking the margins of AWS, Azure, and GCP (and thus limiting their own potential marginal revenue from the service) while also potentially increasing their own need for future capital investments.
This service announcement is amazingly ambitious and I, like many others, am so curious to see how this all plays out once R2 goes GA.
Those closest to me know my thinking here…if either Cloudflare and/or Fastly get ambitious enough, we likely mention them in the same breath as AWS & Azure in 10 years. The question is if they have the vision & ambition. Cloudflare apparently does. Fun times https://t.co/htqfSFqfjc
— Jason Warner (@jasoncwarner) September 28, 2021
Disclosure: AWS, Cloudflare, and Storj are RedMonk clients.
Image credit: Willrow Hood – stock.adobe.com