Why Facebook doesn't Provide Public Cloud Service
This is my answer to a Zhihu question Why doesn't Facebook Provide Public Cloud Service:
1. Public cloud is an enterprise market, while FB is purely a consumer company now.
It has to start from 0 in terms of customer relationship. While Amazon, MS and IBM have been playing in the enterprise market for tens of years. It would be a fundamental shift for FB in all aspects. Building a public cloud is a massive system engineering instead of a single point of technology like a sophisticated algorithm. It will not happen in one or two months, it requires many years of heavy investment before being profitable.
As an example, Google had a hard time entering the public cloud market, because in the early days they lag behind in enterprise customer facing features, and they need to establish customer relationships from scratch. But, I'm optimistic about the future of Google Cloud for their long term competitive advantage in infrastructure (datacenter technology and software stack), user-facing features are short term advantages which can be easily caught up in a short period of time.
2. FB has no competitive advantage over other providers.
Amazon is the market leader with largest market share and richest features. MS has huge existing customers on MS stack. Google has the state-of-the-art infrastructure. FB can't compete in any aspect.
3. Cloud is an ecosystem.
Amazon, MS and Google are unique in one way or another, they'll have their own ecosystems, like iOS vs Android in the mobile world today. Once ecosystems are formed, no chance for new comers with the same generation technology. My estimation is the ecosystems will be formed in 5 years, it's too late for FB.
4. The culture of FB is pretty much a top-down style, which is not a long term advantage.
This point is actually not so related to the question, but I consider important.
In the early days, they have a slogan "Move fast and break things". Top-down style means focus on user-facing features, choose whatever the convenient technologies/tools to meet the needs. The upside is it allows you to quickly develop a feature an iterate incrementally over time. It's mostly correct for a SNS company as the time to market is critical for them. But the downside of top-down is that it doesn't establish long term competitive advantage in fundamentals. In contrast, the culture of Google is bottom-up, it builds stuff layer by layer, every layer is reusable, every layer adds value. Here is a little bit comparison of infrastructure stack between Google and FB What does it take for Google to work at scale.
At the end, I'd like to highlight that innovation often happens in a bottom-up way. People might have dreamed iPhone before Steve Jobs made it in 2007, but it's only a dream, because every new stuff stands on the shoulder of old stuff, only when the supporting technologies are ready, you can start doing the next big thing.
