05 Jan 2022
In 2021, there was a lot of talk about moving workloads out of the cloud. Call it “cloud repatriation” or the “cloud diaspora”, the central idea is that enterprises are getting disillusioned with the cloud, and large players are moving their workloads back out of it.
The story goes that paying per use for resources is a great idea early in a company’s lifecycle but, as time goes on, cloud costs actually become larger and more unpredictable than the equivalent resources in an on-premises data center, or a colocation space.
The idea was explained well in May 2021 by Sarah Wang and Martin Casado at Andreessen Horowitz, who coined the “cloud repatriation” term: “While cloud clearly delivers on its promise early on in a company’s journey, the pressure it puts on margins can start to outweigh the benefits, as a company scales and growth slows.”
The big benefit of cloud is it’s quick. If you’re a startup, you can get something going fast in the cloud, get ahead of the competition, and win those new customers. Later in your company’s lifecycle, there aren’t so many new customers to be had, so you have to be more efficient, and you start to realise you could save money by running your own IT - and you might have to if you want to turn a profit.
They see a paradox: “You’re crazy if you don’t start in the cloud; you’re crazy if you stay on it.”
A large software company told Wang and Casado that cloud costs are 80 percent of its cost of revenue (COR). They estimate that most companies can halve this cost by moving workloads out of the cloud, and reckon that the top 50 public software companies operating in the cloud are losing $100 billion in potential market value by not doing so.
They also explained why repatriating loads isn’t easy: “Because this shift happens later in a company’s life, it is difficult to reverse as it’s a result of years of development focused on new features, and not infrastructure optimisation. Hence a rewrite or the significant restructuring needed to dramatically improve efficiency can take years, and is often considered a non-starter.”
How hard is it to move workloads back from the cloud? Well, consider the fact that there really is only one poster-child for this move. Dropbox started its Magic Pocket project to move workloads off AWS, back in 2016. It saved the company shedloads of money, and yet we’ve not seen a huge and public movement out of the cloud. Why?
Greg Moss at UpStack told me companies are stuck with cloud dependency for long periods, because of skills, politics and inertia, and the fact that they only just got into the cloud. “No one's going to pull out of a massive migration or massive decision after just doing it - regardless of the outcome. So that's why the early adopters are now exiting. Finally! After 10 or 12 years.”
They have troubles because they don’t have staff with on-prem skills any more, he said, and there are sluggish gatekeepers: “There’s a person who's been there 15 years, who just doesn't want to do more than what he's doing. I've seen large companies dismiss $50 million a year savings because the gatekeeper, a $150,000 employee, just doesn't let the management team know that there's an opportunity.”
Rich Hoyer of SADA, put the counter case at VentureBeat - which really amounts to saying, as Wang and Cassado do, that the issue is more complicated. “If you’re running large workloads in the public cloud, it’s not time to panic,” says Hoyer. “It’s highly unlikely you are wasting half or two-thirds of your infrastructure costs by running in the cloud without any incremental benefits to show for it.”
My colleague Sebastian Moss spoke to a leader of the Dropbox Magic Pocket project last year and, sure enough, there’s more to it. What Dropbox did is to determine which loads are best in each location, and optimise them there, which makes a lot of sense.
It was a huge investment for Dropbox but, given the company’s costs must be almost entirely on its storage and IT, it had huge payoffs. What does that mean for more average companies? Let’s return to that in a future column...
All the best for the New Year to you all!
Note: You can read more about Cloud or Colo in Kao Data's latest whitepaper, available for download here.