The other day while I was scrolling on Twitter, I saw this awesome tweet from Colm MacCárthaigh, a VP and Distinguished Engineer at Amazon Web Services (AWS).
He said that consistently adding new regions is AWS’s secret sauce and that it moves the needle for its business more than anything else. And it got me thinking.
My first thought was that it means that AWS is going to continue to open regions. While they have 26 regions launched with eight more announced (double the count of any other hyperscaler), it means more will come. And it’s not just AWS. All three hyperscalers are adding new regions and availability zones all the time. Synergy Research Group forecasts that more than 300 new data centers are in the pipeline and the global number is expected to hit 1,200 by 2026. (Though if I were Synergy, I’d discount the IBM contribution to the count as they aren’t real cloud.) This is great news for telco. The takeaway: if there isn’t a usable region near you now, there will be soon—you can bet on it.
But the other thought I had was regarding the common mistake technologists make when it comes to technical decisions: they bet on where a technology (or in this case, a hyperscaler region) is at the time of the decision. They don’t factor in the future. When selecting technology vendors, of course you review where their tech is today; but you also need to calculate their trajectory and velocity. Where are they going? How fast will they get there? Will they get you there faster than if you did it on your own? Should you invest your money to build it, or let them spend their money and ride their coattails?
With the investment that all three hyperscalers are making in their regions and tech, I think it’s a no-brainer: you should bet on their innovation and investment trajectories.
Don’t build a #fakecloud
At TelcoDR, telcos ask us all the time about the timing of the move to the cloud. “Hey, DR, there’s not a region near me. I’m going to build a private cloud instead.” After my facepalm moment, I shake my head and tell customers—no, that’s not how you should think about it. Instead, you should think that it’s coming soon, and that you should use this time to prepare for the day when it is available. You should thank your lucky stars that you’re not yet behind on your move to the public cloud. You have time to get ready, a luxury that other telcos in regions that have the public cloud already don’t have. Use it to your advantage!
Every time I see the news that a telco has chosen to build a private cloud, I mentally add five years to their public cloud journey (c-ya in 2027!). That’s because they’re going to have to spend significant CapEx money on the hardware. With every server you buy, your finance department is going to want you to squeeze every last ounce of use from it to maximize the investment. While this makes financial sense, it will add five years to your move to the public cloud.
The other thing private cloud does is lock your organization into the 10,000 tiny tech decisions your IT team makes that you have to live with in the private-cloud zoo. (I talk about the zoo in this video, at 6:15.) Which container system to use? Database? Orchestration product? How often do you upgrade? Instead of relying on the excellence of the Amazon / Microsoft / Google technologists, you’re relying on your team to make all decisions. I’m not knocking your team, but the hyperscalers have been managing mega data centers for 15+ years now while your team is building a private cloud for the very first time. Who would you rather have make those decisions?
Finally, going with a #fakecloud will waste time and money you could have spent training your teams on public cloud while they learn a custom private cloud tech stack. Private cloud and public cloud are two very different things. Don’t fool yourself into thinking these are reusable skills that cross over into the public cloud. While there is *some* overlap, it is minimal. Get your team ready for the move to the public cloud today.
Start hiring and training your teams now
Public cloud is a whole new tech stack: databases, APIs, services, orchestration. Each public cloud has its own set of tools that likely no one on your team knows how to use. I always recommend to telcos that they start slow, with non-mission critical applications and workloads. Move “easy” applications to the public cloud to gain experience with managing costs and working with finance to budget variable usage. Build new processes for provisioning instances, optimizing application design, and teaching your team how to build with the cloud. This is a better investment of time than learning how to manage some Red Hat / IBM Cloud cluster that you’re just going to throw into the trash in five years’ time.
Don’t forget about the new talent you need to hire! You’re going to need a whole swath of new skills in your org that you’ve never hired for. Your HR department is going to need to recruit with new hiring profiles and source talent from new pools outside of telco. How will you screen and interview for the public cloud? Which certifications and experience do you need? Cloud expertise is the most in-demand tech skill in the market today and you’re going to have to adjust your compensation packages to attract top talent. Is your organization ready for that?
Don’t have access to a public cloud? Start with Outposts, Azure Stacks, or Google Distributed Cloud
All three hyperscalers have come around and realized that they can’t build regions everywhere all at once. So they’ve compromised and come up with a solution: we’ll put our public cloud in your data center. AWS Outposts, Microsoft Azure Stack, and Google Distributed Cloud (just launched!) allow ANYONE to put a piece of the public cloud right in their own managed data centers—and in AWS’ case, it even provides the hardware. While you won’t get access to 100% of the services they offer in their mega data centers, there are some good basic offerings you can use to start the learning process for your teams and the movement of workloads to the tech stack of the public cloud.
Again, I recommend this option over building a private cloud. While this approach is theoretically the same (workloads on your data center), there’s one key difference: your team will learn how to use the tech stack of the public cloud in the process, which sets you up for the day when you inevitably move to the public cloud.
Deciding to take a calculated risk on future technology is a tough call. Fortunately, the hyperscalers are investing as much CapEx in networks to build out their services as telcos do. Just as crazy as it sounds for someone to decide to build their own network instead of buying services from you, that’s how crazy it is to build your own private cloud. The public cloud has come to telco—time to jump on the train and start to use it.