If you looked closer you’d see they’re talking about scale, not functionality. Anybody can build the functionality, few others have infrastructure that will keep latency low for users across the planet.
And yes they could use many local providers, except that significantly increase engineering costs
Would it really increase engineering costs? Who knows what they are doing now. is everything running in containers on Fargate? Is it EKS? Is it a bunch of EC2 instances in one zone behind load balancers? Are they exclusively using CloudWatch? Do they have DataDog collecting logs everywhere?
Without specifics, I wouldn’t immediately claim it’s impossible to pick providers across the globe or find some kind of other solution that doesn’t involve a high dependence on a single provider.
Unless all those local providers offer equivalent APIs / management interfaces and comparable hardware you’re going to have to deal with a lot of middleware
If you looked closer you’d see they’re talking about scale, not functionality. Anybody can build the functionality, few others have infrastructure that will keep latency low for users across the planet.
And yes they could use many local providers, except that significantly increase engineering costs
Would it really increase engineering costs? Who knows what they are doing now. is everything running in containers on Fargate? Is it EKS? Is it a bunch of EC2 instances in one zone behind load balancers? Are they exclusively using CloudWatch? Do they have DataDog collecting logs everywhere?
Without specifics, I wouldn’t immediately claim it’s impossible to pick providers across the globe or find some kind of other solution that doesn’t involve a high dependence on a single provider.
Unless all those local providers offer equivalent APIs / management interfaces and comparable hardware you’re going to have to deal with a lot of middleware