Comment on Disney+ cancellation page crashes as customers rush to quit after Kimmel suspension
BCsven@lemmy.ca 2 weeks agoOh right, I skipped a part. It is not really a dev complexity prep issue. You build the database that serves the comments etc in as of in one place, then you deploy cache servers for scaling. They self replicate, so a comment in California gets commited to the dbase, the server in new York pulls the info over from the Cali change, it sends back that it is synced with the change. And vice versa. The caching servers do the work, not your program.
DreamlandLividity@lemmy.world 2 weeks ago
That entirely depends on your application. What you described is one possible approach, that will only work in specific circumstances.
BCsven@lemmy.ca 2 weeks ago
Besides application specifics, its how the internet works currently to give low latency. AWS, Azure, Linode etc have data centers across the globe to replicate data near where the people are.
DreamlandLividity@lemmy.world 2 weeks ago
Again, yes and no. While you are right pretty much every larger website will use a cache server in some way (at least in form of a CDN), cache servers really don’t help you in any way for things like a customer canceling their subscription, which is what this post is about. That is all back-end work. Yes, those are probably those app specifics you mention but glossing over them misses the point why solving this is not as easy as enabling auto-scaleing.