> In 2003 I was a developer on a website that served a million pages per hour using 2 databases, 4 webservers, and associated networking. Most of that hardware was there for automated failover.
fwiw, Stackoverflow runs on a similarly sized/deployed stack.
But that is because Stackoverflow is fine with their employees running a cheap, un-complex system that works (I can't remember the last SO outage!), and their employees are also happy with this.
Others, especially startups? Gotta chase these hipster credentials and use the free AWS/Azure/GCP credits somehow... and the employees don't want boring old school stuff, they want the newest hipster crap that's barely alpha tested, but in production. Utterly ridiculous. Most of them would be fine with three sub-$10k AMD servers and a bit of work by a seasoned Linux sys-admin. But eww. Linux management. Not cool/hip.
But that is because Stackoverflow is fine with their employees running a cheap, un-complex system that works (I can't remember the last SO outage!), and their employees are also happy with this.
Great example!
Far too often we introduce complexity to manage complexity. Then we fail to manage the complexity that we created, and conclude that we need even more complexity!
Taking the simple approach seems impossible and utterly scary. But it works more often than most would believe.
They do NOT need a stack of daily traffic analytics hardware. As I said above, remote RPCs take an order of magnitude more resources than function calls. If you are used to using various distributed tools to do analytics, you are used to requiring an order of magnitude more hardware than a simpler solution. And if there are inefficiencies in your solution, it is easy for a second order of magnitude to sneak in. Therefore your experience is going to seriously mislead you about the likely limits of a simpler solution.
Making this concrete, if your job requires a Hadoop cluster of less than 50 machines, odds are good that it could be run just fine on a single machine with a different technology choice. But if you are fully utilizing hundreds of machines in your cluster, then you actually need to be distributed. (Though you could achieve considerable savings for some of your jobs could be run in a more efficient architecture. Amazon in particular makes excellent use of this kind of hybrid architecture in places.)
Google is making excellent decisions for the scale that they operate on. You just are unlikely to operate at a scale that makes their decisions make much sense for you.
For what the hell do you need traffic analytics beyond parsing access logs? Current ad-tech/analytics is a mixture of outright fraud, fraud sold as "AI", fraud sold as "big data", middlemen claiming to solve said fraud, and privacy invasions on a scale that would make the Stasi drool.
Stackoverflow doesn't need this kind of shit and frankly, no one else does. The GDPR was a first start in the right direction and I'm seriously hoping coronavirus has at least one positive side and eliminates those fraudsters who survived the GDPR wave.
Stackoverflow logs every request into SQL Server which uses columnstore tables to handle all their analytics. It's fast, efficient and more than enough.
fwiw, Stackoverflow runs on a similarly sized/deployed stack.
But that is because Stackoverflow is fine with their employees running a cheap, un-complex system that works (I can't remember the last SO outage!), and their employees are also happy with this.
Others, especially startups? Gotta chase these hipster credentials and use the free AWS/Azure/GCP credits somehow... and the employees don't want boring old school stuff, they want the newest hipster crap that's barely alpha tested, but in production. Utterly ridiculous. Most of them would be fine with three sub-$10k AMD servers and a bit of work by a seasoned Linux sys-admin. But eww. Linux management. Not cool/hip.