Grokking the System Design Interview
Vote

0% completed

​

Serverless Architecture vs Traditional Server-based

A photo sharing app resizes every image that users upload. Most uploads happen around lunchtime, and almost nothing happens at night. The server that runs the resizing code is busy for about three hours a day. But the team pays for all 24 hours.

This is where a common design question appears: who manages the machines, and do you pay for time or for use? This lesson compares serverless platforms with traditional servers, and explains how to choose by looking at the shape of your traffic.

Two Ways to Run Your Code

A traditional server is capacity that you choose and keep

.....

.....

.....

Like the course? Get enrolled and start learning!
Frances Pierce

Frances Pierce

· 14 days ago

I think this lesson conflates "serverless vs server-based" with "cloud vs on-premise". If you run your service on cloud compute like EC2, then you get the benefits of pay-as-you-go and automatic scaling, and a lot of the server maintenance stuff is done for you, but you still have some control/responsibility for managing the servers.

Andre Hildinger

Andre Hildinger

· 6 months ago

This explanation seems to be confusing “serverless” with “serverless functions.” Serverless is an infrastructure management model, while Lambda-style functions are just one implementation pattern within it.

ECS with Fargate is serverless and, when configured with a minimum of 1 task, it operates just like a traditional server with no cold start.

Show 1 reply

Reading Progress

0%

Vote for new content