Let’s think it through · Start with a story
Many readers arrive at a library at once. The service desk doesn’t make everyone stand still in one endless line. How can a server organize its work?
Break down the process
- Receive requestsKeep connections distinct
- Schedule tasksWork concurrently or wait
- Reuse resourcesFor example, use a cache
- Share the loadAdd servers when needed
Servers handle many requests using concurrency, asynchronous processing, and multiple worker threads or processes.
While one task waits for a disk or network, the system can work on others. When demand grows, several machines can share the load.
“Serving at once” does not mean one CPU core executes every instruction at the exact same instant.
Your turn
If a server can serve many people, does that mean the number of connections can grow without limit?
Have an answer? Click to see the explanation
No. CPU, memory, network, and backend services all have limits, so they need monitoring and, when appropriate, scaling.
Takeaway
Servers handle many requests using concurrency, asynchronous processing, and multiple workers.
Try covering the diagram and explaining each step to someone at home. If a step is hard to explain, go back and look at that card again.