In the Chapter “Designing Yelp or Nearby Friends” of “Grokking the System Design Interview”, there is the following sentence in Section 10. Load Balancing (LB):
We can add LB layer at two places in our system 1) Between Clients and Application servers and 2) Between Application servers and Backend server.
I find it confusing that the diagram right before section 9 does not actually show the Application Servers and associated load balancers. In addition, could you please explain why we need both application servers and aggregation servers here? It seems to me that all the application servers will do is to call the aggregation servers.
Thanks for your help.