Why Skipping Load Balancing Cripples Web Services: Layer‑4 vs Layer‑7 Explained

No Load Balancing Used

In this case users connect directly to the web server, which becomes a single point of failure; if the web server crashes, subsequent connections fail, and under high traffic the busy server may reject some users, causing loss of many users. Therefore this simple architecture is only suitable for small web sites or prototype development.

Layer‑4 Load Balancing

The diagram shows that the backend providing the web server service consists of multiple machines, and the external interface is no longer the server itself but a front‑end device performing load balancing. This device forwards user requests to a specific backend web server based on the IP and port in the packet and load‑balancing algorithms, using the transport‑layer (layer‑4) of the network stack, hence called layer‑4 load balancing. A similar layer‑7 (application‑layer) load balancing exists later. The forwarding process is similar to routing: the load‑balancer rewrites the user’s IP and port to those of a selected backend server and also modifies the server’s response packets to hide the real web server address.

Layer‑7 Load Balancing

The diagram for layer‑7 load balancing looks similar to the layer‑4 one, but the difference is that the backend web servers are divided into two groups based on request content such as / and /blog. User requests are dispatched to different backend machines according to the request content. Since the actual content must be read, a TCP connection must first be established, so the client creates separate TCP connections with both the front‑end load‑balancer and the final server. Consequently, layer‑7 load‑balancers have higher requirements and behave somewhat like proxies, unlike layer‑4 which only needs a connection to the backend server.