Comparison
Layer 7 load balancingvsSidecar proxy
Layer 7 load balancing
the proxy reads the request, routes on the path, retries the failed one on another backend and adds a trace header.
Distributing traffic per request at the application layer, which requires terminating and parsing the protocol. It buys path and header routing, per-request balancing, retries, timeouts, rewriting and observability, and it costs CPU, latency and the need to handle certificates. Nearly everything a service mesh or an ingress does that sounds clever is this distinction and nothing more.
Full entry →Sidecar proxy
every connection your service makes actually goes through a proxy in the same pod, which is where the retries and the TLS happen.
The per-workload proxy that intercepts inbound and outbound traffic, giving a mesh its identity, encryption, routing and telemetry without application changes. It is the data plane of a service mesh, and it is why mesh features apply uniformly to services written in five languages. Its costs are a hop of latency, memory and CPU on every pod, and a genuinely difficult startup and shutdown ordering problem with the application beside it.
Full entry →