System Atlas
the atlas
NET · Networking6 minApplied

Forward proxy vs reverse proxy

Two middlemen with mirror-image jobs: one fronts the clients, one fronts the servers.

the gist

A forward proxy sits in front of clients and acts on their behalf — filtering, caching and anonymising outbound requests. A reverse proxy sits in front of servers and acts on their behalf — load balancing, terminating TLS, caching and shielding them. Same physical position, opposite side of the conversation.

ANALOGY // A forward proxy is a chauffeur who runs errands for you so shops never see who sent them; a reverse proxy is a company's front desk that greets every visitor and decides which department they go to.
living diagram01 / 05
ClientClientForward proxyrepresents clientsInternetReverse proxyrepresents serversServerServerServer
01

Forward proxy fronts the clients

A forward proxy sits on the client side of the network and every outbound request from those clients passes through it first. To the wider internet the traffic appears to originate from the proxy, not the individual machines behind it. This is the setup a company or school runs so that all employee or student browsing funnels through one controlled point.

02

It represents the clients

Because it stands in for the clients, the forward proxy is where you enforce policy on their behalf: block disallowed sites, log and filter traffic, cache popular downloads so they are fetched once, and hide the clients' real IP addresses for anonymity. The destination server has no idea who the real requester was — it only ever sees the proxy. Its job is to control and conceal the people making requests.

03

Reverse proxy fronts the servers

A reverse proxy is the mirror image: it sits on the server side, and every inbound request from the outside world hits it before reaching any backend. Clients believe they are talking directly to the website, but they are actually talking to the proxy, which never reveals the real servers behind it. This is what nearly every production website puts at its edge.

04

It represents the servers

Standing in for the servers, the reverse proxy is where you distribute load across many identical backends, terminate TLS so backends can speak plain HTTP internally, cache and compress responses, and absorb attacks before they reach application code. It presents one clean public address while hiding an entire fleet behind it. Its job is to protect and represent the machines answering requests.

05

Same spot, opposite sides

Both proxies sit physically between client and server and both relay traffic, which is why they get confused, but they guard opposite ends of the conversation. The forward proxy protects and represents whoever is making requests; the reverse proxy protects and represents whoever is answering them. Ask 'which side is this middleman working for?' and the label becomes obvious.

the full picture
01

Two middlemen, mirror-image jobs

A proxy is simply an intermediary that sits between a client and a server and relays requests between them. Both a forward proxy and a reverse proxy do exactly that, which is why the terms get muddled — physically they occupy the same slot in the middle of a conversation. The difference is entirely about *which side of the conversation the proxy is working for*.

A forward proxy works for the clients. It sits at the edge of the client's network, and outbound requests pass through it on their way to the wider internet. From the destination's point of view, the request appears to come from the proxy, not from the actual user behind it.

A reverse proxy works for the servers. It sits at the edge of the server's infrastructure, and inbound requests from the internet pass through it before reaching any backend. From the client's point of view, the proxy *is* the website; the real servers behind it are invisible.

Hold onto that single question — *is this middleman representing the requester or the responder?* — and every other property of the two follows naturally from it.

02

The forward proxy — representing the client

A forward proxy is deployed in front of a group of clients so that their traffic to the outside world is funnelled through one controlled point. The clients are usually explicitly configured to use it, and the servers they contact never learn who the real requester was — they see only the proxy.

That vantage point on *outbound* traffic makes it the natural home for a handful of jobs:

  • Content filtering and access control — a corporate or school proxy can block disallowed categories of sites and enforce an acceptable-use policy for everyone behind it.
  • Caching — if many users request the same external resource, the proxy can cache it and serve subsequent requests locally, saving bandwidth.
  • Anonymity — because the destination only sees the proxy's IP, the clients' real addresses are hidden. This is the mechanism behind privacy proxies and, in a related form, VPNs.
  • Monitoring and logging — a single choke point where all outbound activity can be inspected and recorded.

The unifying theme is that a forward proxy exists to control and conceal the clients. It is the guardian of the people making the requests.

03

The reverse proxy — representing the server

A reverse proxy is the mirror image, deployed in front of one or more servers so that all *inbound* requests hit it first. Clients connect to it believing it is the actual website, and it decides what to do with each request and which backend, if any, should handle it. The real servers are never exposed directly to the internet.

Sitting at the server edge makes it the ideal place for a different set of jobs:

  • Load balancing — spread incoming requests across a pool of identical backends so no single machine is overwhelmed, and route around any that fail.
  • TLS termination — handle the expensive HTTPS handshake once at the edge, letting backends communicate over simpler plain HTTP internally.
  • Caching and compression — serve cached responses and shrink payloads before they go out, offloading work from the application servers.
  • Security shielding — hide the backends' addresses and topology, and absorb or filter malicious traffic (rate limiting, WAF rules, DDoS mitigation) before it ever reaches application code.

The unifying theme is that a reverse proxy exists to protect and represent the servers. It is the guardian of the machines answering the requests. Nginx, HAProxy and Envoy are the canonical tools, and a CDN is essentially a globally distributed reverse proxy.

04

Who each one protects

The cleanest way to keep the two straight is to ask who benefits from the proxy being there, because they protect opposite parties.

A forward proxy protects and serves the clients and, by extension, the organisation that runs the client network. Its owner wants control over what those users can reach, insight into what they do, and privacy for their identities against the outside world. The servers on the internet are strangers it is reaching out to.

A reverse proxy protects and serves the servers and the organisation that runs them. Its owner wants to keep the backend fleet available, fast and hidden, and to filter hostile traffic before it lands on application code. The clients on the internet are strangers it is receiving requests from.

This 'who does it protect?' framing resolves almost every real-world example instantly. A company gateway that stops employees reaching social media is guarding the clients, so it is a forward proxy. Nginx sitting in front of a web app spreading load and terminating TLS is guarding the servers, so it is a reverse proxy. Same category of technology, opposite allegiance.

05

Where they show up, and how they combine

In the real world you meet both constantly, often without noticing. Forward proxies appear as corporate web gateways, school and library content filters, privacy and anonymising proxies, and — in spirit — the outbound side of a VPN. Reverse proxies appear at the front of essentially every production website: an Nginx or Envoy tier, a cloud load balancer, an API gateway, and CDNs like Cloudflare or Fastly, which are reverse proxies distributed across the globe.

Crucially, they are not alternatives to choose between — a full request can pass through both. Imagine an employee loading a website: their request first goes out through their company's forward proxy (which checks policy and logs the visit), crosses the internet, and arrives at the website's reverse proxy (which terminates TLS and load-balances it to a backend). Two proxies, each doing its own job, at opposite ends of the same request.

The practical guidance is to always name the proxy by the side it serves. When you are configuring or debugging one, fix in your mind whether you are trying to control *outbound* traffic on behalf of clients or manage *inbound* traffic on behalf of servers — that framing tells you which tool you need, which features apply, and where in the path it belongs.

what to remember
01

A forward proxy represents the CLIENT: outbound requests flow through it to the internet.

02

A reverse proxy represents the SERVER: inbound requests flow through it to the backends.

03

Forward proxy jobs: content filtering, access control, caching, and hiding client identity (anonymity).

04

Reverse proxy jobs: load balancing, TLS termination, caching, compression, and security shielding.

05

Rule of thumb: a forward proxy protects and represents the clients; a reverse proxy protects and represents the servers.

check yourself

Quick quiz

Q1

A school routes all student browsing through a single server that blocks certain sites and hides student IPs. What is that server?

Q2

Which task is a classic job for a reverse proxy, not a forward proxy?

follow the thread