Server types you should know
Nine specialised servers, and the one job each of them exists to do.
'Server' is a role, not a box — a program that answers requests for a specific job. Learn the common specialisations: web and application servers handle HTTP and business logic, proxies and load balancers sit in front to route and spread traffic, DNS resolves names, and database, cache, mail and FTP servers each own one kind of resource. Real systems chain many of them together.
Web server
A web server listens for HTTP/HTTPS requests and returns web content — historically static files like HTML, CSS, images and JavaScript. Nginx and Apache are the classics. It's the front door for browser traffic and often also acts as a reverse proxy passing dynamic requests deeper into the system.
Application server
An application server runs your actual program logic — executing code, querying databases, and generating dynamic responses the web server can't produce on its own. Think of a Node, Java or Python app process behind the web tier. The web server handles the protocol and static assets; the application server does the thinking.
Proxy / reverse proxy
A proxy is an intermediary that forwards requests on someone's behalf. A forward proxy sits in front of clients (filtering or anonymising outbound traffic); a reverse proxy sits in front of servers, hiding them, terminating TLS and caching. It's the general pattern the next two servers specialise.
Load balancer
A load balancer is a specialised reverse proxy whose job is to spread incoming requests across a pool of identical backend servers. It picks targets by round-robin, least-connections or hashing, and health-checks the pool so it can route around any instance that fails. This is what makes horizontal scaling and high availability real.
DNS server
A DNS server translates human-readable names like example.com into the IP addresses machines actually route to. It's the phone book of the internet, and every other request implicitly depends on it resolving first. Without DNS, clients would have no way to find any of the servers above.
Database & cache servers
A database server stores and serves persistent, structured data durably, answering queries and enforcing consistency (PostgreSQL, MySQL). A cache server keeps hot data in memory for microsecond reads (Redis, Memcached), absorbing repeated lookups so the database isn't overwhelmed. One owns the source of truth; the other owns speed.
Mail & FTP servers
A mail server sends, receives and relays email using protocols like SMTP, IMAP and POP3. An FTP server transfers files between machines over FTP/SFTP. Both are specialised servers that own one specific kind of exchange, and knowing they exist rounds out the map of what 'a server' can mean.
A server is a role, not a machine
The first thing to unlearn is the image of "a server" as a specific humming box in a rack. In systems terms, a server is a role: a piece of software that listens for requests and responds to them, providing some resource or service to *clients* that ask. The counterpart is the client, which initiates the request. This is the client–server model, and it's the backbone of almost all networked computing.
Because it's a role, one physical machine can host many servers at once, and one logical server can be spread across many machines. What distinguishes the "types" below isn't hardware — it's the *job* each one specialises in and the *protocol* it speaks. Learning them is really learning a vocabulary: when someone says "put a cache server in front of the database" or "the reverse proxy terminates TLS," you should immediately picture which job is being done and where in the request's journey it sits. Almost every real system is several of these roles chained together, each handling the one thing it's best at.
Serving the web: web and application servers
The two most familiar roles sit at the front of a web system, and beginners constantly conflate them.
A web server listens for HTTP and HTTPS requests and returns web content. Classically its job is serving *static* resources — HTML files, CSS, JavaScript bundles, images — straight from disk, as fast as possible. Nginx and Apache are the archetypes. A web server is optimised for handling huge numbers of concurrent connections and shuffling bytes efficiently; it is not where your application's logic lives.
An application server is where that logic *does* live. It runs your actual program — a Node.js, Java, or Python process — executing business rules, talking to databases, and generating *dynamic* responses that can't just be read off disk. The division of labour is clean: a request for a static logo is answered by the web server directly, while a request to "check out my cart" is passed back to the application server, which computes the answer. In practice the web server usually doubles as a reverse proxy, fronting the application server, absorbing static traffic and forwarding only the dynamic requests inward.
Directing traffic: proxies and load balancers
Between clients and your servers sit roles whose job is *routing* rather than producing content.
A proxy is an intermediary that forwards requests on someone's behalf. The direction matters. A forward proxy sits in front of *clients* — a corporate web filter or a privacy proxy that mediates outbound traffic. A reverse proxy sits in front of *servers*, so the outside world talks only to the proxy and never learns the real backend addresses; along the way it can terminate TLS, cache responses, compress payloads and present a single clean hostname.
A load balancer is a reverse proxy specialised for one purpose: spreading incoming requests across a pool of identical backend servers. It chooses a target using an algorithm — round-robin, least-connections, or consistent hashing — and continuously health-checks the pool, quietly steering traffic away from any instance that has failed. This is the component that turns a fleet of interchangeable servers into a single service that scales horizontally and survives individual crashes without anyone noticing.
Finding things and storing things: DNS, database and cache
Three more roles handle naming and data.
A DNS server is the phone book of the internet. It translates human-readable names like example.com into the numeric IP addresses machines actually route to. Every other interaction implicitly depends on it: before a browser can reach a web server, it must resolve the name to an address, so DNS is the silent first step of nearly every request.
A database server owns durable, structured data. It stores information persistently, answers queries, and enforces consistency and integrity — PostgreSQL and MySQL are canonical examples. It is the system's source of truth, and precisely because durability is hard, it's usually the most carefully protected tier.
A cache server owns *speed*. It keeps frequently-accessed data in memory so reads take microseconds instead of milliseconds — Redis and Memcached are the standards. By absorbing repeated lookups, the cache shields the database from the bulk of read traffic. The two are partners: the database guarantees the data is correct and never lost, while the cache makes reading it fast.
Specialised exchanges: mail and FTP, and the bigger picture
Two older but still-common roles round out the map, each owning a specific kind of exchange over its own protocol.
A mail server sends, receives and relays email. It's really a small family of servers speaking different protocols: SMTP to send and relay messages between servers, and IMAP or POP3 to let clients fetch mail from a mailbox. When you "send an email," a chain of mail servers routes it to the recipient's mailbox.
An FTP server handles file transfer between machines using the File Transfer Protocol, or its secure variants FTPS and SFTP. It exists to move whole files around — the sort of bulk transfer that predates the web and still underpins plenty of data exchange between organisations.
Step back and the value of this vocabulary is the *composition*. A single user action might resolve a name at a DNS server, hit a load balancer, pass through a reverse proxy and web server, reach an application server that consults a cache server and then a database server, and separately trigger a mail server to send a confirmation. None of these is exotic — each is just a program doing one well-defined job — and understanding a system means being able to name which role owns which responsibility along that path.
A 'server' is a role: software that listens for and responds to requests for some resource or service.
Web server serves HTTP content; application server runs the business logic behind it.
Proxy/reverse proxy and load balancer sit in front to hide, route and spread traffic across backends.
DNS server resolves human names to IP addresses so everything else can be reached.
Database and cache servers own persistent and fast-access data respectively.
Mail and FTP servers handle email delivery and file transfer over their own protocols.
Quick quiz
What's the core difference between a web server and an application server?
A load balancer is best described as…
Load balancer vs API gateway vs reverse proxy
Three boxes that sit in front of your servers — and do very different jobs.
How DNS works
Turning a human name into a machine address by walking a global hierarchy.
Forward proxy vs reverse proxy
Two middlemen with mirror-image jobs: one fronts the clients, one fronts the servers.
Why Nginx is so popular
One event loop that quietly handles tens of thousands of connections at once.