Servers

A server is a computer or program whose job is to provide a service to other devices (clients). “Server” describes the role, not the machine: one physical box can run a web server, a mail server and a file server at once. Different services have different needs, so for each server below think about three qualities: scalability (can it cope with more users?), reliability (what if it fails?) and security (who could abuse it?).

A web server stores a site's files (HTML, CSS, JavaScript, images) and sends them to browsers when asked, over HTTP or HTTPS. For dynamic pages it runs server-side code (Python, PHP, Node.js) to build the page first. Popular examples are Apache, Nginx and IIS. Scaling: several servers behind a reverse proxy. Reliability: if the one server fails, every site on it goes dark, so busy sites keep spares. Security: HTTPS, patching, and limiting what each user can reach.

People remember names; networks need numbers. A DNS (domain name system) server turns a name such as www.example.com into an IP address. No single server knows everything, so it works as a hierarchy:

1. Your device asks the local DNS server (often your provider's). If it has the answer in its cache, it replies and that is the end.
2. If not, it asks a root server, which knows who handles each ending (.com, .org, .uk) and points to the right top-level domain server.
3. The TLD server points to the authoritative server for that domain, which holds the real answer.
4. The answer comes back, and the local server caches it for its time to live, so the next person asking is answered instantly.

DNS scales because the data is spread over thousands of servers and heavily cached; it is made reliable by having several servers for every zone. It is also a prime target for attack, so key servers are well protected.

The DHCP server (A2.1.4) keeps a pool of addresses and leases one, with the subnet mask, gateway and DNS address, to each device that joins. It is often the home router itself. If it fails, new devices cannot get an address, so large networks run a backup DHCP server.

A file server is shared storage on the network. Users see it as a drive or folder on their own computer (using protocols such as SMB or NFS). It scales by adding drives and faster network cards, stays reliable with RAID and backups, and stays secure with read and write permissions on each folder.

A mail server acts as a digital post office: it accepts outgoing mail with SMTP, passes it to other mail servers, and stores incoming mail until you collect it (with POP3 or IMAP). It must cope with large attachments, filter spam and phishing, and be backed up, because people treat email as a permanent record.

A proxy server sits between clients and the internet. Requests go to the proxy first, which can filter them (block sites, require login), cache pages so repeat requests are fast and save bandwidth, log activity, and hide the clients' addresses, since the website only sees the proxy. A reverse proxy turns that around: it sits in front of the servers, shares incoming requests among several of them (load balancing), caches common responses and shields the real servers from the open internet.