Chapter 1

Network operating systems

Last updated November 20264 sectionsabout 9 min read

Key takeaways

A network operating system manages and coordinates resources across a network, extending the single-machine work of managing processes, memory and files to many connected computers at once.
A server provides resources or services to other computers, a terminal is the device or interface used to reach it, and a directory is the central store of information about resources, users and devices.
Server roles give a server a specific function such as file sharing, authentication or name resolution; one server can run several roles, or a large organisation can dedicate a server to each.
The networking features a NOS provides are multiprocessing, resource sharing, backup and web services, internetworking, network security and directory services.
Client-server centralises administration and security and scales to thousands of users; peer-to-peer needs no dedicated server but is practical only for a handful of machines.

A network operating system (NOS) is an operating system specifically built to manage and coordinate resources across a network, rather than for a single standalone machine, extending the general operating-system functions covered in Introduction to Computing System, managing processes, memory and files, to manage them across many connected computers at once. This chapter defines the vocabulary this paper and the two after it are built on, the features a NOS provides, and the two fundamental architectures a network can be built around.

Key terms

Server

A computer that provides resources or services (files, authentication, DNS, web content) to other computers on a network, typically running a network operating system and dedicated hardware built for continuous operation.

Terminal

A device or interface used to access and interact with a server or network system, historically a simple screen-and-keyboard device with no processing of its own, now more often a full client computer or a remote-desktop session acting the same role.

Server roles

Server roles assign a server a specific function on the network, such as file sharing, authentication, or name resolution, rather than a server being a single undifferentiated machine. One physical server can run several roles at once (common in a small organisation) or a large organisation can dedicate a separate server to each role for performance and resilience. Each role named here is covered in full, with its own configuration steps, in Windows Server Roles and Features.

Features of networking a NOS provides

Multiprocessing

Using more than one processor (or core) to handle multiple tasks in parallel, important for a server handling many simultaneous requests from client machines rather than one user's own work.

Resource sharing

Making files, printers and storage centrally accessible to network users, the most immediately visible benefit of a NOS to an ordinary user.

Backup and web services

Built-in support for scheduled backups of server and client data, and for hosting web-accessible resources such as an intranet site or a portal.

Internetworking

Connecting and coordinating across multiple interconnected networks, so a NOS's services can be reached from, and can reach, more than one local network.

Network security

Authentication, authorisation and access control enforced centrally by the NOS rather than configured separately, and inconsistently, on every individual machine.

Directory services

A centralised store of user, computer and resource information, the backbone of centralised administration: a password changed once in the directory takes effect everywhere that account is used.

Types of network operating system

Every network is built on one of two fundamental architectures, and the choice shapes almost everything else about how it is administered, secured and scaled.

Client-server systems

Dedicated servers provide services (files, authentication, printing) to many client machines. Centralised, scalable, and the dominant model for organisations of any real size, because administration, security policy and backup are all applied in one place rather than machine by machine.

Peer-to-peer network operating system

Every computer can act as both client and server, sharing its own resources directly with others, with no dedicated central server. Simpler and cheaper to set up, needing no server hardware or licence, but harder to secure (each machine manages its own accounts and shares) and harder to scale, since there is no single place to apply a policy change across every machine.

Aspect	Client-server	Peer-to-peer
Administration	Centralised, one place to manage users and policy	Distributed, each machine managed separately
Security	Consistent, centrally enforced	Inconsistent, depends on each user
Cost	Server hardware and licensing required	No dedicated server needed
Scalability	Scales to hundreds or thousands of users	Practical only for a handful of machines
Typical use	Any organisation with more than a few staff	A home network or a very small office

Choosing an architecture for a growing veterinary practice

A veterinary practice starts with three computers sharing files directly between them (peer-to-peer): the receptionist's machine, the vet's machine, and one in the back office, each sharing its own folder with the others.

As the practice grows to twelve computers across two branches, keeping every user account and shared folder synchronised by hand across twelve separately managed machines becomes unreliable: a password changed on one machine does not change anywhere else, and a leaver's access must be removed from every machine individually or it is missed on one.

Moving to a client-server architecture, one server running Active Directory, file services and centralised backup, fixes both problems: every account, permission and password lives in one directory, changed once and effective everywhere, and new machines simply join the domain rather than needing their own local account setup.

The point at which peer-to-peer stops being practical is not a fixed number of machines but the point at which inconsistent, unmanaged accounts and shares become a real security and administration risk, which for most organisations arrives well before twelve computers.

Self-check

(a) Define a network operating system and explain how it extends a single-machine operating system's functions. (3 marks)

(b) Distinguish a server, a terminal and a directory. (3 marks)

(d) Distinguish a client-server architecture from a peer-to-peer architecture on administration and scalability. (4 marks)

(e) A small office of four staff wants the cheapest way to share one folder and one printer, with no plans to grow. Recommend an architecture and justify it. (2 marks)

(f) Explain why a password changed in a directory service takes effect on every machine at once, and why the same change is unreliable on a peer-to-peer network. (2 marks)

Show worked answer

(a) A network operating system is an operating system built to manage and coordinate resources across a network rather than for one standalone machine. It extends the single-machine functions (managing processes, memory and files) to manage them across many connected computers, adding centralised authentication, resource sharing and network security.

(b) A server is a computer that provides resources or services to other computers on a network. A terminal is a device or interface used to access and interact with a server or network system. A directory is a structured store of information about network resources, users and devices.

(c) Multiprocessing: handles many simultaneous client requests in parallel. Resource sharing: makes files, printers and storage centrally accessible. Network security: enforces authentication and access control centrally and consistently. Directory services: holds user and resource information in one place so a change takes effect everywhere. Any four.

(d) Client-server centralises administration in one directory and one set of policies applied everywhere, and scales to hundreds or thousands of users; peer-to-peer distributes administration to each machine individually, which is unreliable to keep consistent and practical only for a handful of machines.

(e) Peer-to-peer, because with only four staff and no growth planned the cost and administrative overhead of a dedicated server is unnecessary; each machine can share the folder and printer directly.

(f) A directory service holds the account centrally, so every machine checks the same single record when the user logs in, and a change to that one record is immediately what every machine sees. On a peer-to-peer network each machine holds its own separate copy of accounts, so a password changed on one machine has no effect on the separate copies stored on the others.

Exam tip

Client-server versus peer-to-peer is the foundational distinction this whole paper builds on: client-server centralises administration and security, the model this diploma's later Windows Server papers assume, while peer-to-peer trades that centralisation for simplicity, suited only to very small networks. Justify a recommendation against the specific organisation's size and growth plans, not architecture alone.