A computer network is a system of interconnected computing devices—ranging from traditional to cloud-based environments—that communicate and share resources with one another.
Networking, or computer networking, involves connecting two or more computing devices (for example, desktop computers, laptops, mobile devices, routers, applications) to enable the transmission and exchange of information and resources.
Networked devices rely on communication protocols—rules that describe how to transmit or exchange data across a network—allowing them to share information over physical or wireless connections.
Computer networks form the backbone of nearly every digital experience—from personal communication and entertainment to cloud-native business operations and global infrastructure. Designed for scalability, speed and IT security, today’s networks support dynamic data flows across both on-premises systems and virtualized cloud environments.
Industry newsletter
Stay up to date on the most important—and intriguing—industry trends on AI, automation, data and beyond with the Think newsletter. See the IBM Privacy Statement.
Your subscription will be delivered in English. You will find an unsubscribe link in every newsletter. You can manage your subscriptions or unsubscribe here. Refer to our IBM Privacy Statement for more information.
Before contemporary networking practices, computer science engineers would have to physically move computers to share data between devices, which was an arduous task at a time when computers were large and unwieldy.
To simplify the process (especially for government workers), the Department of Defense funded the creation of the first functioning computer network (eventually named ARPANET) in the late 1960s. This milestone laid the groundwork not only for the internet but also for cloud networking, which today supports globally distributed infrastructures and application services.
Since then, networking practices and the computer systems that drive them have evolved tremendously. Today’s computer networks facilitate large-scale interdevice communication for every business, entertainment and research purpose. The internet, online search, email, audio and video sharing, online commerce, live-streaming and social media all exist because of advancements in computer networking.
In enterprise settings, this progress caused more flexible networking models centered on cloud infrastructure. Organizations increasingly rely on hybrid cloud and multicloud networking strategies, where applications and data flow seamlessly between on-premises infrastructure and cloud environments provided by cloud service providers. Some well-known providers are AWS, Microsoft Azure, IBM Cloud® and Google Cloud Platform. This cloud-first networking strategy enables businesses to scale resources dynamically, reduce infrastructure costs and access advanced services without maintaining physical hardware.
Today, artificial intelligence (AI) and machine learning (ML) are further transforming networking by enabling smarter, more adaptive systems. These technologies help automate network management, enhance security through anomaly detection, and optimize performance by predicting and responding to traffic patterns in real time.
Using email as a basic example, here’s how data moves through a network.
When a user wants to send an email, they first write the email and then press the “send” button. When the user presses “send,” an SMTP or POP3 protocol uses the sender’s wifi to direct the message from the sender node through the network switches. Here it is compressed and broken down into smaller and smaller segments (and ultimately into bits, or strings of ones and zeros).
Network gateways direct the bit stream to the recipient’s network, converting data and communication protocols as needed. When the bit stream reaches the recipient’s computer, the same protocols direct the email data through the network switches on the receiver’s network. In the process, the network reconstructs the original message until the email arrives in human-readable form in the recipient’s inbox (the receiver node).
To fully understand computer networking, it is essential to review networking components and their functionality, including:
Switches: A switch is a device that connects network devices and manages node-to-node communication across a network, ensuring that data packets reach their intended destinations. Unlike routers, which send information between networks, switches send information between nodes within a network.
Therefore, “switching” refers to how data is transferred between devices on a network. Networks rely on three main types of switching:
Circuit switching establishes a dedicated data communication path between nodes in a network so that no other traffic can traverse the same path. Circuit switching ensures that full bandwidth is available during every transmission.
Message switching sends whole messages from the source node to the destination node, with the message traveling from switch to switch until it reaches the destination.
Packet switching involves breaking down data into independent components to make data transmission less demanding of network resources. With packet switching, packets, instead of entire data streams, travel through the network to their end destination.
While traditional networking components (for example, routers, switches, ports, gateways) remain foundational to network operations, cloud environments have transformed how these elements are deployed and managed.
In cloud settings, many of these traditional components are virtualized and offered as managed services, enabling organizations to build robust network infrastructures without the need to maintain physical hardware. Cloud providers abstract the underlying complexity while still relying on the same fundamental networking principles—but with enhanced scalability, flexibility and global reach.
Modern networking increasingly relies on cloud-native components that extend and enhance traditional networking capabilities. These components include:
Typically, geographical areas define computer networks. A local area network (LAN) connects computers within a defined physical space, while a wide area network (WAN) can connect computers across continents. However, networks are also defined by the protocols they use to communicate, the physical arrangement of their components, how they manage network traffic and the purpose they serve in their respective environments.
The most common and widely used computer network types fall into three broad categories:
The network types in this category are distinguished by the geographical area the network covers.
A LAN connects computers over a relatively short distance, such as those within an office building, school or hospital. LANs are typically privately owned and managed.
As the name implies, a WAN connects computers across large geographical areas, such as regions and continents. WANs often have collective or distributed ownership models for network management purposes.
A cloud network is a type of wide-area network (WAN) where networking resources—such as routers, firewalls and switches—are virtualized and delivered through public or private cloud platforms. Unlike traditional WANs, which depend heavily on physical infrastructure, cloud networks offer on-demand scalability, automation and global availability. These networks are ideal for businesses running applications in hybrid or multicloud environments because they allow seamless and secure connectivity between cloud services and on-premises systems.
A software-defined wide area network (SD-WAN) is a virtualized WAN architecture that uses SDN principles to centralize the management of disconnected WAN networks and optimize network performance. SD-WAN enables an organization to share data and applications across branch offices, remote workers and authorized devices that span vast geographical distances and multiple telecommunications infrastructures.
MANs are larger than LANs but smaller than WANs. Cities and government entities typically own and manage MANs.
A PAN serves one person. If a user has multiple devices from the same manufacturer (an iPhone and a MacBook, for instance), it’s likely they've set up a PAN. The PAN shares and syncs content, text messages, emails, photos and more, across devices.
Network nodes can send and receive messages by using either wired or wireless links (connections).
Wired network devices are connected by physical wires and cables, including copper wires and Ethernet, twisted pair, coaxial or fiber-optic cables. Network size and speed requirements typically dictate the choice of cable, the arrangement of network elements and the physical distance between devices.
Wireless networks eliminate the need for cables by using infrared, radio or electromagnetic-wave transmission across wireless devices equipped with built-in antennae and sensors.
Computing networks can transmit data by using a range of transmission dynamics, including:
In a multipoint network, multiple devices share channel capacity and network links.
Network devices establish a direct node-to-node link to transmit data.
On broadcast networks, several interested “parties” (devices) can receive one-way transmissions from a single sending device. Television stations and radio stations rely on broadcast networks.
A VPN is a secure, point-to-point connection between two network endpoints. It establishes an encrypted channel that keeps a user’s identity and access credentials, as well as any data transferred, inaccessible to hackers.
Computer network architecture establishes the theoretical framework for a computer network, encompassing design principles and communication protocols.
The primary types of network architectures include:
In a P2P architecture, two or more computers are connected as “peers,” meaning they have equal power and privileges on the network. A P2P network doesn’t require a central server for coordination. Instead, each computer on the network acts as both a client (a computer that needs to access a service) and a server (a computer that provides services to clients).
Every peer on the network makes some of its resources available to other network devices, sharing storage, memory, bandwidth and processing power across the network.
Within a research-intensive organization, for instance, team members might use a decentralized file-sharing system to exchange large datasets directly between their workstations, eliminating the need for a central server.
In a client-server network, a central server (or group of servers) manages resources and delivers services to client devices on the network. Clients in this architecture don’t share their resources and interact only through the server. Client-server architectures are often referred to as tiered architectures due to their multiple layers.
For example, in a corporate environment that uses a client-server architecture, employees (clients) often have access to a central human resources system (server). This server allows them to manage personal data, submit leave requests and view internal documents.
Hybrid architectures incorporate elements of both the P2P and client-server models. Many businesses require both centralized services (such as user authentication) and peer-to-peer capabilities (such as local file sharing) to optimize performance and resource use.
Whereas architecture represents the theoretical framework of a network, topology refers to the practical implementation of that framework. Network topology describes the physical and logical arrangement of nodes and links on a network. It includes all hardware (for example, routers, switches, cables), software (for example, apps, operating systems) and transmission media (for example, wired, wireless connections).
Common network topologies include:
In a bus topology, every network node is directly connected to a main cable.
In a ring topology, nodes are connected in a loop, so each device has exactly two neighbors. Adjacent pairs are connected directly, and nonadjacent pairs are connected indirectly through intermediary nodes.
Star network topologies feature a single, central hub through which all nodes are indirectly connected.
Mesh topologies are more complex, defined by overlapping connections between nodes. There are two types of mesh networks, full mesh and partial mesh.
In a full mesh topology, every network node connects to every other network node, providing the highest level of network resilience. In a partial mesh topology, only some network nodes connect, typically those nodes that exchange data most frequently.
Full mesh topologies can be expensive and time-consuming to run, which is why they’re often reserved for networks that require high redundancy. However, partial mesh provides less redundancy but is more cost-effective and simpler to run.
Regardless of subtype, mesh networks have self-configuration and self-organization capabilities, and they automate the routing process, so the network finds the fastest, most reliable data path.
Whether it’s the Internet Protocol (IP) suite, Ethernet, wireless LAN (WLAN) or cellular communication standards, all computer networks follow communication protocols. These protocols are sets of rules that every node on the network must follow to share and receive data.
The IEEE (Institute of Electrical and Electronics Engineers)—which sets global standards for networking technologies—develops and manages many of these protocols, including Ethernet (IEEE 802.3) and wifi (IEEE 802.11). Network protocols also rely on gateways to enable incompatible devices to communicate (a Windows computer attempting to access Linux servers, for instance).
Many modern networks run on TCP/IP models, which include four network layers:
While TCP/IP is the protocol suite used in most networks today, the Open Systems Interconnection (OSI) model is a standardized framework that defines how data moves through a network in seven layers.
Each layer has a specific role—from sending raw bits over cables at the physical layer to managing user applications at the top layer. This layered approach helps network engineers design, troubleshoot and standardize communication across diverse systems. Although OSI itself is not a set of protocols used in practice, its model remains foundational for understanding how different networking technologies work together.
From global enterprises to everyday users, computer networks underpin virtually every digital experience, connecting devices, data applications and users across the world. In business, they underpin operations, enabling cloud services, real-time collaboration and secure data exchange. Here are some of the most common computer network use cases:
Networking enables every form of digital communication, including email, messaging, file sharing, video calls and streaming. Networking connects all the servers, interfaces and transmission media that make business communication possible.
Without networking, organizations would have to store data in individual data repositories, which is unsustainable in the age of big data. Computer networks help teams keep centralized data stores that serve the entire network, freeing up valuable storage capacity for other tasks.
Common network-based storage solutions include storage area networks (SAN) and network attached storage. SAN offers high-speed block storage, typically used for mission-critical applications like databases and virtualization, while NAS provides file storage accessible over a standard network.
Users, network administrators and developers alike benefit from how networking simplifies resource and knowledge sharing. Networked data is easier to request and fetch, so users and clients get faster responses from network devices. Networked data also provides benefits on the business side, making it easier for teams to collaborate and share information as technologies and enterprises evolve.
AI and algorithms help automate complex tasks, such as network monitoring, traffic analysis, anomaly detection and incident response, manual intervention reduction and overall network security strengthening.
For instance, many organizations in industries like telecommunications, financial services and manufacturing rely on a network operations center (NOC) to constantly monitor and manage network performance, availability and security.
Not only are well-built networking solutions more resilient, but they also offer businesses more options for cybersecurity and network security. Most network providers offer built-in encryption protocols and access controls (such as multifactor authentication) to protect sensitive data and keep bad actors off the network.
IBM Cloud Pak for Network Automation is a Cloud Pak that enables the automation and orchestration of network infrastructure operations.
Cloud networking solutions from IBM provide high-performing connectivity to power your apps and business.
Consolidate datacenter support with IBM Technology Lifecycle Services for cloud networking and more.