Computer network - Network Architecture and Types

Network Topologies and Types Understanding Network Topology A network topology describes the logical interconnection pattern of nodes (computers, devices, switches, etc.) in a network. Think of it as the "map" showing how devices connect to each other. Topology is crucial because it directly affects two critical network properties: Throughput: How much data can flow through the network Reliability: How well the network continues functioning if a link or node fails The general principle is that more interconnections increase robustness—if one path fails, data can take alternative routes. However, this reliability comes at a cost: more interconnections mean more cables, switches, and installation expenses. Bus Topology In a bus topology, all nodes connect to a single, shared transmission medium—imagine a long cable running through your network with computers tapping into it along the way. Key characteristics: Simple and inexpensive to install All nodes share the same medium, so they must take turns transmitting (only one can send at a time) If the bus cable is damaged, the entire network fails (not reliable) Limited scalability—as you add more nodes, available bandwidth per node decreases Bus topology was historically common in early Ethernet networks but is rarely used in modern networks because of its poor reliability. Star Topology In a star topology, every single node connects to a central device—typically a network switch or wireless access point. The central device acts as a hub that routes traffic between nodes. Key characteristics: If any single node's connection fails, only that node is affected; the rest of the network continues working More reliable than bus topology The central device becomes a potential bottleneck (if it fails, the entire network fails) Requires more cabling than bus topology (you need a cable from every node to the center) Most common topology in modern Ethernet networks Star topology offers a good balance: it's reasonably reliable, scalable, and relatively simple to manage. Ring Topology In a ring topology, each node connects to exactly two neighbors—its immediate left and right—forming a closed loop. Data flows around the ring in one (or sometimes both) directions. Key characteristics: Each node only needs two connections, so cabling is moderate Data travels a predictable path, making it easy to monitor If any single connection breaks, the entire ring is broken (poor reliability) Data must traverse through intermediate nodes to reach distant nodes, potentially creating delays Historically used in token ring networks, but rarely used today Mesh Topology In a mesh topology, each node connects to multiple neighboring nodes, creating numerous paths between any two points. A fully meshed network has every node connected to every other node. Key characteristics: Highly redundant: multiple paths exist between any two nodes, so the network can survive multiple failures Very reliable, but expensive in terms of cabling and equipment More complex to manage and configure Partially meshed networks (where not every node connects to every other) offer a compromise between cost and reliability Mesh topologies are often used in important backbone networks or wireless networks where reliability is critical. Tree and Hierarchical Topologies A tree topology arranges nodes hierarchically, like an organizational chart. Parent nodes connect to multiple child nodes, which may themselves have children. This creates a branching structure. Key characteristics: Scalable: you can easily add branches Breaks the network into logical segments at each level Common in large Ethernet networks with multiple switches A link failure in an upper-level branch affects all nodes below it Most large real-world networks use a tree topology because it balances scalability, manageable complexity, and reasonable reliability. Physical vs. Logical Topology Here's an important distinction that confuses many students: physical topology (the actual physical layout of cables and devices) can differ from logical topology (how data actually flows). For example, a network might be physically wired in a star configuration—all computers plugged into a central switch—but the switch might internally route traffic as a logical ring or bus. The devices don't "know" about this difference; they just send and receive data according to the logical arrangement. This flexibility is one reason star topology is so popular: you get the physical simplicity of star wiring, but the switch can implement whatever logical topology is most efficient. Overlay Networks An overlay network is a virtual network built on top of another physical network. Think of it as a layer of abstraction above the underlying infrastructure. How they work: Nodes in an overlay network are connected by logical links Each logical link may actually traverse multiple physical links and devices in the underlying network The overlay network hides the underlying physical topology from its users Real-world examples: Peer-to-peer (P2P) networks: BitTorrent, blockchain networks, and file-sharing applications create logical connections between users, even though data physically travels through the Internet The Internet itself: It's an overlay on top of physical fiber-optic cables, copper telephone lines, and wireless links Content Delivery Networks (CDNs): Create virtual networks for efficient content distribution Advantages: Services can be added without modifying underlying infrastructure Provide specialized services like quality-of-service routing or distributed hash tables Allow flexibility: nodes can be added/removed without reconfiguring physical hardware Overlay networks are fundamental to modern internet services and represent a powerful abstraction principle in networking. Geographic Scale of Networks Networks are also classified by the physical distance they span. Understanding these categories helps you recognize what technologies and design considerations apply to different network sizes. Personal Area Networks (PANs) A Personal Area Network connects devices very close to a single person, typically within about 10 meters (roughly 30 feet). Common technologies: Wired: USB, which is both a data and power connection Wireless: Bluetooth (phones, headsets, smartwatches), infrared Examples: A smartphone communicating with wireless earbuds, a laptop syncing with a printer via Bluetooth. PANs require minimal infrastructure and are designed for convenience and low power consumption. Local Area Networks (LANs) A Local Area Network connects computers and devices within a limited geographic area—a home, office building, school, or campus. Characteristics: Typically uses wired Ethernet technology for stationary devices Can span a few hundred meters (large buildings) to several kilometers (large campuses) Supports very high data rates: modern Ethernet LANs exceed 100 gigabits per second Fast response times because of short distances Usually owned and operated by a single organization Multiple LANs can be connected through routers to form larger networks Home Area Network (HAN): A residential type of LAN connecting personal computers, printers, smart devices, and media centers. Often shares a single broadband Internet connection. Ethernet is the dominant LAN technology because it's reliable, standardized, and cost-effective. Storage Area Networks (SANs) A Storage Area Network is specialized: it provides block-level access to data storage devices (disk arrays, tape libraries, optical storage) over a dedicated network. Key concept: SANs make centralized storage appear to servers as if it's locally attached, even though it's physically distant. Characteristics: Usually separate from general data networks with their own equipment and switches Not accessible through the regular LAN Optimized for fast, reliable data access to shared storage Common in data centers where multiple servers need to share storage This separation allows organizations to manage storage independently and prevent storage traffic from interfering with regular network traffic. Types of Networks by Function and Scale Campus Area Network (CAN) A Campus Area Network interconnects multiple local area networks within a limited geographical area—typically a university campus, corporate office park, or military base. Characteristics: Larger than a single LAN but smaller than a WAN Interconnects multiple buildings or facilities using dedicated high-speed links Usually owned by a single organization Often uses Ethernet as the backbone connecting multiple LANs Think of it as: many LANs (one per building) connected together to form a campus-wide network. Backbone Network A backbone network serves as the high-capacity central "spine" through which different local area networks or subnetworks exchange information. Key principles: Designed with greater capacity than the networks it interconnects to prevent congestion Acts as the main highway, with LANs as side roads Must be carefully designed considering performance, congestion points, and future scalability The Internet Backbone: A global system of fiber-optic cables and optical networking equipment that carries the bulk of Internet traffic between major network hubs. This is the "backbone of backbones." Backbone networks are critical infrastructure—their design directly impacts the entire network's performance. Metropolitan Area Network (MAN) A Metropolitan Area Network interconnects users and resources across a city or large town—larger than a LAN but smaller than a WAN. Characteristics: Typically spans several kilometers to tens of kilometers Often uses fiber-optic or high-speed Ethernet links May be operated by a single organization or a service provider Commonly used to connect multiple office buildings in a downtown area MANs fill the gap between local area networks (single building) and wide area networks (global). Wide Area Network (WAN) A Wide Area Network covers very large geographic areas: a city, country, or intercontinental distances. The Internet is the largest WAN. Characteristics: Uses diverse transmission media: telephone lines, coaxial cables, fiber-optic cables, satellite links, and wireless Often relies on transmission services from common carriers (telephone companies, ISPs) Operates primarily at the physical layer, data link layer, and network layer of the OSI model Data rates are typically lower than LANs due to distance and cost More complex routing and addressing schemes Key point: WANs connect geographically distant networks and are the backbone of global communications. <extrainfo> Enterprise Private Network An enterprise private network is built by a single organization to interconnect all its facilities: offices, manufacturing plants, headquarters, remote sites, and retail locations. Purpose: Share computer resources, data, and applications securely across the entire organization. Implementation: May use combinations of LANs, WANs, and private leased lines to create a unified network infrastructure. </extrainfo> Virtual Private Network (VPN) A Virtual Private Network creates a secure, private overlay network where some connections are carried over open/public networks like the Internet. How it works: Data link layer protocols are tunneled (encapsulated) through the underlying public network Uses encryption and authentication to create a "private" connection over public infrastructure Allows remote employees to securely access company networks as if they were in the office Important clarification: Not all VPNs provide encryption or authentication—some only provide tunneling. Security depends on the specific VPN implementation. VPNs are cost-effective because you leverage existing public infrastructure (the Internet) rather than building private dedicated lines. Organizational Scope of Networks Networks can also be classified by who controls them and what access they provide. Intranet An intranet is a set of networks under the control of a single administrative entity (usually an organization). Characteristics: Uses standard Internet protocols (TCP/IP) and applications (web browsers, email, file transfer) Internal to the organization; not accessible from the public Internet Often combines LANs and WANs under unified administration Can scale to thousands of computers and devices Example: A corporation's internal network where employees access email, shared drives, and internal websites. Intranets provide organizational benefits by using familiar, standardized technologies while maintaining internal control and security. <extrainfo> Extranet An extranet extends an intranet's resources to external partners, customers, or suppliers while maintaining separate security controls. Characteristics: Still controlled by a single organization, but allows limited access to outsiders Creates a middle ground between a private intranet and a public internet Often implemented using WAN technologies Requires careful access control and authentication Example: A supplier portal where external vendors can submit orders and check inventory. </extrainfo> Internet The Internet is the largest internetwork in the world, connecting governmental, academic, corporate, public, and private networks globally. Fundamental characteristics: Built on the Internet Protocol (IP) suite—a standardized set of communication rules Uses global backbones of copper and optical fiber cables Enables diverse services: the World Wide Web, email, video streaming, the Internet of Things, cloud computing, and countless others Decentralized: no single entity controls it; instead, network operators exchange routing information How routing works: Network operators use the Border Gateway Protocol (BGP) to communicate reachability information. This creates a redundant worldwide mesh of transmission paths, allowing data to flow around congestion and failures. Key insight: The Internet's strength comes from its redundancy and decentralization. Even if major portions fail, alternate routes allow communication to continue.