Computer Networks And Cybersecurity

A communication system contains a sender (the source of data), a receiver (the destination), a message (the information), a transmission medium (the path), and a protocol (the agreed rules). A network also needs interfaces, such as a Network Interface Card (NIC), to convert computer data into signals. Protocols define such details as addressing, timing, error handling, and the format of transmitted units.

A topology describes how devices are arranged. In a bus topology, all devices share one main cable. It uses little cable, but a break in the backbone can disrupt the entire network and simultaneous transmissions cause collisions. A star topology connects each device to a central switch. It is easy to expand and troubleshoot, although failure of the central switch affects its attached devices. A ring topology gives each device two neighbors; controlled circulation can reduce collisions, but a break may interrupt communication unless a second ring or bypass exists. A mesh topology provides multiple paths. Full mesh connectivity among n devices requires n(n - 1)/2 links, so four devices require 6 links; this gives resilience but becomes expensive quickly. Modern Ethernet local networks commonly use an extended star: access switches connect end devices, and higher-level switches connect the access layer.

Layered Models, Media, and Network Devices

The OSI reference model divides communication into seven layers. The Physical layer carries raw bits as signals. The Data Link layer organizes bits into frames and manages local delivery. The Network layer provides logical addressing and routing. The Transport layer supports end-to-end delivery, reliability, and flow control. The Session layer manages communication sessions. The Presentation layer handles representation, compression, and encryption formats. The Application layer provides network services to applications.

A useful encapsulation view is:

text Application data ↓ Transport segment ↓ Network packet ↓ Data-link frame ↓ Physical bits/signals

At the destination, headers are removed in reverse order. The OSI model is primarily a reference framework: real protocols do not always fit one layer perfectly. The TCP/IP architecture is the practical Internet model. Its Link layer combines OSI Physical and Data Link functions; the Internet layer corresponds mainly to Network; the Transport layer remains similar; and the Application layer combines OSI Session, Presentation, and Application functions.

| OSI layer | Main responsibility | TCP/IP counterpart | Example | |---|---|---|---| | 7 Application | Network services for applications | Application | HTTP, DNS | | 6 Presentation | Format, compression, encryption representation | Application | TLS encoding | | 5 Session | Session control | Application | Session management | | 4 Transport | Process-to-process delivery | Transport | TCP, UDP | | 3 Network | Logical addressing and routing | Internet | IP | | 2 Data Link | Frames and local delivery | Link | Ethernet | | 1 Physical | Signals and media | Link | Fiber, radio |

A twisted-pair cable uses copper wires twisted to reduce electromagnetic interference. Unshielded Twisted Pair (UTP) is common in offices; Category 6 cable can support high-speed Ethernet over typical maximum horizontal runs of 100 meters. Coaxial cable has a central conductor, insulation, shielding, and an outer jacket. It resists interference better than ordinary unshielded copper but is less common for modern local Ethernet. Optical fiber carries light through glass or plastic. It offers high bandwidth, low attenuation, and resistance to electromagnetic interference. Single-mode fiber supports long distances; multimode fiber is economical for shorter building links. Wireless media use radio waves and avoid cabling, but they share spectrum, experience interference, and require careful authentication and encryption.

A hub is a Layer 1 repeater: it copies incoming signals to every port, so all connected devices share one collision domain. A bridge operates at Layer 2 and forwards frames using Media Access Control (MAC) addresses. A Layer 2 switch is a multiport bridge that learns which MAC address is reachable through each port. A Layer 3 switch can also forward using IP addresses, often at high speed inside a campus network. A router connects different IP networks and selects paths using routing information. A gateway is a broader term for a device or software system that connects unlike networks or protocols; a default gateway is commonly the router used to leave a local subnet.

Switching determines how network capacity is shared. Circuit switching reserves a complete path before data transfer. Traditional telephone networks used this approach: bandwidth remains allocated even during silence. It provides predictable delay after setup, but reserved capacity can be wasted. Packet switching divides data into packets, each carrying addressing information. Packets from many users share links statistically. This improves efficiency for bursty computer traffic, but queueing can cause variable delay, loss, or reordering. The Internet uses packet switching because web browsing, streaming, and cloud applications do not continuously require a fixed-rate circuit.

Ask yourself whether a switch should forward traffic to another local host or whether a router is needed. If both hosts are on the same IP subnet, local delivery uses a frame and a switch. If they are on different subnets, the sender forwards the frame to its default gateway, and the router handles the next network-level decision.

Worked Example: Choosing a Design and Tracing Encapsulation

Consider a small laboratory with 24 computers, one server, and an Internet connection. The computers require 1 Gbit/s Ethernet, the server is in the same building, and the uplink to the router must support communication with another IP network.

1. Select the topology. A star topology is chosen because each computer can connect independently to a central switch. A failed cable affects one computer rather than the whole laboratory.

2. Select the medium. Use Category 6 UTP for the 24 access links, provided each run is no longer than 100 meters. Use fiber for a 300-meter building-to-building uplink because copper Ethernet at this distance would exceed the normal horizontal-cabling limit.

3. Select devices. One 24-port Layer 2 switch connects the computers and server. A router connects the laboratory subnet to the other network and the Internet. A hub is rejected because it would repeat every frame and create one shared collision domain.

4. Trace a request. Suppose computer A requests a web page. The application creates an HTTP request. Transport software places it in a TCP segment. IP places that segment in a packet addressed to the web server. Ethernet places the packet in a frame addressed first to the router’s MAC address, because the server is outside the local subnet.

5. Forward locally. The switch examines the destination MAC address. If it has learned the router’s address on port 1, it sends the frame only to port 1. It does not inspect the web request itself.

6. Route onward. The router removes the Ethernet frame, examines the destination IP address, selects an outgoing interface, and creates a new link-layer frame for the next network. The IP packet may remain logically addressed to the web server while the link-layer addresses change at each hop.

7. Compare switching behavior. With packet switching, the laboratory’s traffic shares the uplink with other users. A 2-megabyte file may be divided into many packets and interleaved with other traffic. A circuit-switched design would reserve a path, but that reservation would waste capacity while the laboratory is idle.

Final result: the appropriate design is an extended star using a Layer 2 switch for local delivery, a router for inter-network delivery, Category 6 UTP for short links, and fiber for the 300-meter uplink. Packet switching is more efficient for this bursty data traffic.

The example also shows why layers matter: the switch makes a local frame decision, while the router makes an inter-network packet decision. Confusing these roles often leads to incorrect equipment choices.

Guided Practice and Answer Key

1. A network has six devices and requires a full mesh. How many direct links are needed? Hint: use n(n - 1)/2 and substitute n = 6.

2. A host sends data to a destination on another IP subnet. Should the first Ethernet frame use the destination host’s MAC address or the default gateway’s MAC address? Hint: distinguish local frame delivery from routed packet delivery.

3. Compare a hub and a Layer 2 switch when four computers transmit at the same time. Hint: consider whether each device receives every signal and whether collisions are shared.

4. A fiber link connects two buildings 2 kilometers apart, while copper links connect desks within each building. Why is this combination reasonable? Hint: compare distance, interference, bandwidth, and installation cost.

5. Explain one advantage and one disadvantage of packet switching compared with circuit switching. Hint: think about shared capacity, setup, delay, and bursty traffic.

Answer Key: Six devices require 15 full-mesh links. The first frame uses the default gateway’s MAC address, while the packet retains the remote destination’s IP address. A hub repeats transmissions to all ports and shares collisions; a switch forwards learned frames selectively and provides separate collision domains per port. Fiber suits the long, interference-resistant building link, while copper is economical for short desk connections. Packet switching shares capacity efficiently but can introduce variable queueing delay; circuit switching offers reserved capacity and predictable service after setup but may waste bandwidth.

A network becomes easier to design and secure when every forwarding decision is tied to the correct layer, medium, and device. The central lesson is not merely that networks contain many components, but that each component has a defined responsibility in moving a bit from a source interface to a destination application.

End of chapter one. 4 more chapters in the full book.