Show understanding of packet switching

14.2 Circuit Switching and Packet Switching

Learning Objective

Demonstrate a clear understanding of packet switching – how it works, its advantages and disadvantages, and how it differs from circuit switching in terms of resource allocation, latency, reliability, security and real‑world use. Relate the concepts to the broader networking topics required by the Cambridge International AS & A Level Computer Science (9618) syllabus.

1. Overview of Switching Techniques

• Circuit Switching – a connection‑oriented method that reserves a dedicated physical path for the whole duration of a communication session.
• Packet Switching – a connection‑less method that breaks data into small packets; each packet is routed independently and shares the same physical links with other traffic (statistical multiplexing).

2. Where the Techniques Are Used (LAN / WAN Context)

• LANs (Local Area Networks) – modern LANs (Ethernet, Wi‑Fi) are packet‑switched. They use switches and routers to forward frames/packets.
• WANs (Wide Area Networks) – early telephone WANs used circuit‑switching (PSTN). Today most WAN services (Internet, MPLS, cellular data) are packet‑switched, although some voice services still use a circuit‑switched core.

3. Network Hardware and Their Role in Each Switching Model

4. The Protocol Stack – Mapping to Switching

The Cambridge syllabus expects students to know the four‑layer TCP/IP model (Link, Internet, Transport, Application). The table below shows where the key protocols mentioned in the notes belong.

5. IP Addressing and DNS – The “Names” of Packet‑Switched Networks

• IP addressing gives each host a unique logical identifier (e.g., 192.168.1.10 or 2001:0db8::1). Subnet masks define the size of the local network.
• DNS (Domain Name System) translates human‑friendly domain names (e.g., www.example.com) into IP addresses, allowing users to locate resources without memorising numbers.
• Contrast with circuit‑switched telephony, where a telephone number is a fixed identifier that routes a call through a pre‑reserved circuit.

6. CSMA/CD and Ethernet Collisions

In a traditional Ethernet LAN (a packet‑switched environment) all stations share the same physical medium. Carrier Sense Multiple Access with Collision Detection (CSMA/CD) works as follows:

1. Each station listens (“carrier sense”) before transmitting.
2. If the medium is idle, it begins to send its frame.
3. If two stations transmit simultaneously, a collision occurs; both detect it, stop, wait a random back‑off time, then retry.

Because a circuit‑switched link is dedicated to a single call, collisions never happen – the path is already reserved.

7. Wireless vs. Wired – How Switching Enables Broadband

• Wired broadband (e.g., fibre‑to‑the‑home, cable) uses packet‑switching to allocate bandwidth dynamically to many users.
• Wireless broadband (Wi‑Fi, LTE, 5G) also relies on packet‑switching; the radio spectrum is shared, and protocols such as CSMA/CA (Wi‑Fi) or scheduling (LTE/5G) manage access.
• Both technologies support “on‑demand” streaming (video, music) because packets can be sent whenever bandwidth is available, unlike a fixed‑bandwidth circuit.

8. How Circuit Switching Works

1. Signalling phase – a request is sent to the network to set up a dedicated circuit.
2. Path reservation – a continuous series of physical links is reserved; a fixed portion of the link bandwidth is allocated for the whole session.
3. Data transfer – once the circuit is established, data flow with predictable, low latency (no queuing).
4. Teardown – when the session ends, the resources are released.

9. How Packet Switching Works

1. Data are divided into packets (typically 500–1500 bytes).
2. Each packet receives a header containing:

   • Source and destination IP addresses
   • Sequence number (for re‑assembly – used by TCP)
   • Checksum/CRC for error detection
   • Optional QoS / priority field (e.g., DSCP)

3. Packets are routed independently; they may travel different paths and experience different queuing delays.
4. At the destination, packets are reordered using the sequence numbers; missing packets are requested again (ARQ – Automatic Repeat reQuest, e.g., TCP retransmission).

10. Latency, Throughput and Efficiency

• Circuit‑switching latency = signalling time + propagation delay. After setup, latency is essentially constant.
• Packet‑switching latency = propagation delay + queueing delay + processing delay at each router. It varies with network load.
• Effective throughput (packet‑switched link):

  \$T_{\text{eff}} = R \times (1 - O)\$

  where R = raw link rate (bits s⁻¹) and O = proportion of bandwidth used for header overhead.

• Because the circuit is reserved even when idle, its utilisation can be low; packet‑switching fills idle periods with other users’ traffic, giving higher overall efficiency.

11. Reliability, Error Handling and QoS

• Packet loss may occur due to congestion or link failure. TCP recovers by retransmitting lost packets (ARQ) and by using acknowledgements.
• Resilience – if a link fails, routers can find an alternate path; the rest of the traffic continues.
• Quality of Service (QoS) – fields in the IP header (e.g., DSCP) allow routers to give priority to latency‑sensitive traffic such as VoIP or video streaming.
• Reliability in circuit‑switched networks is high while the circuit exists, but a single link failure breaks the whole call.

12. Security Considerations in Packet‑Switched Networks

• Because packets travel through many intermediate devices, they can be intercepted or altered. Security mechanisms include:

  • Encryption – TLS/SSL, IPsec encrypt payloads to provide confidentiality.
  • Authentication & Integrity – digital signatures, HMACs ensure packets are genuine and unmodified.

• Circuit‑switched voice calls are less exposed to interception (the path is dedicated), but modern telephony still uses encryption (e.g., SRTP for VoLTE).

13. Comparison of Circuit Switching and Packet Switching

14. Advantages and Disadvantages of Packet Switching

Advantages

• Efficient use of bandwidth – idle capacity is utilised by other users.
• Robust to failures – alternate routes can be found automatically.
• Supports bursty, continuous and interactive traffic.
• Scales easily as the number of users grows.
• QoS mechanisms allow prioritisation of latency‑sensitive traffic.
• Enables broadband services (Wi‑Fi, LTE, fibre) and on‑demand streaming.

Disadvantages

• Variable latency can affect real‑time applications.
• Header overhead reduces effective data throughput.
• Packet loss may occur; retransmission mechanisms (ARQ) are required.
• Routing, congestion control and QoS management are more complex than in circuit‑switched networks.
• Security concerns – packets travel through many devices and must be protected.

15. Real‑World Examples

• Circuit‑switched: Public Switched Telephone Network (PSTN), legacy ISDN lines, dedicated leased‑line links for corporate sites.
• Packet‑switched: The Internet (TCP/IP), Voice over IP (VoIP) services, mobile data networks (3G/4G/5G), MPLS backbones used by ISPs.
• Hybrid: Modern cellular networks – voice may still traverse a circuit‑switched core (e.g., 2G), while data uses packet‑switched LTE/5G; MPLS adds label‑switched paths on top of a packet‑switched infrastructure.

16. Key Points to Remember

1. Circuit switching reserves a dedicated path (resource reservation) before any data are sent.
2. Packet switching shares the same links among many users (statistical multiplexing) and does not require prior set‑up.
3. Each packet carries a header with addressing, sequencing, error‑checking and optional QoS information.
4. Packets may travel different routes, arrive out of order, and be re‑assembled at the destination.
5. Header overhead reduces net throughput but provides flexibility, resilience and the basis for routing.
6. Latency in packet‑switched networks includes queueing delay; ARQ and QoS are used to mitigate loss and delay for real‑time traffic.
7. Security measures (encryption, authentication) are essential in packet‑switched environments.