Circuit switching, packet switching

When two devices communicate across a network, the data does not travel in a straight line — it crosses many connected devices and links. Switching is the method the network uses to decide which route the data takes from the sender to the receiver.

There are two methods you must know for 9618:

Method	Core idea
Circuit switching	Reserve one dedicated path for the whole communication.
Packet switching	Split the message into packets that travel independently.

Both get the data there, but they manage the network very differently — and each is better suited to different situations. The rest of these notes builds up each method, then compares them.

In circuit switching, before any data is sent a single dedicated physical path (a circuit) is set up between the sender and the receiver. This circuit is reserved for the entire communication: nobody else can use those links until the communication finishes, when the circuit is released (torn down).

The classic example is a traditional landline telephone call: when you dial, a circuit is established for your whole call.

Benefits

• Guaranteed, constant data rate — because the path is reserved, the full bandwidth is always available, giving low and predictable delay (no congestion from other users).
• Data arrives in order along the single path, so no reassembly is needed.
• Excellent for real-time two-way communication where steady timing matters.

Drawbacks

• The circuit is reserved even when no data is being sent (e.g. silences in a call), so bandwidth is wasted — an inefficient use of the network.
• There is a setup time to build the circuit before any data can flow.
• If a link on the path fails, the whole communication drops — there is no automatic alternative route.

Where it is applicable: situations needing a guaranteed rate and predictable timing for the whole connection — for example traditional phone calls.

In packet switching, the message is divided into small blocks called packets. Each packet is sent independently across the network and may take a different route to the others. No path is reserved — the network links are shared between many communications at once.

Every packet is built from three parts:

Part	What it contains
Header	Control/addressing data: source IP address, destination IP address, packet/sequence number, number of packets.
Payload	The actual chunk of the message (the data) being carried. Also called the body.
Trailer	An error-check (e.g. a checksum) and a marker showing the end of the packet.

Benefits

• Efficient, shared use of the network — packets use spare capacity on the links, so the network is not left idle as it is with a reserved circuit.
• Resilient/robust — if a link or router fails or is congested, packets are simply re-routed along another path.
• No setup time — packets can start being sent immediately.

Drawbacks

• Variable delay (jitter) — packets queue at routers and take different routes, so timing is unpredictable; this can be a problem for some real-time uses unless buffering is added.
• Packets may arrive out of order (or be lost) and must be reassembled at the destination.

Where it is applicable: sending data across large shared networks, especially the internet, and bulk transfers where efficiency and reliability matter more than perfectly steady timing.

A router is the device that makes packet switching work. Its job is to connect networks together and forward packets towards their destination.

When a packet arrives at a router:

1. The router examines the packet's header and reads the destination IP address.
2. It looks this address up in its routing table — a table mapping destinations to the best available next route.
3. It forwards the packet to the next router or network along that route. This single step from one router to the next is called a hop.

A packet usually passes through several hops, router to router, until it reaches the destination network and finally the destination device.

Crucially, the router makes this decision independently for every packet. This means:

• different packets of the same message can be sent along different routes, and
• the router can avoid a congested or failed link by choosing an alternative route — which is exactly what makes packet switching so resilient.

The diagram below shows three packets of one message leaving the sender, taking different routes through the routers, and arriving at the receiver — where they are reassembled.

Because packets are routed independently, they travel different distances and experience different delays. As a result they may arrive out of order, and occasionally a packet may be lost or corrupted.

This is where the sequence number in each header earns its keep:

1. Every packet's header records its sequence (packet) number — its position in the original message.
2. The receiving device buffers (collects) the packets as they arrive.
3. It uses the sequence numbers to put the packets back into the correct order — this is packet reassembly.
4. The trailer's error-check lets the receiver detect a corrupted packet, and a missing or corrupt packet can be requested again.
5. Once all packets are present and correctly ordered, their payloads are joined together to recreate the original message.

So the message is rebuilt reliably even though no two packets had to follow the same path — the property that makes packet switching both efficient and robust.

Put side by side, the two methods make opposite trade-offs:

Aspect	Circuit switching	Packet switching
Path	One dedicated path for the whole call	Independent packets, possibly different routes
Bandwidth	Reserved → wasted when idle	Shared → efficient use
Data rate / delay	Guaranteed, constant, predictable	Variable (queue at routers, different routes)
Resilience	Link fails → whole call drops	Routers re-route around failures
Order of arrival	In order, no reassembly	May arrive out of order → reassembly
Setup	Setup time needed first	No setup needed

Choosing a method (the exam's favourite task): match the method's properties to the scenario's needs.

• A live two-way video conference is real-time and needs steady timing → a circuit-switched style dedicated path with a guaranteed constant rate avoids lag (though in practice the internet uses packet switching with buffering).
• A large overnight backup is not time-critical but needs efficient, reliable bulk transfer → packet switching uses shared capacity and re-routes around failures, and the variable delay does not matter.

Exam tip: in a 'compare' question, write linked statements (reserved vs shared, guaranteed vs variable, drops vs re-routes). Listing facts about each method separately does not count as comparing.