AQA A Level Computer Science (7517): Fundamentals of communication and networking – Study Guide

This topic explains how data moves between devices, from a single wire inside a computer to the global internet. You need the vocabulary of communication (baud rate, bandwidth, latency), the layered TCP/IP model, how addresses and ports identify a destination, and the main security measures. Calculations are short but they appear regularly, particularly subnet masks and transfer times.

Where this fits in the specification

This is section 4.9 of AQA 7517 and is assessed on Paper 2. It builds on the encryption ideas in Fundamentals of data representation and underpins the later work on client-server databases. Expect definitions, a diagram-labelling or topology question, a short calculation and an extended explanation of how a message crosses the internet.

Communication methods

In serial transmission, bits are sent one after another along a single wire; it avoids the skew between wires that limits fast parallel transmission, where several bits travel at once on separate wires, and it is cheaper over distance. Parallel suits short, fast links within a device.

Synchronous transmission uses a shared clock signal (or clock recovered from the data) so sender and receiver stay in step; it suits large, continuous transfers. Asynchronous transmission has no shared clock: each character is framed by a start bit and a stop bit, so the receiver can recognise where it begins and ends. This is simple, but the extra bits add overhead.

Key measures:

• Baud rate: the number of signal changes per second. Where each change carries more than one bit, the bit rate is higher: bit rate = baud rate x bits per signal. A line at 2,400 baud using 4 bits per signal change has a bit rate of 9,600 bits per second.
• Bandwidth: the range of frequencies a channel carries, which in practice limits the maximum data rate; often quoted loosely as bits per second.
• Latency: the delay between sending and receipt. A link can have high bandwidth and high latency, such as a satellite connection.
• Protocol: an agreed set of rules for communication, including format, speed, error checking and handshaking.

Network topologies and types

In a physical star network every device connects to a central switch. It is easy to extend, a cable fault affects only one device, and traffic is limited to the relevant link, but the switch is a single point of failure and more cable is needed. In a physical bus network devices share one backbone cable, which is cheap and simple, but a break disrupts the whole network, collisions increase with traffic and performance falls as devices are added. A logical bus describes how data is passed between nodes regardless of cabling, as in a bus topology emulated over a star.

Wired networks are generally faster, more stable and more secure; wireless networks give mobility and avoid cabling but share a medium, suffer interference and are easier to intercept.

Client-server networks use a central server to provide services, which gives central control, backup and security, but the server is a point of failure and costs more. Peer-to-peer (P2P) networks have equal devices sharing resources directly, which is cheap and resilient but harder to manage and secure. A thin client relies on a server for processing and storage; a thick client does most processing locally.

Wi-Fi

A wireless network is set up through a wireless access point. Security measures include a changed SSID (network name), which can also be hidden, encryption with WPA2 or newer, and MAC address filtering. Because radio is shared, devices use CSMA/CA (carrier-sense multiple access with collision avoidance): a node listens before transmitting and, if the channel is busy, waits. With RTS/CTS, a device sends a request to send and waits for a clear to send before transmitting, which reduces collisions caused by the hidden node problem, where two devices cannot hear each other but both reach the access point.

The TCP/IP model

Communication is split into four layers, each with its own protocols and a defined interface to its neighbours, so each can be changed independently.

Layer	Role	Example protocols
Application	Provides services to the user’s program	HTTP, HTTPS, FTP, SMTP, POP3, IMAP, SSH
Transport	Breaks data into packets, numbers them, manages delivery between processes	TCP, UDP
Network (Internet)	Addresses and routes packets between networks	IP
Link	Transmits frames across the local physical link, using MAC addresses	Ethernet, Wi-Fi

A port number identifies the application at the destination, and an IP address plus a port form a socket. Well-known ports include 80 (HTTP), 443 (HTTPS), 21 (FTP control), 22 (SSH), 25 (SMTP), 110 (POP3) and 143 (IMAP).

Packet switching and routing

In packet switching, a message is split into packets, each containing a header with the source and destination addresses and a sequence number, plus the data and error-check information. Each packet may take a different route; routers read the destination and forward the packet using their routing tables, and the packets are reassembled in order at the destination. A packet also has a hop count (time to live) so that packets stuck in a loop are discarded. This makes efficient use of the network and allows rerouting around failure, but packets can arrive out of order or late. A gateway connects networks that use different protocols, and a router connects different networks together.

IP addressing

An IPv4 address has 32 bits written as four decimal numbers, and IPv6 uses 128 bits to provide far more addresses. An address has a network identifier and a host identifier; the subnet mask shows the split, with 1s over the network part. CIDR notation such as /26 gives the number of network bits. Private ranges (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16) are not routed on the internet; NAT translates between them and a public address. DHCP assigns addresses automatically. DNS translates domain names to IP addresses.

Worked example: subnets and transfer time

(a) A host has the address 172.16.37.77 with mask /26. Find the network address, broadcast address and number of usable hosts.

A /26 mask has 26 ones, so the mask is 255.255.255.192 (the last octet is 11000000). The host part is 6 bits. The last octet of the address is 77 = 01001101. Keep the top two bits: 01000000 = 64. So the network address is 172.16.37.64. Setting all six host bits to 1 gives 01111111 = 127, so the broadcast address is 172.16.37.127. There are 2^6 = 64 addresses, minus the network and broadcast addresses, so 62 usable hosts (172.16.37.65 to 172.16.37.126). The address 172.16.37.130 is not in the subnet because its last octet is above 127.

(b) How long does a 450 MB file take to send at 60 Mbit/s?

450 MB = 450 x 8 = 3,600 Mbit. Time = 3,600 / 60 = 60 seconds. If only 70 per cent of the rate is achieved in practice, the time is 3,600 / (60 x 0.7) = 85.7 seconds, so 86 seconds to the nearest second.

Network security

Threats include malware, social engineering (phishing), weak passwords, misconfigured access rights, unpatched software and removable media. Measures:

• A firewall examines packets and blocks those that break rules (packet filtering), and may keep track of connections (stateful inspection).
• A proxy server acts for clients, can cache pages and filter content, and hides internal addresses.
• Encryption with public and private keys: a message encrypted with the receiver’s public key can only be decrypted with their private key. A digital signature is made with the sender’s private key and checked with their public key, proving who sent it and that it was unchanged. A digital certificate, issued by a trusted authority, links a public key to an identity.
• Penetration testing by authorised staff finds weaknesses before attackers do.

Common mistakes

• Equating baud rate with bit rate when more than one bit is sent per signal change.
• Forgetting to subtract two addresses (network and broadcast) from the host count.
• Mixing bits and bytes in transfer-time calculations.
• Describing the IP address as identifying the application. The port does.
• Saying a firewall removes viruses. It filters traffic.

Exam technique

For layers, state the layer’s job, not just the protocol name. For a calculation, show the conversion and the units. For “explain how a packet reaches its destination”, cover splitting, headers, routing at each router, and reassembly, in order.

Check your understanding

1. Why is asynchronous transmission given start and stop bits? Answer: so the receiver can tell where each character begins and ends without a shared clock.
2. A line runs at 1,200 baud with 2 bits per signal. State the bit rate. Answer: 2,400 bits per second.
3. Give one advantage and one disadvantage of a physical star topology. Answer: a fault on one cable affects only one device; the central switch is a single point of failure.
4. What does CSMA/CA do? Answer: it makes a device listen and wait for a clear channel before transmitting, to avoid collisions.
5. What is a socket? Answer: an IP address combined with a port number.
6. How many usable hosts are in a /28 subnet? Answer: 2^4 - 2 = 14.
7. What does a digital signature prove? Answer: who sent the message and that it has not been altered.
8. How long to send 90 MB at 24 Mbit/s? Answer: 90 x 8 / 24 = 30 seconds.

Frequently asked questions

What is the difference between baud rate and bit rate? Baud rate is the number of signal changes per second, while bit rate is baud rate multiplied by the number of bits per signal change. A 2,400 baud line with 4 bits per signal change has a bit rate of 9,600 bits per second.

How many usable hosts are in a subnet? Raise 2 to the number of host bits and subtract 2 for the network and broadcast addresses. A /26 subnet has 6 host bits, so it has 64 − 2 = 62 usable hosts.

What is the difference between an IP address and a port number? The IP address identifies the destination device, while the port number identifies the application on that device. An IP address combined with a port forms a socket.

How does CSMA/CA with RTS/CTS work? A device listens and waits for a clear channel before transmitting. With RTS/CTS it sends a request to send and waits for a clear to send, which reduces collisions caused by the hidden node problem.