2.1 Networks including the Internet 网络与互联网

Syllabus learning goals 考纲学习目标

完成本章后，学生需要能够：

• 解释把设备连接成 network 的目的和好处。
• 比较 LAN 与 WAN 的 characteristics。
• 解释 client-server、peer-to-peer、thin-client 和 thick-client models，并根据场景选择。
• 理解 bus、star、mesh 和 hybrid topologies，说明 packets 如何传输并为场景选择 topology。
• 解释 cloud computing、public cloud 与 private cloud 的优缺点。
• 比较 wired 与 wireless networks，以及 copper、fibre-optic、radio/WiFi、microwave 和 satellite media。
• 描述 LAN hardware：switch、server、NIC、WNIC、WAP、cables、bridge 和 repeater。
• 描述 router 的 role and function。
• 解释 Ethernet 与 CSMA/CD。
• 比较 real-time 与 on-demand bit streaming，并解释 bit rate、broadband speed 和 buffering。
• 区分 Internet 与 World Wide Web。
• 描述 modem、PSTN、dedicated line 和 cell phone network。
• 解释 IPv4、IPv6、subnetting、public/private IP 和 static/dynamic IP。
• 解释 URL 如何定位 WWW resource，以及 DNS 的作用。

Chapter 2 的 Explain 题通常每一分对应一个动作或一条因果关系。先识别题目中的设备和场景，再用独立 bullet points 回答。

1 Purpose and benefits of networking 网络的目的

Network 允许两台或更多设备交换数据并共享服务。常见好处包括：

• Sharing files/data between connected devices。
• Communication between users or devices。
• Sharing resources such as printers, storage and software applications。
• Central management of users, security, software and access rights。
• Central backup，使数据保护更一致。
• 使用一个 shared internet connection。

回答“benefits to this organisation”时，要把通用好处放回情境。例如不要只写 “share resources”，而应写 “employees can share the network printer, so a separate printer is not required for every computer”。

2 LAN and WAN

Two-mark comparison pattern

• A LAN covers a small geographical area, whereas a WAN covers a large geographical area.
• A LAN normally uses privately owned/dedicated infrastructure, whereas a WAN often uses external or public infrastructure.

不要把 “LAN is faster” 当作必然定义；速度取决于实际硬件和连接。

3 Client-server and peer-to-peer models

Client-server

Server 集中保存或提供 data、resources 和 services。Client 发出 request，server 接收并处理 request，再返回 response。

典型四分流程：

• The client sends a request to the server.
• The server receives and processes the request.
• The server accesses or updates the required data。
• The server sends the result/response back to the client.

优点：

• Central management of accounts, files, security and backups。
• Data and software can be kept consistent。
• Server hardware can be designed for high availability and many users。

缺点：

• Server and administration increase cost。
• Server failure may make centrally stored services unavailable。
• Heavy demand can create a bottleneck。

Peer-to-peer (P2P)

Computers have equal status and can act as both client and server, sharing their own files or resources directly。

优点：cheap and simple for a small network；no dedicated server is required；peers share resources directly。

常见缺点与解释：

• No central management of security，so the network may be only as secure as its weakest computer。
• No central management of backup，so data may be lost if each user does not back it up。
• No central file/software control, so versions may become inconsistent。
• A peer may respond more slowly while other computers access its resources。
• The computer containing a shared file must be switched on for that file to remain available。

四分题写两个 drawback，每个都要有 consequence，而不是列四个同义短语。

4 Thin-client and thick-client

Thin-client 优点：client hardware can be cheaper；central software updates and data management；less local storage。缺点：requires reliable connection；server failure or overload affects many clients。

Thick-client 优点：can continue more independently；uses local processing；may work offline。缺点：higher client specification；software and security may need maintaining on each device。

5 Network topologies

Bus

All devices share one central/backbone cable。Packets travel along the shared cable and may be seen by all nodes；the addressed device accepts them。

优点：less cable；cheap for a small simple network。缺点：backbone failure affects the whole network；shared medium causes more collisions；performance falls as traffic increases。

Star

Every node has its own connection to a central switch/device。

Packet path [3]：

• The sending device sends the packet to the switch.
• The switch checks the destination address using its stored address information。
• The switch forwards the packet only to the intended receiver.

优点：one cable/node failure usually affects only that node；easy to add devices；fewer collisions with a switch；faults are easier to isolate。缺点：central device failure affects the network；more cable and central hardware are required。

Mesh

Nodes have multiple paths between them。Full mesh connects every node directly to every other node；partial mesh provides more than one route for selected nodes。

优点：multiple routes；high fault tolerance；traffic can use an alternative path。缺点：many links make it expensive and complex to install/manage。

Hybrid

Combines two or more topology types, allowing different parts of a network to use the most appropriate structure；however design and troubleshooting are more complex。

6 Cloud computing

Cloud computing provides storage, applications or processing on remote servers accessed over a network, usually the Internet。

Public and private cloud

• Public cloud infrastructure is operated by a third-party provider and shared logically between customers。
• Private cloud provides dedicated/bespoke services or storage only available to one organisation or selected users；it may be hosted internally or externally。

Benefits of private cloud instead of public cloud

这类 compare 题要求 private 相对 public 的具体优势，不能只写两者共有的 “access files anywhere” 或 “less local storage”。Mark scheme 的独立得分点包括：

• Not reliant on a third party，especially when the organisation hosts/manages the private cloud itself。
• Greater control over security and privacy，because the organisation controls users、access rules and protection of its dedicated system。
• Greater control over backup，including when、where and how company data is backed up and restored。
• Storage can be tailored/scaled to company requirements，for example the amount available to employees or the way files are shared。

Three-mark answer pattern

The company is not reliant on a third-party public-cloud provider. It has greater control over security and privacy, because access to the dedicated system can be managed for its own employees. It also has greater control over backups, including when and where its data is backed up. The storage can be tailored or scaled to the company's requirements.

三分题从上面选择三个独立 points 即可。不要把 security 和 privacy 拆成两个近义分点；mark scheme 把它们视为同一项 control benefit。

Benefits

• Access data/applications from different locations with an Internet connection。
• Provider manages hardware, updates, backup or security services。
• Capacity can be scalable and adjusted to demand。
• Reduced need to purchase and maintain local server hardware。
• Easier collaboration and file sharing。

Drawbacks

• Requires a reliable Internet connection。
• Reliance on a third party for availability, backup, privacy and security。
• Less direct control over remote infrastructure/data location。
• Ongoing subscription or usage costs。
• Provider outage may make services unavailable。

Public cloud often offers lower setup cost and provider-managed maintenance；但题目若明确问 private instead of public，应使用上面的 control、third-party reliance、backup 和 tailored requirements 得分点。

7 Wired, wireless and transmission media

Wired vs wireless

Wired connections usually provide higher and more consistent performance, less interference and stronger physical control over access；but they require cabling and restrict mobility。

Wireless advantages：

• Devices can be mobile because no physical connection is required.
• Easier installation because less cabling is needed。
• Easier to add devices and support different portable device types。

Wireless drawbacks：interference and obstacles can reduce signal；shared radio bandwidth；shorter range；signals may be intercepted outside the building, so encryption/authentication are important。

Transmission media

比较题每一点要形成同一维度的对比，例如 “fibre has less attenuation over long distances, whereas copper needs repeaters sooner”。

8 Hardware used to support a LAN

注意 switch 使用 MAC address 在 LAN 内转发；router 主要根据 IP address 在 networks 之间转发。不要互换。

9 Router

Router connects different networks and forwards packets between them。

Core packet-forwarding functions

• Receives packets from a device/network。
• Examines the destination IP address.
• Uses/updates a routing table to choose a route, possibly the most efficient available route。
• Forwards the packet towards its destination.

其他可能功能：acts as a gateway；assigns private IP addresses using DHCP；presents a public IP address to the Internet；performs NAT、protocol conversion or firewall filtering。

三分题最稳妥写 receiving、checking destination IP、routing-table forwarding 三个独立动作。

10 Ethernet and CSMA/CD

Ethernet is a protocol/suite for data transmission over standard wired network connections。Data is carried in frames containing source/destination addresses and error-checking data。

CSMA/CD 用于 shared Ethernet medium 检测和处理 collisions：

• A node listens to/senses the communication channel before transmitting.
• If the channel is busy, it waits；data is sent only when the channel is idle。
• Two nodes may still begin at nearly the same time, causing a collision。
• If a collision is detected, transmission is aborted and/or a jamming signal is sent。
• Each node waits a different random back-off time。
• It then attempts to retransmit；after repeated collisions, the waiting range may increase。

Four-mark process answer

Each device listens to the channel before transmitting. It sends only when the channel is idle. If a collision is detected, it aborts transmission and sends a jamming signal. Each device then waits a random time before retransmitting.

Drawbacks of CSMA/CD

CSMA/CD 在 traffic 或 node 数量增加时会出现以下限制：

• After repeated collisions, the random back-off range/time increases，so a node may face a very long or potentially indefinite/infinite waiting time。
• Repeated collisions can cause constant jamming signals，so no node successfully completes a transmission。
• Certain nodes cannot be prioritised；urgent traffic may have to compete in the same way as other traffic。
• It can have high power consumption because devices repeatedly listen、collide、wait and retransmit。
• It is only suitable for a short-distance / limited-distance network。
• It is not scalable：adding more nodes creates more contention/collisions and can produce exponentially longer waiting times。

Two-mark drawbacks answer

Repeated collisions make the random waiting time increase, so a node may wait indefinitely before transmitting. CSMA/CD is not scalable, because adding more nodes causes more collisions and much longer waiting times.

Why vague answers lose marks

• Only works on wired internet is not the mark-scheme point。Ethernet is a wired-network context, but the accepted drawback is that CSMA/CD is only suitable for short-distance/limited-distance networks。
• Performance may be decreased does not state a technical cause。Write more collisions → repeated random back-off/retransmission → longer waiting time。

11 Bit streaming

Bit streaming sends media continuously as a series of bits, allowing playback before the entire file has arrived。

Real-time streaming

Media is captured and transmitted from the source as it happens。It normally cannot be paused and replayed in the same way as stored content, and delay must be kept low for calls/conferences。

On-demand streaming

Media is pre-recorded and stored on a server。The user chooses when to play it and can usually pause, rewind or replay；sections may be downloaded ahead of playback。

Streaming process

• Media is often compressed before transmission。
• It is uploaded/sent to a media server and transmitted continuously as bits。
• The receiving device stores incoming data in a buffer。
• The buffer holds data until it can be processed/played, absorbing short differences between download and playback rates。

Real-time video conference: complete process

For a live video conference, the mark-scheme sequence is：

• Video and audio data are compressed before transmission，reducing the amount of data that must be sent。
• The video and audio are transmitted continuously as a series of bits。
• The captured stream is uploaded to a media server。
• For receiving participants, the media server sends/downloads the data to a buffer on the receiving device。
• The buffer temporarily stores data from the server until the receiving device can process it。
• The receiving device obtains/processes the bit stream from the buffer and presents the video/audio。

Four-mark video-conference answer

The video and audio data are compressed before transmission and sent continuously as a series of bits. The stream is uploaded to a media server. The media server sends the data to a buffer on the receiving device. The buffer stores the data until the device can process it, and the device receives the bit stream from the buffer for playback.

Why a short definition is not enough

只写 “continuous flow of bits, so the whole file does not need to be downloaded” 只说明了 bit streaming 的一般定义。题目问 how it is used，还需要具体写出：

• compression before transmission；
• upload to a media server；
• server sends data to the receiving buffer；
• buffer stores data until processing；
• receiving device reads/processes the stream from the buffer。

Bit rate, broadband speed and buffering

For smooth playback, available download speed should remain at least as high as the media bit rate。If data arrives more slowly than it is consumed, the buffer empties and playback pauses/buffers。Higher bit rate can improve quality but requires more bandwidth/data and may increase buffering on slower connections。

12 Internet and World Wide Web

• The Internet is the global network infrastructure/interconnected networks that transmit data using networking protocols。
• The World Wide Web is a service on the Internet: linked web pages/resources accessed using browsers, URLs and HTTP/HTTPS。

Email, file transfer and online games can use the Internet without being the WWW。Webmail uses both：

• Internet because data is sent over the network infrastructure。
• WWW because the user accesses the email service through a website/browser。

13 Hardware and systems supporting the Internet

Modem

Modulates/converts digital data into a form/analogue signal suitable for a communication line and demodulates the received signal back into digital data。

PSTN

The Public Switched Telephone Network contains many interconnected communication lines and switching centres：

• It uses different types of communication lines, so digital data may need to be converted into a different form/analogue signal。
• It supports duplex/full-duplex data transmission：data can be transmitted in both directions at the same time。
• Communication passes through different switching centres and/or Internet Service Providers (ISPs)。

PSTN answer pattern

The PSTN contains different types of communication lines, so digital data may need to be converted into another form/an analogue signal. It supports duplex transmission, so data can travel in both directions at the same time. The communication is routed through switching centres/ISPs.

只写 “data travels in both directions” 不够精确，因为 half-duplex 也能在两个方向传输，但不能同时传输。得分词应写 both directions at the same time 或 full-duplex/duplex transmission。

Dedicated line

Provides a direct/private, always-available connection between locations；typically gives more predictable/faster transmission but costs more than a shared/public connection。

Cell phone network

• Land is divided into cells。
• Each cell has a base station/tower and antenna。
• Phone and tower exchange data wirelessly using low-power radio signals/frequencies。
• Network routes the data onward；multiple devices can use a tower through channel/frequency/time allocation。
• As a device moves, its connection can be handed over to another cell。

14 IP addressing and subnetting

IP addresses identify network interfaces so packets can be addressed and routed to their destination。

IPv4 and IPv6 format

Public and private IP addresses

• Public IP：globally routable/used for direct Internet communication；exposing a service publicly increases the need for firewall and security controls。
• Private IP：used inside a LAN and not directly routed on the public Internet；normally assigned by the router/DHCP and translated through a public address using NAT。

Static and dynamic IP addresses

• Static IP remains fixed；useful for servers/resources that must be found at a consistent address, but requires management and may make a target easier to locate repeatedly。
• Dynamic IP may change when a device reconnects or a lease changes；efficient automatic allocation and easier administration, but unsuitable when a permanently known address is required。

Subnetting

Subnetting divides a network into smaller logical networks/subnetworks。Devices in one subnet share a network identifier；each device has its own host identifier。

Network ID and host ID

An IP address in a subnet contains two logical parts：

Three-mark definition pattern

An IP address is divided into a network ID and a host ID. The network ID identifies the subnetwork and is the same for devices in that subnetwork. The host ID uniquely identifies each device within the subnetwork.

不要写 “network ID identifies the computer”；identifying the individual device is the host ID 的作用。

What a subnet mask does

Subnet mask 告诉设备 IPv4 address 中哪些 bits 属于 network ID，哪些属于 host ID：

• Mask bit 1 selects a network bit。
• Mask bit 0 selects a host bit。
• 255 in one mask group is binary 11111111，so all eight corresponding IP bits belong to the network part。
• 0 in one mask group is binary 00000000，so all eight corresponding IP bits belong to the host part。

Common masks：

CIDR shorthand helps show the number of network bits, but Cambridge questions may provide only the dotted-denary subnet mask。

Past-paper example: identify the two parts

The 2022 Winter Paper 13 question used：

IP address:  10.10.12.1
Subnet mask: 255.0.0.0

255 marks the first group as network bits；the three 0 groups mark host bits：

Network ID: 10
Host ID:    10.12.1

The mark scheme asked only for the network ID and awarded 10。

The same question also used：

IP address:  192.168.12.4
Subnet mask: 255.255.255.0

Network part: 192.168.12
Host ID:      4

The mark scheme answer for the requested host ID was 4。The complete network address is commonly written 192.168.12.0，because all host bits are replaced by 0。

Calculating the network address with AND

To calculate a network address, perform a bitwise AND between each IP bit and the corresponding subnet-mask bit：

IP address AND subnet mask = network address

For a /24 example：

192.168.12.4
AND 255.255.255.0
=   192.168.12.0

The first 24 mask bits are 1，so the first three IP groups remain；the final eight mask bits are 0，so the final group becomes 0。

If the mask splits one 8-bit group, convert that group to binary。Example：

IP last group:    70  = 01000110
Mask last group:  192 = 11000000
                        --------
Bitwise AND result:     01000000 = 64

Therefore：

IP address:      192.168.1.70
Subnet mask:     255.255.255.192
Network address: 192.168.1.64

The host identifier is formed from the remaining six host bits：000110，which is denary 6。

Deciding whether devices are in the same subnet

• Apply the subnet mask to each IP address using bitwise AND。
• Compare the resulting network addresses。
• Same network address → devices are in the same subnet。
• Different network addresses → traffic must be forwarded between subnets, normally by a router/layer-3 device。

Example with mask 255.255.255.0：

Why subnetting is used

Benefits must include both point and consequence：

• Reduces broadcast traffic within each subnet, improving performance。
• Isolates departments/groups, improving security and access control。
• Limits the area affected by faults or attacks。
• Organises address allocation and makes network management/troubleshooting easier。

Four-mark benefits pattern

Subnetting reduces network traffic/congestion because local data only travels through the necessary subnet rather than the whole network. It can improve security because devices and traffic in one subnet can be isolated from other parts of the network.

Other accepted reasoning includes easier expansion, better control of network growth and easier maintenance because one subnetwork can be changed or taken down without stopping the entire network。

15 URL and DNS

URL identifies the location of a resource on the WWW。A URL can contain protocol/scheme, domain name, optional port, path, filename, query and fragment。

Example：

https://www.example.com/course/page.html
protocol   domain name       path/resource

Browser access process

• Browser parses/splits the URL into its components and may check its cache。
• It needs the IP address associated with the domain name。
• The browser queries a DNS server.
• DNS searches its records/hierarchy and returns the matching IP address。
• Browser establishes a connection with the web server at that IP address。
• It sends a request for the resource identified by the path。
• Server returns the resource；browser renders/displays it and may cache relevant data。

Four-mark answer pattern

The browser extracts the domain name from the URL and sends it to a DNS server. DNS returns the matching IP address. The browser connects to the web server at that IP address and requests the resource identified by the URL path.

DNS does not send the web page and a router does not normally translate a domain name；DNS resolves names, router forwards packets。

16 Exam-answer method 答题方法

Device-function questions

使用“input/action/output”组织：device receives 什么、checks/uses 什么、forwards/returns 什么。

Advantages and drawbacks

不要列同义词。每一点写成：

feature/change -> technical consequence -> benefit/drawback in the scenario

例：Wireless requires no physical cable，so students can move between rooms while staying connected。

Compare questions

使用同一维度成对比较：coverage 对 coverage、ownership 对 ownership、speed 对 speed。避免一边写 cost，另一边写 security。

Process questions

按时间顺序写独立动作，例如 CSMA/CD、streaming、DNS lookup 和 packet routing。四分题通常至少准备四个可独立评分步骤。

17 Final checklist 真题检查表

• LAN/WAN 是否写到 geographical area 和 infrastructure ownership？
• Client-server 是否出现 request、process、response？
• P2P drawback 是否带具体 consequence？
• Private cloud vs public cloud 是否写了 third-party reliance、security/privacy control、backup control 或 tailored requirements？
• Star packet path 是否经过 switch，并检查 destination address？
• Wired/wireless 和 media comparison 是否在相同维度？
• Switch 是否使用 MAC address，router 是否使用 IP address？
• CSMA/CD process 是否包含 sense、idle、collision、random wait/retransmit？
• CSMA/CD drawback 是否明确写出 repeated collisions/back-off、constant jamming、no priority、high power、limited distance 或 poor scalability？
• Real-time bit streaming 是否写了 compression、continuous bits、media server、buffer 和 receiving device processing？
• Streaming 是否说明 buffer 的作用以及 bit rate 与 broadband speed 的关系？
• Internet 与 WWW 是否写成 infrastructure 与 service/resource system？
• PSTN 是否写了 signal conversion、switching centres，以及 both directions at the same time / duplex transmission？
• IPv4/IPv6 的 bits、groups、separator 和 number base 是否准确？
• Network ID 是否识别 subnetwork，host ID 是否唯一识别该 subnet 内的 device？
• Subnet mask 的 1 是否对应 network bits，0 是否对应 host bits？
• 求 network address 时是否逐位计算 IP address AND subnet mask？
• 判断同一 subnet 时是否比较两个计算后的 network addresses？
• Subnetting 是否写出 benefit 及其解释？
• DNS 是否只负责把 domain name 解析为 IP address？

---

本材料依据 9618 syllabus Chapter 2 的 15 项要求，并综合本页所列 2021-2025 past-paper questions 与 mark schemes。简答模板优先使用评分方案中反复出现的 device action、technical consequence 和 scenario justification。