Learn
A2.1.2 | MODERN DIGITAL INFRASTRUCTURES
01 | THE SUPPORTING SYSTEMS BEHIND A DIGITAL SERVICE
A student opens a learning platform on a phone. A roadside computer estimates traffic from sensor data. A cryptocurrency network shares ledger updates. Each depends on more than the application visible to its user.
Digital infrastructure is the underlying computing, storage, communication and supporting services that make digital activity possible. Physical equipment, networks and managed software services work together.
This lesson covers the internet, cloud computing, distributed systems, edge computing and mobile networks. These are different ideas, not five mutually exclusive alternatives. A mobile app can use a cellular network to reach a cloud service that runs across distributed servers.
For each, describe its purpose, a relevant benefit and a limitation. Then connect these features to an example, rather than assuming that newer or more complex always means better.
02 | THE INTERNET: CONNECT DIFFERENT NETWORKS
The internet is a global interconnection of networks using common internet protocols. Devices can communicate across networks administered by different organisations, with routers forwarding traffic along available paths.
Its purpose is general communication between connected systems. Web access, messaging, online games and many cloud services can use internet connectivity, but the internet does not itself supply every application’s data or features.
A major benefit is broad interoperability and reach: a school can access resources hosted elsewhere without building a private physical connection to every provider. The underlying paths still use physical links and equipment.
Access depends on local connectivity, providers, routing, capacity and destination services. Congestion, outages, filtering or security incidents can interrupt a particular path or service. The fact that the internet has many paths does not guarantee that a user’s single access link can never fail.
03 | THE WORLD WIDE WEB IS A SERVICE, NOT THE WHOLE INTERNET
The World Wide Web is an information system of linked resources and applications commonly accessed using browsers and web protocols over the internet. A browser requests a resource from a server and receives a response it can display or process.
The internet supplies communication infrastructure; the web supplies a particular way of publishing and accessing information. Other internet services need not be web browsing. A school can also provide web-based resources on a private internal network.
The web benefits users by connecting information through links and supporting access across many kinds of devices. Its operation still depends on working networks, servers, application software and permissions.
A webpage failing to load does not prove that the entire internet has failed. The problem might be a local connection, an address lookup, the remote server or the application itself. Keep the service and its supporting infrastructure distinct.
04 | CLOUD COMPUTING: ACCESS MANAGED RESOURCES ON DEMAND
Cloud computing provides network access to configurable computing resources and services, commonly from a shared managed pool that can be provisioned and adjusted as required. These resources run on real hardware in data centres; they are not stored in a non-physical cloud.
An organisation can obtain storage, application services or processing capacity without buying and managing all the supporting equipment itself. A school might use a hosted learning platform; a developer might provision resources for an application.
Benefits can include flexible scaling, remote access and reduced local infrastructure administration. SaaS, PaaS and IaaS describe different service responsibilities, as introduced in A1. Scaling still has limits, configuration requirements and costs.
Limitations include network dependency, provider outages, ongoing charges and concerns about access, data handling and provider dependence. The customer still has responsibilities for appropriate account and data choices. Using cloud storage does not automatically establish a suitable backup or retention policy.
05 | DISTRIBUTED SYSTEMS: SEVERAL COMPUTERS COOPERATE
A distributed system has networked components on different computers cooperating to deliver a service or computation. Work or data is spread across nodes rather than all being handled by one computer.
One design divides a large job between workers. Another replicates data across machines so that more than one copy exists. Some systems combine task distribution and replication. A distributed system can operate in local premises or cloud infrastructure.
Benefits can include increased capacity, geographical reach and resilience to selected failures. Those benefits require suitable design: placing identical servers behind one failed power supply does not remove that shared dependency.
Limitations include communication delay, partial failure and coordination complexity. A node may be working while another is unreachable. Replicated copies may not all receive an update at exactly the same moment, so software needs clear rules about consistency and recovery. “Distributed” is not the same as “automatically decentralised ownership” or “impossible to fail”.
06 | EXAMPLE: A CRYPTOCURRENCY BLOCKCHAIN
A cryptocurrency blockchain is an example of a distributed ledger arrangement. Participating nodes communicate to validate and share ledger information according to the system’s rules. Blocks are linked using cryptographic information, and agreement mechanisms determine which updates are accepted.
The purpose is to maintain a shared transaction history under specified trust assumptions. Distributing relevant ledger information can reduce reliance on one authoritative storage computer, while cryptographic links make unauthorised changes detectable under the design’s assumptions.
Coordination has costs: communication, validation and storage are needed, and accepted updates may take time. Energy use and performance depend on the chosen agreement mechanism; not all blockchains use proof of work.
A blockchain does not prove that every outside-world claim supplied to it is true, nor does it remove software, account-key or governance risks. The example illustrates distributed infrastructure, not a guarantee of investment value or infallibility.
07 | EDGE COMPUTING: PROCESS NEAR THE DATA SOURCE
Edge computing places relevant processing close to where data is produced or used, such as on a device, in a local gateway or at a nearby site. It describes placement relative to the workload, not one fixed type of box.
A roadside processor can analyse local sensor observations and send a compact result to a traffic controller. It can also send summaries elsewhere for longer-term analysis. Not every raw observation must take a round trip to a distant data centre before a local result is available.
Benefits can include reduced communication delay, lower upstream data volume and continued local operation when an external path fails, provided the needed local resources and dependencies remain available. Local processing may also reduce unnecessary transfer of sensitive raw data.
Limitations include constrained local capacity, physical exposure and the difficulty of updating and securing many separate devices. Keeping data nearby does not itself guarantee privacy. A local application that still requires a remote login or model service may remain dependent on external connectivity.
08 | EDGE AND CLOUD CAN WORK TOGETHER
An edge/cloud hybrid gives different jobs to different locations. Local processing can provide a timely observation, while a central service combines summaries across sites or runs more demanding analysis.
For smart traffic infrastructure, local operation can use sensor results under the controller’s permitted rules, while a wider service analyses patterns across junctions. A lost external link need not remove every local function if the system was designed for that situation.
However, global coordination or central updates may become unavailable. Independence must be designed and tested; drawing an edge box does not prove that all local functions can safely continue.
09 | MOBILE NETWORKS: CONNECT WHILE USERS MOVE
A mobile cellular network provides wireless access over areas served by cells. A device communicates by radio with suitable base-station equipment, which connects through supporting transport and core-network services.
As a device moves, a handover can transfer its connection between serving cells or access arrangements. This supports communication beyond the range of one local wireless access point. Exact behaviour depends on the network and device.
Benefits include mobility, wide-area access and connectivity where a fixed local connection is not convenient. A school trip can use cellular access for a hosted service, and roadside equipment can use mobile connectivity where suitable.
Limitations include coverage gaps, indoor signal attenuation, shared radio capacity, congestion, battery demand, cost and provider dependency. A phone showing a signal does not guarantee that a remote application will work. Wi-Fi is a different access technology; a mobile device can use Wi-Fi without using its cellular data path.
10 | EXAMPLE: A SCHOOL USES SEVERAL INFRASTRUCTURES
Inside school, devices use local networks to reach printers, servers or external links. The internet connects the school to services elsewhere. A browser-based learning platform uses web technologies, while its provider may run the service on cloud resources.
The provider can distribute storage or application work across servers. Some local activities, such as analysing sensor readings in a robotics room, can happen near the data at the edge. A teacher away from school can use a mobile network to reach permitted services.
Each layer has a role. Losing the school’s external connection may stop a cloud-only workflow while an independent local printer or sensor activity still works. A cached page may remain visible even though new updates cannot be fetched.
A good description identifies the actual dependencies. It does not assume the word “online” tells you where all data is processed or which functions are available offline.
11 | TEST PLACEMENT, DELAY AND AN EXTERNAL OUTAGE
A toy roadside processing task supplies an occupancy estimate to a local controller. Its illustrative deadline is 50 ms. Compare three plans: cloud processing, edge processing, or local edge processing with a separate cloud-summary upload.
The edge calculation takes 8 ms. Cloud processing takes 4 ms plus the selected round-trip network delay. The chosen delay is already a round trip, so do not double it. Local sensing, interface and actuation delays are omitted in every plan.
The local I/O path remains available. The edge-processor switch changes only local computation, while the external-link switch changes access to the cloud. There is no automatic fallback beyond the selected plan.
These invented constants illustrate dependency and timing, not measured performance or a real traffic-signal deadline. Sending a summary in the hybrid plan is a separate background task and does not delay its local result.
| Plan | Local result available? | Modelled result delay | 50 ms target | Cloud work |
|---|
Try 40 ms, then 120 ms. Disconnect the external link and compare the hybrid local result with its unavailable central summary.
12 | PURPOSES, BENEFITS AND LIMITATIONS SIDE BY SIDE
| Infrastructure | Purpose | Potential benefit | Relevant limitation |
|---|---|---|---|
| Internet | Communication between interconnected networks | Broad reach and common communication rules | Path, provider and destination dependencies |
| Cloud computing | Managed network-accessible resources and services | Flexible provisioning and reduced local administration | Costs, connectivity, provider and data-management concerns |
| Distributed systems | Cooperation across networked components | Capacity and resilience with suitable design | Coordination, consistency and partial-failure complexity |
| Edge computing | Processing near data production or use | Lower upstream traffic or result latency | Limited local resources and dispersed maintenance |
| Mobile networks | Cellular communication while devices change location | Mobility and broad-area access | Coverage, congestion, power and provider limits |
The benefit is conditional. Distribution needs a failure-tolerant design, edge processing needs local capacity, and a cloud service needs appropriate provisioning and access. Use “can” or state the conditions instead of promising universal performance.
13 | RESILIENCE, SECURITY AND DATA RESPONSIBILITY
Resilience is the ability to continue or recover appropriate service when failures occur. Replication, independent links and local fallback can contribute, but only if they address the dependencies that can actually fail.
Security responsibilities also cross the infrastructure. Devices need protection, services need access rules, updates need management and transmitted data may need suitable encryption. Moving a task to the cloud or edge does not remove these responsibilities.
Data placement affects who can access information, how much must be transferred and how it is retained. Sending summaries rather than raw video may reduce transfer and exposure, but the summaries themselves can still be sensitive.
Cost includes more than purchasing equipment: provider charges, network traffic, maintenance, energy and staffing can matter. No single infrastructure choice removes every operational tradeoff.
14 | IDENTIFY THE FEATURE BEING DESCRIBED
Choose the infrastructure concept that best matches the specific feature. Several concepts can coexist in the same real service, so the questions ask what a feature describes rather than forcing an entire application into one category.
15 | WORKED SCENARIOS: WHAT EACH SYSTEM CONTRIBUTES
WWW: the internet carries communication while web servers and browsers provide access to linked resources and applications. Cloud or distributed services may support the server side, but the web and internet remain different concepts.
Cryptocurrency blockchain: networked nodes exchange and validate ledger information under agreement rules. Distribution supports a shared ledger, while communication, storage and consensus impose costs.
Smart traffic infrastructure: nearby computation can supply local results, with wider connections carrying summaries or coordination data. Time, faults and safe operation must be handled by the control design.
School: local networking, internet links, hosted platforms, distributed provider services and mobile access can support different parts of learning. Describe the arrangement and failure dependencies rather than only naming technologies.
16 | CHECK THE COMMON MISCONCEPTIONS
| Claim | More accurate description |
|---|---|
| The internet and web are the same. | The web is a service/information system commonly using internet communication. |
| Cloud data has no physical location. | Cloud services use real computing and storage infrastructure. |
| Distributed means no component can fail. | Partial failures and shared dependencies need explicit handling. |
| Every distributed system is a blockchain. | Blockchains are one particular ledger arrangement. |
| Edge means no external dependency ever. | Only suitably self-contained local functions can continue independently. |
| A mobile phone always uses cellular data. | It can use Wi-Fi or other connections instead. |
| Replication is automatically a recoverable backup. | Recovery needs protection against deletion, corruption and other relevant failures. |
| A latency result proves a safety deadline. | Real end-to-end bounds require all relevant delays and conditions. |
17 | DESCRIBE THE PURPOSE AND A CONDITIONAL BENEFIT
For describe, give a detailed account of what the infrastructure does and its relevant features. Include an example, explain the benefit it can provide and state a limitation that follows from the arrangement.
For edge computing: “A roadside processor analyses nearby sensor data, so local results do not require a cloud round trip. This can reduce response delay and upstream traffic. However, the local processor has finite capacity and must still be maintained and secured.”
For cloud: “A school accesses a managed platform over a network, reducing local server administration and permitting flexible resource provision. Access still depends on working paths and services, and provider charges and data responsibilities remain.”
The next lesson describes the network devices supporting these arrangements. Here, keep purpose, processing location, communication and application service distinct.
Further reading: The World Wide Web · Computation near network data sources