All subjects · Information Technology · WIT13 2 hour written exam on data, enabling technologies and systems development
Big Data infrastructure and the impact of storing it
Edexcel IAL Information Technology · WIT13 2 hour written exam on data, enabling technologies and systems development
Big Data needs special systems to collect, store and transmit it, and storing it has consequences for access, processing time, transmission time and security (spec 12.3.2 and 12.3.3).
Collection
Data is collected automatically from many sources: point-of-sale terminals and card payments, loyalty cards, website and app activity (clicks, searches, cookies), social media and email, sensors and IoT devices, GPS, CCTV, and public or purchased data sets. Organisations may also buy or license data from other organisations. Collection must follow the laws and regulations on personal data that apply where the data subjects live (you need to know the types of law, such as data protection and privacy laws, rather than their details).
Storage and processing
- Distributed file systems and processing frameworks (Hadoop is a well-known example) split data across many servers and process the parts in parallel.
- NoSQL databases (such as Cassandra and MongoDB) store semi-structured and unstructured data without a fixed schema and scale out across many machines.
- Massively parallel processing (MPP): many processors each work on part of the data at once.
- Cloud computing: storage and processing rented on demand, so capacity scales up and down and there is no large up-front hardware cost.
- Data warehouses (cleaned, structured, integrated historical data for analysis) and data lakes (raw data in its original format, stored until needed).
- Data centres need power, cooling, fire protection and physical security.
Transmission
Moving huge volumes needs high bandwidth: fibre optic backbones, compression, and sometimes alternatives to standard TCP. TCP's acknowledgement and congestion control can underuse very fast long-distance links, so protocols such as UDT and QUIC are used to move data faster. Wireless links suffer variable bandwidth and interference. Very large one-off transfers are sometimes made by physically shipping storage devices.
Impact of storing Big Data
| Impact | Explanation |
|---|---|
| Access | Many users and systems need fast access; data must be indexed, distributed and replicated; access must be controlled |
| Processing time | Analysing petabytes on one machine would take too long, so work is split across clusters; real-time analysis needs in-memory and stream processing |
| Transmission time | Large volumes take a long time to move; bandwidth costs money; processing is moved close to the data (edge or in the same cloud region) |
| Security | A large, valuable store is an attractive target; breaches affect millions; needs encryption, access control, monitoring and legal compliance |
Worked example
A Hadoop-style cluster splits a 12 TB job evenly across 400 servers, each of which processes 1 TB per hour. How long does the job take, compared with one server?
- Data per server = 12 / 400 = 0.03 TB
- Time per server = 0.03 TB / 1 TB per hour = 1.8 minutes
- One server alone: 12 hours
- Speed-up = 12 h / 1.8 min = 400 times (ignoring overheads)
Answer: 1.8 minutes instead of 12 hours
Check yourself
- Why are NoSQL databases often used for Big Data?
- They only accept structured tables
- They store semi-structured and unstructured data without a fixed schema and scale across many servers
- They prevent all security breaches
- They do not need any storage
Show answer
They store semi-structured and unstructured data without a fixed schema and scale across many servers. Flexible schema plus scale-out.
- What is the main difference between a data lake and a data warehouse?
- A lake is always smaller
- A warehouse holds only images
- A lake cannot be stored in the cloud
- A lake holds raw data in its original format; a warehouse holds cleaned, structured data ready for analysis
Show answer
A lake holds raw data in its original format; a warehouse holds cleaned, structured data ready for analysis. Raw versus processed.
- Which is the main advantage of using cloud computing to store Big Data?
- Data never needs to be transmitted
- Capacity can be scaled up or down on demand without buying hardware
- It removes all legal obligations
- It makes data more accurate
Show answer
Capacity can be scaled up or down on demand without buying hardware. Elastic, pay-as-you-go capacity.
- A job takes 30 hours on one server. It is split evenly across 120 servers. How many minutes does it take, ignoring overheads?
Show answer
15 minutes. 1800 minutes / 120 = 15 minutes.
- Why does storing Big Data increase security risk?
- A large store of valuable data is an attractive target and a breach affects many people
- Big Data cannot be encrypted
- Big Data is always public
- It is stored on paper
Show answer
A large store of valuable data is an attractive target and a breach affects many people. Scale raises both the value and the impact.
- What is massively parallel processing?
- One fast processor works through the data in order
- Data is processed only at night
- Many processors each work on part of a task at the same time
- Data is printed and checked by hand
Show answer
Many processors each work on part of a task at the same time. Split and run simultaneously.
- Explain one reason why transmission time is an issue for Big Data.
Show answer
The volume of data is so large that moving it over a network takes a long time and uses a lot of bandwidth, which delays analysis and costs money.. Volume, so time or bandwidth.
Sources
- Pearson Edexcel International Advanced Level in Information Technology specification (first teaching 2018), Issue 2, February 2021
- Pearson Edexcel IAL Information Technology: scheme of work for Unit 3
- Pearson Edexcel IAL Information Technology WIT13/01 examiners' report, June 2025 series
Last updated 8 November 2026. Spotted a mistake? Tell us.