Course 5, lesson 49 of 100, Ages 11+
Big data and the cloud
Storing and crunching enormous datasets
Like I’m 5
Some datasets are so big they wouldn't fit on a million laptops. They live in huge buildings full of computers called data centres, often called the cloud.
The big idea
Big data means datasets too large or too fast-growing for one computer: every card payment in a country, or every video uploaded in a day. To handle it, work is split across thousands of machines.
The cloud lets anyone rent that computing power over the internet instead of buying it. Training large AI models uses enormous clusters of specialised chips, which costs a lot of money and electricity.
Examples
- Online payments: Fraud detection checks millions of payments every minute.
- Video sites: Hundreds of hours of video are uploaded every minute.
- Renting compute: A start-up rents cloud computers to train a model for a week.
How it works
- Store the huge dataset across many machines.
- Split the work into pieces that run at the same time.
- Combine the results into one answer.
Check your understanding
- What is a data centre?
- Options: A building full of computers that store and process data; A shop that sells data; A library of paper books.
Answer: A building full of computers that store and process data. Data centres power the cloud. - How is big data processed quickly?
- Options: By splitting work across many machines at once; By using one very slow computer; By printing it all.
Answer: By splitting work across many machines at once. Parallel processing divides the job so it finishes fast.
Remember
Big data is split across many machines in data centres, which you can rent as the cloud.
Talk about it
Which apps you use might create big data every day?
Go deeper
Distributed systems like MapReduce, Spark and modern data warehouses split processing across clusters. Large AI training uses data and model parallelism across thousands of GPUs or TPUs.