Data Lake
A data lake is a massive pool of object storage that holds structured, semi-structured and unstructured data side by side — built for the next generation of analytics and machine learning.
One repository, any type of data
A data lake uses object storage to hold all data types. Each piece of data carries a metadata tag and a unique identifier, so applications can find every object containing a given term and retrieve it in a viewable format, or import it into a database. By leveraging inexpensive object storage and open formats, a lake lets applications discover information that would otherwise sit lost inside petabytes of storage — breaking down the data silos that build up across an organisation, without the historical restrictions of data transformation.
- No schema required — data is stored “as is,” from raw to fully refined
- Foundation for advanced analytics, machine learning and data-driven decisions
- Data scientists and analysts can run their own tools directly against it
- Deployable on-premise or in the cloud
A natural fit for IoT
A data repository for IoT can use edge computing devices to process and analyse data locally before sending it on to the data lake.
See Edge Computing →The case for a cloud-based data lake
Lower TCO
Cloud-based data lakes reduce total cost of ownership compared with early on-premise deployments.
Faster time to value
Cloud infrastructure accelerates how quickly a data lake starts paying off for the business.
Business agility
A single, central repository improves how quickly the business can act on its own data.