The biggest challenge for companies today is that more and more data is coming every day.

Businesses face a difficult challenge: data is growing tremendously and a lot of data is structured information, 80% (texts, videos) and at least 50% is stored outside companies, creating unprotected and security interruptions.

Data lake is a centralized set of data warehouses, raw (unprocessed), both structured (conventional databases) and unstructured, described by metadata. A data lake can easily scale up storage capacity (using platforms such as Hadoop or Apache Spark). Can recognize, analyze and generate data reports collected by a data lake.

Compared to Datawarehouse and Data Marts, a data lake does not describe the data format (no schema available) until it is used.

Features of a date lake:

• Scalable architecture
• Adapting data governance technology
• Have a centralized and indexed cataloging of data and metadata inventory
• Make it as a single repository, easy to use, and easy to share with all applications.
• Supporting lightweight analytical relationships both in the data lake and through multiple analytical approaches and data flows.
• Ensure efficiency, with the ability to delete data and compress and duplicate existing data.
• Accept that the data does not go to the process, but that the process goes to the data.

To create a data lake, consider the following steps:

• Data acquisition
• Data Curation / Grooming data
• Data provision
• Data protection

Source of information: itainnova.es