Harshit Ahluwalia — January 7, 2022
Beginner Data Engineering Data Warehouse

Introduction

All data mining repositories have a similar purpose: to onboard data for reporting intents, analysis purposes, and delivering insights. By their definition, the types of data it stores and how it can be accessible to users differ. This article will discuss some of the features and applications of data warehouses, data marts, and data lakes.

 

What is a Datawarehouse?

A Datawarehouse works as a central repository that merges information from different sources and consolidates it through the extract, transform and load process, also known as ETL process, into a comprehensive database used for analytical and business techniques. At a very high level, the ETL process extracts the data from multiple sources, transforms it into a cleaned format to be used for business processes, and finally loads that data into the data repository. A data warehouse stores current data and the historical information that has been cleansed, conform categorized. When data gets loaded into the data warehouse, it is modelled and structured, ready for a specific purpose. Moreover, a data warehouse was traditionally used for storing data from transactional databases such as CRM, ERP, HR and Finance applications. But with the advancement in technology like NoSQL technologies and new data sources, non-relational databases are also used for data warehousing. Typically a data warehouse has a 3-tier architecture.

Datawarehouse

The bottom tier of the architecture includes the database servers, which could be relational or non-relational or maybe both, that extract data from multiple sources and consolidate it into one.

The second tier of this architecture includes OLAP Server, a software category that allows users to proceanalyzeanalyze data/information coming from multiple databases server. The topmost level of this architecture is the Client Front end layer. This layer includes all the applications and tools used for reporting, queryianalyzingalyzing data. In response to the rapid growth and today’s sophisticated analytical tools, data warehouses that once resided in on-premise centres are moving towards the cloud. Compared to an on-premise data warehouse, a cloud warehouse provides many benefits that include lower cost, limitless storage and computer capabilities, scale on a pay as you go basis, faster disaster recovery. Organization station, you should opt for a cloud-based data warehouse as it has many benefits compared to the on-premises data warehouse. Popularly used data warehouses include Teradata, Oracle Exadata, IBM DB2 Warehouse on the cloud, Amazon Redshift, Big Query by Google Cloudera’s Enterprise Datahub and Snowflake Cloud Data warehouse.

Companies using Data warehouse

Read more about Data warehouse on our blog.

What are Data Marts?

The data mart is a subsection of the data warehouse built for a particular operational task or a business function, purpose, or community of users. There are specific reasons for creating data marts; the first is that we can easily access frequent data. Secondly, end-user response time is improved. Thirdly, easy creation of data mart, because for creating the data warehouse a lot of resources and work is to be done while the creation of data mart is far easy compared to the data warehouse. There is a minimal cost that is related to data mart.

There are three types of data marts: dependent, independent, and hybrid. Dependent data marts are built by drawing data from the existing central warehouse. In contrast, independent data marts are created by drawing from operational or external data sources or both. Dependent data mart offers analytical capabilities for a restricted data warehouse area. Moreover, it also provides isolated security and solo performance. A hybrid data mart combines input from a data warehouse, operational systems and external systems. The difference lies in how the data is extracted from the source, how data has been transformed that needs to be applied, and how the data has been transported into the mart. Dependent Data Mart pulls data from the enterprise data warehouse, which has already been cleaned and transformed. Independent data mart has to carry out the cleansing and transformation of data as the data has been coming from the operational systems and external sources. Whatever the type, the main aim of the data mart is to provide data to the end-users that are most relevant to them when they need it. These data marts accelerate the business process and offer a cost and time-efficient way to take the data-driven decision.

What is a Data Lake?

A Data Lake is a repository of data that stores all types of data, whether structured, unstructured, or semi-structured. It holds a large amount of data in its native format.

Data Lake

Source: Image

While a data warehouse stores data that has been adequately cleaned, transformed and ready for business operations or analytical tasks. While dumping the data into a data lake, data can be loaded without defining the structure and schema of the data. In simpler terms, we can say that a data lake is a repository that contains raw data in its native form, straight from the source. It does not mean that a data lake is a place to dump your data without governance. While the data is appropriately classified in the data lake, it is protected and governed. Data lakes can be deployed using cloud object storage such as Amazon S3 or large scale distributed systems such as Apache Hadoop to process big data. They can also be deployed on a relational database management system or NoSQL data repositories. Data lakes offer many benefits, such as they can store all types of data such as structural, unstructured and semi-structured data. The second benefit is saving time in defining the data’s structure, schemas, and transformations as it is imported into the data lake in its raw format. Vendors that provide technologies, platforms, and reference architectures for data lakes include Amazon, Cloudera, Google, IBM, Informatica, Microsoft, Oracle, SAS, Snowflake, Teradata, and Zaloni.

 

End Notes

This article taught us data mining repositories’ capabilities, such as data warehouses, data marts, and data lakes. While they all have a similar goal, they need to be evaluated within the use case and technology infrastructure to select the best one for the organization’s needs.

 

About the Author

Harshit Ahluwalia

Our Top Authors

Download Analytics Vidhya App for the Latest blog/Article

Leave a Reply Your email address will not be published. Required fields are marked *