ETL, or Extract, Transform, Load is the process of integrating data from multiple applications (systems), converting them to a single format or structure and then loading the data into the target, often a data warehouse. This process is essential for data analysis, business intelligence, and other related tasks – particularly in businesses with a wide range of data sources and formats to consider.
Selecting the right tool to do so is integral to ensuring the success of not only the specific action, but also for the overall goals and efforts of the business. To learn more about ETL and get a better understanding of how businesses use it, visit our ETL blog post.
The ETL Ecosystem
In the past it was common for businesses to have several ETL tools that operated independently of one another. Today, it is becoming increasingly common to have a single ETL tool as part of a greater data integration effort. To do so allows these processes to be seen as contributing to the overall profitability of the organization – instead of as a siloed data project.
To help businesses who are looking towards a greater data integration strategy, it can be helpful to isolate single components to better understand their part in the whole – as is the case with ETL tools.
Choosing an ETL Solution
The world of ETL tools has greatly evolved over the years to include a much wider range of capabilities and setups. Many come in cloud based versions, giving a greater degree of scalability, availability, and security; with lower infrastructure costs.
There are a few criteria that can help you when evaluating potential ETL tools, it is important to decide which of these will be most essential for your business needs. However, generally speaking some of the most important factors to consider include:
Depending on the needs of your business, the importance of certain functionalities over others will vary. Day-to-day tasks such as data conversion, joining records, filtering, grouping, and combining data should be included with any tool. Some come with the capacity for more advanced tasks such as web methods, rebuilding indexes, handling arrays, and processing unstructured data.
Any ETL solution must be able to connect to Excel, SharePoint, FIX, Salesforce, Hadoop, FTP, and others. Without this functionality, the processing power of the tool is irrelevant as it will not be usable. However, keep in mind that all tools can connect to a database/RDBMS, but only some have native client drivers that enable greater performance when compared with ODBC.
Being able to create effective workflows to organize and connect all of these tasks is key. Some of the most important workflows to establish include: constraint (criteria), branching, grouping, and looping (repeating).
Being able to understand how an ETL package runs is essential – this includes how long it takes, when it started (and ended), who began the progress, if it was successful or not, and in the case of failure, what the error message received was. Execution also includes the capacity to run at predetermined times, restart in the case of failure, and limit the duration of the execution.
This again is where the needs of your business will greatly impact your decision. For those who need greater capacity, many ETL tools include features such as bulk loading or the ability to cache the lookup table, to name a few.
This can mean anything from being able to configure packages to run at the same time, to setting alert frequency, and creating different users and setting their permissions.
The value of each of these criteria against each other will vary depending on the size of your business, the goals you have for your data, and other similar factors.
Leading ETL Vendors
Major tech companies have developed tools with incredible functionality to suit the needs of a wide range of organizations. However, a number of growing tech companies are starting to offer even more features and capabilities such as data profiling, data quality, and metadata for specialized needs and requirements.
Below, you can find a slightly outdated list of ETL tools. We have the latest and greatest version of this list with a much better interface under aimultiple.com. The updated list allows you to sort/filter the results and learn more about the products, hope you enjoy it.
|Name||Year Founded||Status||Additional Features|
|Informatica||1993||Public||-Range of prebuilt transformations
-Embeddable engine for real-time and batch data execution
|Stitch||2016||Private||-Was created from RJMetrics|
|IBM: Infosphere Information Server||2008||Public||-Netezza integration for faster loading|
|Oracle Data Integrator (ODI)||2006||Public||-Separation of declarative rules from implementation details
-ELT architecture can use RDBMS engine
|ETLeap||2013||Private||-Data wrangling enables working off sample data alone|
|SAP Business Objects Data Services (BODS)||2007||Public||-Structured and unstructured data integration
-Web-based DI administrator for repository management
|CloverETL||2002||Private||-Open source based on Java
-Has its own transformation language for complex validation rules
|Microsoft SQL Server Integration Services (SSIS)||2005||Public||-Transformation is processed in the memory, making the integration process in SQL server much faster|
|SAS Data Management||2006||Public||-Access to Hadoop via Impala or Pivotal HAWQ
-Role-based GUI with drag-and-drop functionality
|Matillion||2011||Private||-Tools built specifically for Redshift, BigQuery, Snowflake|
|Talend Open Studio||2005||Public||-Open source ETL architecture|
Virtual ETL, also called data virtualization, is a method for data management that allows data scientists to augment ETL processes by eliminating the data centralization approach before analysis.
ETL tools are good for physical data consolidation projects where data scientists duplicate data from original sources and load it into the enterprise data warehouse. Data scientists focus on data cleaning operations which take quite a while. Though the ETL process can provide you analytics capabilities as you want, it may not meet your expectations for real-time analysis.
On the other hand, in decision support applications data recency is important, virtual ETL solutions are necessary for faster data access.
Leading Virtual ETL Tools
The application of data virtualization to ETL enables organizations to solve the most common ETL tasks of data migration and application integration for various data sources. The leading vendor in the market is Informatica with its tool called PowerCenter.
Some leading Virtual ETL tools include:
|Vendor||Tool||Year Founded||IPO Status||Additional Features|
|Actifio||Actifio Sky||2009||Private||-Delivered as an OVA, VHD, AMI, or as an image from other cloud marketplaces|
|AtScale||AtScale Intelligent Data Virtualization||2013||Private|
|Data Virtuality||Logical Data Warehouse||2012||Private||-Supports standard APIs such as JDBC, ODBC, REST to deliver data to the data consumers.
-You can connect your data source in XML, JSON, CSV, xSV formats and manage data in SQL
|Denodo||Denodo Platform||1999||Private||-Available on leading cloud marketplaces such as Amazon Web Services (AWS), Microsoft Azure and Docker.
-Supports OAuth 2.0, SAML, OpenAPI, OData 4
|IBM||Cloud Pak for Data||1911||Public||-Embedded governance capabilities such as automated data discovery and classification, data masking, data zones and data lifecycle management|
|Informatica||PowerCenter||1993||Private||-Compatible with XML, JSON, PDF, Microsoft Office, and Internet of Things machine data|
|Oracle||Oracle Data Service Integrator||1977||Public||-Provides a virtual relational database interface to applications via JDBC or ODC|
|Red Hat||Red Hat Virtualization Platform||1999||Private|
|SAS||SAS Federation Server||1976||Private||-Compatible with popular relational databases, including DB2, Oracle, SAP, SQL Server, Teradata, and Greenplum|
|Stone Bond Technologies||Enterprise Enabler||2001||Private||-Simple drag & drop interface to auto-generate virtual models that can be consumed by BI reports, web services, and applications|
Interested in finding out more modern ways to manage your data? Our blog covers a wide range of related topics to help you discover what new technologies could help you the most. And we have a series of posts covering data in enterprise.
How can we do better?
Your feedback is valuable. We will do our best to improve our work based on it.