Skip to content

XAware Community

Get the Flash Player to see this player.
Flash Image Rotator Module by Joomlashack.
XAware 5.6
Need Help
Webinars and Events
Advanced Tutorials
Webinars and Events

You are here: Home arrow Blogs

XAwareWeb 20SOAETLData WarehouseData ServicesBusiness Intelligence 1 Feb 2008 6:50 AM
Composite Data Services is not ETL by billm

As you might guess, we are often asked about how XAware is different from Talend, and other ETL providers. After all, both companies are referred to as "Open Source Data Integration". So, these products must be competitve, right? Well, not really.

I know this can get pretty confusing. How can these both be data integration software and yet not really be competitive. This is because they have been designed with different kinds of problems in mind. Composite Data Services versus ETL.

XAware has been designed to simplify, for applications developers, implementers, and maintainers, the problem of enabling transactional read, write, and update access to, from, and between multiple disparate data sources in real time, and to be able to easily implement and manage data access as services - in the SOA and Web 2.0 context. We often refer to this design approach as "Composite Data Services". This approach is more the progeny of, and much more closely related to the "EII" (i.e. Enterprise Information Integration) branch of the Data Integration family as opposed to the "ETL" (i.e. Extract, Transform, and Load) branch of the family.

Talend, on the other hand, comes from the ETL branch of the Data Integration family. Now there is nothing wrong with that, they are just a little different in that branch. The ETL branch of the family focuses on moving lots data between data stores while operating on that data to deal with "impedance mismatches", clean up the data on the way, and change around formats so it can be used for some other purpose. ETL is mostly used for loading a data warehouse, maybe for a Business Intelligence, Reporting, or Data Mining use. ETL designers think about design time process flow definition (think Extract then Transform then Load), batch operations, very high data transfer rates (think millions of records per second), operations that check data for consistency and fix format differences in the process of moving the data (i.e. Data Quality). This is an important set of use cases for sure. Our ETL cousins have been very successful in the data warehouse and batch data movement fields. We are very happy for them.

We at XAware are also very proud of on own success at providing run time applications the ability to connect to and work with multiple external data sources using Composite Data Services. We think more about data views and services design, transactional performance and integrity, and service managment and reuse.

Sometimes we do run into use cases where the approach to use is not so clear, for example a Business Intelligence application that needs access to data from multiple sources on-demand and in real time (our ETL cousins do try to be a little more transactional, more like us, by implementing what they call "change data capture" - I'm told that imitation is the sincerest form of flattery). These cases are rare though. Most use cases are a clear fit for one or the other approach and skill set. Still, because we all go by the same family name of Data Integration, we know it can be confusing. I am sure some confused users do sometimes find the wrong one of us in the phone book and end up with the wrong approach - which can be troublesome and costly. We will certainly try to make it clear what we do very well and what is better left to our cousins.

While most of the time we don't really run into our distant ETL cousins (e.g. Talend) when we are out pursuing our day job, we do tend to bump into them at those cocktail parties where the analysts, journalists, and financial investment types hang out. They can be counted on to ask: "Hey, you two look a lot alike, you have the same family name on your name tags, are you two related? Do you work in the same field?" We reply: "Related? Why yes, but distantly. We don't think we really look that much alike. No, we do very different kinds of work, but we both do what we do very well and there is plenty of work for both of us."

Comments (0)add comment

Write comment
You must be logged in to a comment. Please register if you do not have an account yet.


Get the Flash Player to see this player.
Flash Image Rotator Module by Joomlashack.
Free Training
QuickStart Packages
Image 4 Title
Image 5 Title


Which data source and BizComponent combinations do you most frequently use?


Recent Entries