This is a common question that we’ve encountered in the past during our PoC process.

Question:

What ETL tools should I use to get good performance?

Answer :

All ETL tools provide varying levels of performance depending upon the transformations.  In a simplest case, where the transformations required are simple, the number of data sources is few, and high speed is essential, a viable option is to write your own code to generate the data.

ETL tools bring a lot of value:

1. Error handling
2. Documentation
3. Easy to use interface
4. System independence (most run on WinServ, Linux, Unix)

To attain the best possible performance, the ETL solution should be hosted on a separate server from the Infobright server.

There are many ETL tools available.  On the open source side, have a look at Pentaho’s Kettle product, or the Talend offering.  On the commercial side, you can look at Informatica, AB Initio, and Datastage among others (most BI tool vendors like SAS, Cognos, and the like offer an ETL tool of some sort).

Talend has always been my favorite ETL tool. The flexibility that their design allows—and their community and support is top notch.

When i need an ETL tool, I usually use Talend http://www.talend.com/ .

Its a good program, it is steady and performs well: you can also do a lot of other things than ETL. I use it for data migration, deduplication, data transfer… I enjoy the good community and the debugging process.