Cost of an ETL tool

We are considering bying an ETL tool to replace our hand-coded scripts. We have no idea about the cost of such a tool. Can somebody give an idea ?

I do not think that there is a fixed price for any of the ETL tool.
You may want to do a tool evaluation process. The vendors will quote you based on your hardware(CPUs/Servers), # of licenses, type of tool(Standard/Enterprise), add ons and how desparately they want your business.
Primary vendors in the etl space are:
Microsoft SSIS
Informatica
IBM-Infosphere (formerly Ascential-Datastage)
Check their websites and call their local sales representative to start the process.

Good Luck

Edwin
...who is biased to 1 of the 3 products mentioned above.

What RDBMS are you using ?

- Jason

I will add Oracle to the above mentioned list. Oracle has pretty decent offerings in ETL/Data Integration, Data Profiling and Data Quality.

Regards,

Dil Mustafa

Don't forget Data Integrator (now owned by Business Objects...now owned by SAP). It plays in the same space as Informatica and DataStage.

jcampbell wrote:What RDBMS are you using ?

- Jason

We use SybaseIQ.

For a tool like Informatica or Data Stage, 100's of thousands of dollars. The good news is that contract help is readily available at reasonable costs unlike ERP tools such as SAP. If you are a small shop, you mileage may vary. I believe these companies have processor based licensing. So if you are using a small Unix box, your price will be less. Also, given the current economy state, I would expect that excellent pricing is available for new customers.

There is also an open source ETL tool called Talend Open Studio. It has the options of outputing code in Perl or Java. I'm looking at it now for use in my own environment over hand coding perl.

Don't forget the cost of training! Much of your "cost" of switching from hand-coded systems -- where you have undoubtedly tons of in-house knowledge and experience with the language you're using -- is in learning how to use the brand new tool. On top of that, each new tools comes with a set of best practices, nuances, and peculiarities that can make using them right from the start an exhausting process, even after you have gone through weeks of training. (SSIS, for example, comes with a brand new expression language. Boy this is stupid, if you ask me. I still get hung up sometimes on it and it has been a major source of frustration and delay on the projects I’ve worked on. But I digress…)

So when figuring out what tool to buy, always consider the man hours it will take to train users on the tool. You can likely find average times by visiting the vendor’s websites and forums.

Add WhereScape RED to your list of options. It's fast, easy to learn, Kimball compliant, metadata driven, creates native database objects (ie. no black box), and is reasonably priced.

Disclosure: I work with the product and do have a bias.

Rob

Up to $250,000, give or take, for one server and depending on the version and platform. You may want to put development, qa and production on one server, but remember your security and development outages will affect production.

Gartner has a Magic Quadrant comparing the products. The last one I saw was a year ago and it put DataStage and Informatica as the leaders (easiest, most functionality, most expensive). Pervasive and Business Objects are shown as the Challengers. iWay, SAS, Microsoft, SAS, Ab Initio, Oracle, Sybase, Tibco, Sun Systems, Cognos, ETI and Embarcadero are also players. If it's not on the Gartner list, it's probably a niche player that does one thing spectacularly, but there will be a small pool of practitioners.

I like DataStage, but that's the one I've used the most. Informatica was easy to learn on my own, but I used Microsoft, DataStage and Sybase before that. If your volume is high, go for the Parallel version of DataStage. If it's low or you have a smaller server, try the Server version (single thread). Server and Parallel have a built in scheduler, but most shops want to make it run along with other server processing.There is also a Mainframe version that works nicely (generates COBOL behind the screen).

It's not going to act like COBOL or shell scripts or MY-SQL queries.

23K USD for Oracle Data Integrator EE per target CPU. This is before any discounts.