Talend vs. SSIS: A Simple Performance Comparison

  • I have to admit that I found this review close to useless. Here's why: It's not because of the tests, although the simplistic tests may or may not be valid. The reason I found this "comparison" between two products so unhelpful was the obvious, and rather flagrant, bias this developer had against non-Microsoft products. Comments like "abhor Java" do more damage in my mind to his argument than it helps. I use SSIS and I like it. It's not perfect but I does a very good job. But I'm not just a Microsoft developer. As well as being a SQL Server developer, I am equally capable (my own assessment of course) as an Oracle developer.

    In spite of the results, this only made me want to find out about Talend. Thanks for the comparison and introducing me to a new product.

  • Just wondering, why in Talend job a tFileOutputMSDelimited (multi - schema) is used instead of a simple tFileOutputDelimited (single - schema). It wouldn't most probably change a lot, but well, it would be more transparent.

    Have to agree with this comment. The tFileOutputMSDelimited component adds a huge amount of overhead and also would account for the out-of-memory error. Using a simple tFileOutputDelimited component would be far more efficient and use only a nominal amount of memory. It would also be more efficient to do the filter in tMap instead of as a separate component.

    However, in general, Talend - because it generates code and is very flexible - has a bit of overhead in tracking the progress of the job through the components.

  • Liked the article. I too was an MS only person for about 15 years. But, have expanded to much more in the last 3-4 years or so (IBM DataStage, Informatica, and Talend). Just getting into Talend for last 1 year, and like it a lot.

    Think the author is trying to get an apples to apples comparison. Though it might be better to have both tools write to a separate database instance as the target (maybe Oracle or DB2 to make it neutral) as most ETL will write to a RDBMS target.

    The biggest thing missing for me is, both tests should be run as scheduled jobs and not using the designer. They will behave much differently when not run through the GUI. The comments about requiring tuning the server or OS for Talend, think are not that import, they are 1 time steps that usually our infrastructure people know how to tune for java/linux. We’re not talking about some tuning you have to do inside each job, just the server as a whole.

    Talend has a lot of benefits, including tons of unique components (with more functionality added each release) that would be hard to do the same thing in SSIS.

Viewing 3 posts - 31 through 32 (of 32 total)

You must be logged in to reply to this topic. Login to reply