Blogger, Article Defend Veteran Data Integration Tools

Loraine Lawson

I hate to write a blog post pointing to a blog post that, in turn, points you somewhere else. But I'm going to. And here's why:

  1. I missed the original article, and it's a good one. It's about the new breed of ETL tools and how they stack up to the old faithfuls.
  2. Deloitte Manager Vincent McBurney's commentary actually adds quite a lot of value to the original Enterprise Systems article, including this rather unsettling metaphor about IBM's integration solution, "You can't teach an old dog new tricks. So IBM shot the dog and bought a new one. Or at least they kept the parts of the old dog they liked and added new parts."

The original article compared the new breed of data integration vendors' tools with the more established ETL players. The article specifically looked at startups Talend and Expressor Software, but there are other newcomers as well.


The new players do not fare well in either the article or McBurney's assessment. The new guys -- who label themselves as "next-gen platforms" - like to claim that they're more flexible and more scalable than the "legacy" players. And it's largely this marketing ploy that the article and McBurney take to task.


The big takeaway from both pieces is that newer doesn't mean better and old doesn't mean obsolete. In fact, as McBurney pointed out, established vendors have managed learn from their mistakes, work out the kinks, and improve their products. Wrote McBurney:

"So old vendors with old products do not remain successful by standing still. They keep those legacy pieces that work - like the outstanding DataStage GUI and the Transformer - and dump the pieces that don't work. This gives them a better product than new vendors who haven't learn from their mistakes yet and haven't put years of development to build components that are massively scalable."

McBurney didn't pull any punches with the new guys. It's pretty obvious he thinks the established vendors have the better product.


The article is a bit kinder, but both pieces are accusing new vendors of playing marketing games rather than offering substantial innovations. Still, they make a compelling case for why you should be skeptical about "next-gen" players and their claims about established solutions. So check them out before you buy.


Then be sure to read the comments -- because I suspect the new data integration vendors will have a response to McBurney's post, Enterprise Systems, and probably this humble pointer post as well.

Add Comment      Leave a comment on this blog post
Sep 18, 2008 4:50 AM Vincent McBurney Vincent McBurney  says:
I got a grumpy comment from Expressor on that blog post but I didn't think I was hard on them. I just tried to refute the FUD in their marketing message. I think it's pretty obvious that a fifteen year old software product with hundreds of millions of dollars of development put into it has a greater range of functionality and better scalability than a new product.Expressor has a great idea with semantic metadata and offer a cheaper option than the larger vendors. Talend and other open source vendors are bringing tool based data integration to the masses. If you don't have a big software budget they are both worth looking at. Reply
Sep 19, 2008 10:32 AM Yves de Montcheuil Yves de Montcheuil  says:
Loraine, Stephen's article is indeed very interesting and highlights many interesting topics. My only issue with it is that when Stephen interviewed SAS' product marketing manager and IBM's Data Stage product strategist to comment on the marketing messages of their new competitors - well, the outcome was more or less expected. But this was the topic of the article, no grudge here.I posted a comment in response to Vincent McBurney's post so I will not redevelop here what I replied to him. Let me just comment on "newer doesnt mean better and old doesnt mean obsolete". You are absolutely correct (and Microsoft proved your point when they released Vista). But newer technologies can still be better than older technologies - and they often are. Why are all these people using Linux when there are great, robust, and solid good old IBM AIX and HP-UX? And aren't you missing this high performance, low-resource VT100? Brand new technologies are often rough around the edges, and need to be polished. But after 2 years on the market, 1.7 million downloads, and over 100,000 active users, Talend Open Studio is starting to be in pretty good shape. Without 20 years of legacy.Yves @ Talend Reply
Sep 22, 2008 1:31 AM bob potter bob potter  says:
Its absurd to think that a small company can win on marketing ploys only. It doesnt work that way. Being a veteran of both start-ups and large enterprise software companies, I can assure you that the only way for the former to succeed is through innovation. And innovation has to happen along two dimensions: product and how it is brought to market.Regarding product, expressor has innovated in two areas; namely around what we call smart semantics and data processing engine performance. We are the only vendor on the market that has developed a semantic metadata abstraction layer enabling us to shield users from having to deal with physical metadata of the source and target systems. We do this through a step we refer to as semantic rationalization, which by the way is highly automated by the algorithms we use in our software. The benefits of smart semantics are manifold ranging from simplifying the development and offering a level of reuse of business rules and other project artifacts not seen in any DI system before. We started with a clean slate and we did learn from our mistakes, which is counter to McBurneys criticism. Our employees are from IBM/Ascential, Informatica and Ab Initio and they experienced firsthand that changing software architectures that are 10 to 20 years old is hard to impossible. Moreover, these software platforms are comprised of software products from different companies, which make the task of unifying these platforms even more challenging. On the second expressor differentiator, product performance, our parallel processing engine is the fastest on the market. Our early adopter customers have benchmarked us against their incumbent technologies, and we have outperformed everyone thus far and welcome additional benchmark opportunities against any competitor any time.Regarding our strategy of selling our product, we dont believe in making software excessively expensive and then requiring dozens of highly skilled consultants to come in and make it work. As Gartner said in August 2007, data integration is too complex and expensive. We can prove business value in one day, while the established vendors need the better part of a week just to install their software.In summary, small vendors cant out market the behemoths. Thats not the way it works. They can only succeed through product innovation, customer adoption, referenceability, and innovative business and pricing models. Lots of small companies have succeeded in this way and we welcome companies to bring us in and check out our marketing claims. What is there to lose, besides a few hundred thousand dollars off your DI software maintenance bill. Reply

Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.