cooper to logan - expressor software's semantic data integration blog

« April 2009 | Main | June 2009 »

May 18, 2009

on being from Omaha and eliminating over-engineering

expressor's not a typical software product and I am not your typical technology founder. Sporting a mop of wild hair, having tattoos and riding a Harley are unusual enough, but being from Omaha and attending the University of Nebraska at Omaha like our most famous native Warren Buffet really puts me on the fringe.

In fact, I was reading an article about Buffet's address to the 2009 Berkshire Hathaway shareholders' meeting recently, and found it interesting that Buffet was so disdainful of complex financial models, at one point saying, 'If you need to use a computer or a calculator to make the calculation, you shouldn't buy it.' This got me thinking about the over-engineering I've seen in too many software products and applications.

Which is why I used a technique that I classify as logic reduction to not simply optimize but to actually remove unneeded logic or processing in expressor. Two useful tools are 'Proof by Contrapositive' and 'Model Order Reduction.' These are absolutely invaluable tools to eliminate unneeded complexity. Google them and see if you can find ways to utilize these tools.

Does it work? Here are some highlights.

A custom consumer matching process that once took 250 CPU hours with a leading, 'high-performance' DI tool is reduced to 1.33 CPU hours using expressor.

A business subsidiary parent/child process that once took 168 CPU hours is reduced to 1 minute.

There is an old saying that 'the devil is in the details.' While this is still very true a new one might be 'over-engineering in IT is purgatory.'

- john russell, chief-scientist and co-founder

Posted by expressor software at 1:45 PM | Comments (0)

May 12, 2009

version 2.0 of the expressor semantic data integration system

We've taken the wraps off the next major update to expressor - which features new reporting and bulk-loading tools, and increased connectivity and platform support.

Click here to read the announcement or check out our expanded product pages for more details on the entire expressor semantic data integration system.

- Steve Casey

Posted by expressor software at 8:15 AM | Comments (0)

May 8, 2009

ETL migration webinar

Have you ever considered migrating an existing data integration application to another ETL platform - but stopped because you couldn't justify the required expense and time?

Then join us next Thursday to learn how you can cut your ETL migration cost and time by 80% with an end-to-end software and services solution from expressor software and BitWise.

We're hosting a webinar at 11:00 am EDT on May 14th that will provide analysis of the technologies and business practices driving down DI costs, an explanation of why and when to migrate to expressor and a demonstration of a tool that converts ETL created on other platforms to run on expressor - and enables you to benefit more quickly from expressor's high-performance, affordable data integration system.

Click here to read more and to register.

- Steve Casey

Posted by expressor software at 9:15 AM | Comments (0)

May 4, 2009

the cost of data integration

I have been working in the data integration space for about 15 years now and let me tell you this stuff is expensive. I have always found it odd that the current leading integration software vendors are not bigger financially. Think about it. There are many software vendors that are much younger and have a lower price point and yet yield more profit than the DI vendors. How is that possible?

Well the first reason is that the truth is that most data integration software is not much better than writing custom code. If you are familiar with the space, there is always a stored procedure or some shell script that need to be run due to holes in the product architecture. Let's face it, these products were first designed, and never rearchitected, when data volumes were much smaller and everything was pretty much relational data. XML and standard message formats like SWIFT and EDI were an afterthought. That deceptive product demo shows a simple GUI, but when the rubber hits the road, the wheels fall off.

The real problem is the cost of the software. Every single add-on cost extra: connectivity, 64-bit, real-time, versioning. Why are these extra when they are fundamental to data integration? Even the speed of the hardware is factored into the cost. Do you have any idea how painful it is to decide to upgrade a server because you know it will also increase your software license costs? And this insult is compounded by the fact that the software already costs more than the hardware in the first place.

The real problem is that software is a commodity market. The old software sales practices from the 1990's no longer apply and should not. Back then, many vendors were selling software for outrageous prices because it could be done. Today, that does not work. Look at the technology debacle of 2000. But when you have a large installed base or even worse, are already public, you can't very well go dropping your prices to accommodate market changes.

Need some validation? Take a look at the SEC filings for your technology vendor. See how much business they are really doing. What you will see if that even though they are adding N new customers per quarter, the installed base is not growing by the same percentage. The reason is that old customers are leaving just about as fast as new ones are joining.

I wonder why that is happening?

- john russell, chief-scientist and co-founder

Posted by expressor software at 1:15 PM | Comments (0)