Irrational Independent Data Models in the Supply Chain

March 1, 2004
by Tom Dziersk, president & CEO, ClearOrbit A traveler goes to book a flight on the Internet. He doesn't know what airline he wants to fly on, so he downloads

by Tom Dziersk, president & CEO, ClearOrbit

A traveler goes to book a flight on the Internet. He doesn't know what airline he wants to fly on, so he downloads all the flight information from all the airlines. With all that data in his hard drive, he checks all the airlines for the best price and arrival/departure times. He chooses the preferred flight on his desktop. Then he runs a complex synchronization routine to match his updated database against all the airline databases, hoping none of their flight information has changed in the meantime and the flight he wants it still available.

Make sense?

No. It's absurd.

What the traveler did was to construct a unique, proprietary "independent data model" to access the information he needed. What's wrong with this picture? The obvious solution is for this traveler to interact with the data where it exists in its native state, not download it all into his computer. Clearly, constructing independent data models is irrational for booking air travel. It's an inefficient waste of time and Internet bandwidth. And, worse, it is error-prone because it demands reconciliation and synchronization of complex databases that can easily slip out of synch as the user makes business-critical decisions. In fact, the scenario we just described is so irrational, it was never even possible. Today, a traveler can, in real time, simply ping multiple airline databases using a simple integrated Web interface. By doing so, he knows he is working with the latest "inventory" of flights. He transacts with them in real time, and he doesn't have to synchronize or reconcile disparate or stale databases.

If this scenario is so clearly irrational for air travel, why do we tolerate it in the supply chain?!

In fact, we do. At great cost.

Multimillion-item databases are regularly constructed to "unify" the databases of disparate trading partners, ERP systems or electronic catalogs, so users can inquire "from one source" about such things as inventory availability, purchase order status and shipping status. The cost incurred, in terms of time, consulting fees, data storage, and network bandwidth, is beyond calculation, to say nothing of how easy it is to entirely avoid it.

Affordable supply chain software "extensions" exist today that allow companies to avoid constructing independent data models by allowing them to interact with their trading partners' data where it exists in its native state, without having to download and constantly update that data in a third-party independent data model. For instance, ClearOrbit, is putting in place a global drop-ship outsource manufacturing system for JDU Uniphase. JDS Uniphase deals with drop-ship outsource manufacturers on a global scale, but rather than download the ERP data from JDS Uniphase's ever-changing database to its trading partners, or upload their data to JDS Uniphase's ERP system (they run Oracle 11i in a hosted environment), we were able to quickly put in place a collaboration software suite that allows all the trading partners to access the same database, which is housed, hosted and controlled by JDS Uniphase, the originator of the purchase order and the sole point of customer contact. The trading partners interact with that data where it resides naturally, and they can access component inventory, part status, shipping status, datasheets and more all through the Web.

This model is the same as the system we use to book airline tickets: A Web interface provides access to a unified, rationalized database; system-to-system integration is accomplished using Web services. Were we to have installed an independent data model for the global JDS Uniphase deployment, all these global trading partners would have to download, and regularly update, JDS Uniphase data, and JDS Uniphase would have to pull down (and regularly update) the data from its trading partners. Today, JDSU retains complete visibility and control over the disposition of inventory during the staging and shipment processes at supplier locations, including the printing of shipping documents, commercial invoices and bar code labels, with all of the necessary customer information printed in JDSU's format, using JDSU's ERP data. JDSU also remains the single point of customer contact, and the brand the customers associate with the products delivered.

Let's put a stop to the needless cost -- in time, dollars and business efficiency -- of independent data modeling. Let's access the well-developed, mature, affordable software solutions available today to interact with the data where it exists, and we can all fly first class for a lot less.

Tom Dziersk is president & CEO of ClearOrbit, which provides real-time supply chain execution (SCE) software solutions. ClearOrbit was founded in 1994 and currently serves more than 250 leading manufacturing and distribution companies including Cisco Systems, Texas Instruments and Applied Materials. For further information, visit ClearOrbit at