XML versus EDI
Mike Hogan explains inter-enterprise communications
4 July 2001, 2 pm GMT
I often see articles talking about XML versus EDI. Many companies say that they are using, or planning to use XML because it is faster or cheaper, but they never really explain why this is. So I thought that I would give a quick rundown on the issues.
EDI (Electronic Data Interchange) is an early solution for inter-enterprise exchange of business documents. Of course, such inter- enterprise interaction is at the heart of B2B.
EDI started with the best of intentions and grand expectations, but it has not lived up to these xpectations, as evidenced by about 2% market penetration.
EDI started with standards for defining and handling the various document types, e.g. a purchase order.
One of the primary shortcomings of EDI, however, is that there is no such thing as a standard PO in the real world. Different industries, and even different companies within those industries, have different data needs. For this reason, EDI devolved from the concept of universal connectivity into the practice of one-to-one connectivity.
As a result, EDI only addresses fixed relationships that enjoy sufficient volume to support the investment EDI requires.
Flexibility and speed, or lack thereof, became impediments in the adoption of EDI-based solutions. Cost has also been an issue since
EDI was initially relegated to expensive and proprietary, albeit secure and reliable, VANs (Value Added Networks). However, VANs are now building their capabilities on top of the Internet and, as a result, VAN costs are dropping.
B2B is also larger in scope than EDI. B2B is focused on three primary integration touchpoints, the integration of data, processes and collaboration.
I'll explain these touchpoints in another comment, but by example, companies might want to share a catalog (data), agree to modify a product for a buyer's specific needs (collaboration) and then turn a PO into an invoice (process).
EDI only addresses the latter issue, the process touchpoint.
So, is XML ultimate solution to all problems? No. XML solves a number of issues but it isn't a panacea. XML also suffers from compatibility
issues due to the proliferation of vocabularies. Additionally, while there are efforts underway, Internet-based XML solutions haven't demonstrated that they are as secure, stable, etc. as VANs yet.
However, XML does provide significant benefits versus EDI. For example, a single XML-based infrastructure can be used to address all three major B2B touchpoints. It can serve as the format for addressing data exchange, process integration and collaboration
Because XML can be processed and acted upon by computers, it is also capable of supporting ad hoc relationships and low-volume relationships much more effectively than EDI. XML, because of transformation or mapping technologies can more easily span multiple vocabularies, which EDI cannot do. As a result, XML is capable of supporting many-to-many relationships.
So will companies systematically trash EDI-based solutions and replace them with XML-based solutions? No again.
Why pull out something that is stable and proven and replace it with new technology?
That makes very little sense to me. Instead, I would recommend complementing existing EDI-based solutions with XML-based solutions for the other two touchpoints.
If I were starting from scratch though, I would go right to XML.
Notwithstanding the foregoing (thought I'd throw in a little legalese for kicks), XML does have an ace in the hole with regard to widespread adoption. Because XML more effectively supports many-to- many connectivity it is capable of achieving the network effect.
The network effect defines a situation where the more people that use the many-to-many network, the more valuable those solutions become.
Unlike mainframe applications that can be hidden in the basement of major companies, inter-enterprise process integration is very public and very dependent on the support of trading partners. For this reason, XML will eventually replace most if not all EDI-based
solutions. But it will take time.
The ironic thing is that VANs may end-up being the big winners in this transition, assuming that they can be fleet-footed and that they are willing to cannibalize their existing business before someone else does. That explanation is yet another comment though.
Mike Hogan is VP Strategy and Bizdev at
Poet Software
www.poet.com
Delicious
Digg
Reddit
Magnoliacom
Newsvine
Furl
Google
Yahoo
Technorati
Comments
Post new comment