Web Services = end of platform integration costs


CNet has an article on ZapThink’s latest report on the demise of companies paying consultants big integration fees because Web services will make integration a non-issue.  ZapThink is a big fan of Web services and Service Oriented Architecture (SOA) but I think their analysis is on target even if they seem a little overly committed to the Web service way.


I didn’t shell out the $995 to read the actual report but I agree with the concept and this is definitely a key reason for the success of Web services now and in the future.  Someday integration will just be a given because of Web services and you won’t even have to think about it.


But what is really cool about Web services today is that the technology and cross-platform support is already on most platforms.  This makes the integration job a snap.  Got a COM+ object, wrap it in a Web service.  Got a Java Bean, wrap it in a Web service.  Got anything that you want cross-platform access for, wrap it in a web service. Job done.


So I see a ton of consulting money over the next couple years simply taking legacy systems and throwing a web service wrapper around them.  This probably won’t be done just for the sake of having a Web service interface, but it will probably be done so that the marketing folks can see that sales data from the mainframe.  The tools are already there so they should be pretty quick jobs.


The SOA vision will really take hold when all these legacy systems that had a Web service wrapper placed around them so that their data could be accessed for one particular reason, will find that accessing the data for other applications will be a breeze.  Add the fact that any new systems developed will pretty much have web service support by default. Suddenly people will be able to pull information together without having to think at all about who or what platform they are talking to.  You will be able to do from within your spreadsheets, in the VB Winform app you put together, from your server side application or web application, from your Word document…wherever.


It will be really cool.

Comments (2)

  1. Matt,

    Stumbled across this from your article ‘Hello from TechEd 2003’. As a freelance consultant with a vested interest in integrating data from disparate systems, I was initially rather concerned to read your analysis of the ZapThink report. (That’s the end of what’s left of my gravy train, I thought !)

    However, as ever with these things, I mulled it over a bit, and while I understand your thinking on this, I think there is a (very) long road to travel before we get to the scenario you have painted.

    Unless there is something I don’t get (which is possible – because unfortunately my nose is pressed to the legacy grindstone more often than not), the big hole IMHO is the metadata/dictionary stuff. The wrapped-services could be ‘discovered and consumed’ easily enough, but how do you ensure that the quality of the integration is there. Let’s say that a marketing person in a global co. wants to compare sales from a division operating in another country (and load the figures into Excel, say). The local sales are in pounds sterling and the comparison country’s sales are in Euros. Are the developers of the wrapped services really going to have had the foresight to see this sort of scenario ? More problematic would be expanding or collapsing differing reporting hierarchies, as appropriate.

    If there are developments that you feel would address these issues, I be interested to hear about them – and any other comments you may have.

    Thanks,

    Graham.