A recent discussion provoked by this SOA Post Mortem article got me thinking about Service-Oriented Architectures and Web Services in the context of our geospatial industry. The general argument in that article is Web services is just the latest buzzword to justify a big round of IT funding if you don’t do the business transformation work to actually redesign your enterprise around the idea of providing information services.
Sounds like it makes sense, but what does that actually mean? To put it in the context of OGC Web services, I’d phrase it as, WMS, WFS, WCS, etc. represent the SOA concept that needs to die. Just because you put a Web Feature Service API on your roads data doesn’t mean you’ve service-enabled anything. You may have taken the first required step towards a service, but the business transformation work isn’t done.
However, let’s say you figure out that what your end users actually need is a service that reports the pavement quality and last maintenance date of all roads in a certain jurisdiction for long-term strategic planning on infrastructure upkeep. Then you build that “service” using your WFS and its ability to handle sophisticated queries. Now you have what I’d call a real service providing real identifiable business value.
So the OGC suite of data delivering Web services is more of a meta-service architecture, or spatial data infrastructure, upon which real value-oriented services can be built. But don’t think you’ve created much value in your organization by simply standing up out-of-the-box OGC services without going the next step and figuring out what applications your data users actually need.
Raj,
I agree with the fact that deploying an OGC service only indicates that an important step toward a complete service has been achieved. This makes common sense to me!
But one important element of a geospatial infrastructure is that its framework components should support an evolutionary approach to distributed computing that provides greater business agility. This is certainly the case when we compare current OGC services with distributed information networks in the 1980s.
Another important element of an SDI is that framework components should be accessible to anyone for meeting business requirements including unknown requirements. Just because I may not know how to build an electrical stove I should not conclude that my electrical generating station is not useful. In addition I may not want to know how that stove is buit! So, business transformations take many different flavour and service providers should not be required to know all about them to deliver an efficient service.
Spot on!
What may be needed is to amend (append?) current metadata to facilitate use of data created elsewhere, and available via OGC services. In other words, a set of metadata tools to make it simpler both to identify and apply data and services accurately in a mashed up environment.
Jeff