Le architetture orientate ai servizi (Service Oriented Architecture: SOA) sono state un tema centrale nel campo dello sviluppo software negli ultimi anni.
Un articolo recente su IEEE Software July/August 2012, Service-Oriented Architectures: Myth or Reality?, di Haresh Luthria e Fethi A. Rabhi, esplora esperienze reali di applicazione dei principi SOA evidenziando luci e ombre.
Le conclusioni degli autori:
“Most of the available literature on SOA is broadly optimistic, leading potential users to mistakenly assume that moving to it is a relatively small step. This general misconception was well captured by the application architect for one firm in our study. Talking about the academic viewpoint and his business team’s expectations for SOA, he said, “They think, ‘Buy SOA, and you’re set. All your problems will be solved.’ This is not so.”
We aren’t suggesting that SOA isn’t all that it’s made out to be. In fact, the analytical arguments promoting it are quite sound in describing its potential benefits. But there are some constraints and issues associated with its implementation. SOA neither exacerbates nor alleviates the lack of proper systems engineering, project management, and program governance. Potential users shouldn’t assume that the burden of such discipline can be eschewed just because the processes have been defined as granular services. SOA is an approach to building business systems that requires significant investment in process management and governance mechanisms for the service life cycle, and it might not necessarily be an instant solution to all an enterprise’s business problems.”