CONTENT
Terug naar community
Magazine
Proceedings
Blogs
Master thesis
Zoeken
THEMES
The CIO speaks
The architect answers
The business decides
Effect of architecture
SOA
BPM
Methods
Principles
Financial services
Public sector
Health sector
Most popular items
 
 
BLOGS
SOA livelyhood
Adrian Grigoriu   
Sunday, 18 January 2009

The good questions about SOA such as the relation to Enterprise Architecture (EA), is it architecture or integration and its success rate have recently re-emerged in dramatic public discussions. Rather than me listing a few of the blogs, please search the web with "SOA death".

With poor results, SOA disappointed the business people. Very true. It was said that SOA has died - this rushed away justifications since SOA pundits felt the sting - while its offsprings such as mashups, BPM, SaaS, cloud computing are alive and kicking. I would like to remark though that BPM, SaaS as ASP, and cloud computing as IT data center outsourcing, existed before SOA came to the fore. Mashups look more like "offsprings" but even they are an implementation of the existing Web Services technology. SOA had its contribution though in this spiral evolution.

Before making any judgment I would like to review what SOA is. It is a style of architecture based on services that can be flexibly orchestrated in end user processes. Integration is only implied here by orchestration. SOA can be applied to a software application or up to a full Enterprise as part of the target Enterprise Architecture. It depends on the scope of your endeavor.

The architecture side of Enterprise SOA, consisting of business services design and their orchestration in processes, would belong to the business people rather than IT. It is their job to "design" the business, if they need it. SOA is something that "you have to do" as pundits say. It is not really "you" in IT, but the business people. SOA as architecture eases integration by providing modularization, encapsulation and "standard" interfaces for services. It does not specify the protocols or the technology choice though.

The integration implied in SOA belongs to the IT even though SOA is primarily about architecture. IT provides, in fact, all the technology - ESB, BPMS... - that facilitates services orchestration, integration, discovery, security, SLAs... This is the SOA you "buy".

SOA failed. The truth is that Enterprise wide SOA projects failed. The reason is simple: it was promoted, manned and driven by IT, solely. Business did not ask for it. It is the result of IT hype and push. It is not the job of IT to design the business.

The EA wide attempts at SOA fail for lack of support from Enterprise Architecture, business people and management. In practice, had business been aware that SOA is about re-engineering the Enterprise operation, their reaction towards IT would be swift and lethal. In fact that lead to the SOA obituary.

From an IT viewpoint, SOA technology is an evolution of distributed components concepts and technology that continues to evolve, and as such, it does not really die. Being not necessarily SOA specific, the technology would survive the premature SOA fate.

The application (suite) SOA should be also well and alive. IT and software houses can engineer applications as software services e.g. SOA at IT applications or suite level, and integrate them with web services and ESB technologies. That is what happens in WOA, the side of SOA that IT can do. This is where SOA was successful, even driven solely by IT. SOA at application level is not dead and will not die soon. For instance,many ERPs are re-engineered now on SOA principles.

What would revive SOA at Enterprise level? Business people look more and more at the Virtual Enterprise, VE,(in business tongue also networked Enterprise, boundaryless company) even if they do not call it as such. The VE is made of partner companies cooperating to implement the processes in the value chain to deliver the products. Parts of a business, links in the value chain, are, in reality, outsourced to partners. This is where SOA needs to be incorporated into the bigger business context to provide the architectural framework for the VE. Cloud computing, SaaS, i.e. outsourcing IT to 3rd parties, become part of the Virtual Enterprise.

Enterprise SOA failed for lack of business support, drivers and proper preparation, given the project size and implications. The technology usually associated with SOA (ESB, BPMS, WS) for orchestration, discovery, integration is alive and well. Idem, SOA at the application and suite levels. Take for instance the trend to design applications suites observing SOA.

The Virtual Enterprise of the future, composed of orchestrated business services provided by partner companies will drive SOA along B2B to provide the agility and technology for interconnecting the networked Enterprise. IT needs to take a lead.






Comments (1)
RSS comments
Written by This e-mail address is being protected from spam bots, you need JavaScript enabled to view it on 18-01-2009 14:40
 
 
Has anyone any experience with a company that is - or has been in the past - successful with enterprise architecture - data driven, or otherwise - but failed with SOA? In other words, is enterprise SOA failing because the architecture style has serious flaws (please keep me posted!) or because organizations are failing to practice architecture on the enterprise level altogether? So, if this is indeed the case, shouldn't we think about implementing better governance practices before we start to abort the precious SOA initiatives prematurely?

 

Only registered users can write comments.
Please login or register.

 

Via Nova Architectura is not responsible for the content of blogs, but authors and readers are asked to adhere the following guidelines. Authors are strongly encouraged to check facts, cite sources, present balanced views, acknowledge and correct errors. Respect copyright, fair use and financial disclosure laws. Please do not disparage organizations, or individuals. Being critical of someone's practice is acceptable, when it is done in a professional manner. Prevent usage of marketing statements. Comments should be relevant to the specific post they are attached to. Spam, flaming, personal attacks, and off-topic comments are not permitted. Readers are requested to notify This e-mail address is being protected from spam bots, you need JavaScript enabled to view it of any violations. The editor holds the right to remove any statements that, in the editors opinion, infringe the above guideline(s). The author receives a notification of this action.