Sometimes it feels like this industry resembles the old game of “whack a mole” just when you have whacked the last one, another pops up. The SOA community has recently risen up against the horrible term SOA 2.0 and I remain hopeful that this term will disappear as quickly as it appeared. Now another “threat” is bubbling under: The strange view (to me at least) that Web2.0 is in some way incompatible or even competing with SOA.

Steve Jones has written recently about this latest phony war in an excellent blog item. I guess in some ways it was inevitable: Both Web2.0 and SOA suffer from multiple interpretations and when combined with the blogsphere's love of a good row, it was just too easy to set up an “either-or” discussion.

SOA has also been suffering from an on-going naming and scoping discussion to confuse matters further - although I believe the smoke is clearing from the battle-field as we seem to finally agreed that SOA is not simply Web Services or an ESB and equally that SOA can use ESBs without being tainted and that while we may or may not like the term SOA that is what we are stuck with.

With regard to the SOA versus Web2.0 debate in particular, there has been unhelpful throwing around of what are intended as positive or negative terms depending on your side of the argument: 'lightweight' when discussing Web2.0 or 'enterprisy' when discussing SOA (a debate well covered by Joe McKendrik here and here ). This is dangerous as it leads to the mistaken view that the enterprise and all of its integration problems can be solved by either only the heaviest grade middleware or by only the most fluid Web2.0 style approaches. This can only be a plausible view to some software vendors who believe that their product can solve all known ailments or technology purists who believe their chosen technology can really make the world this simple! Instead as Steve Jones points out, there is a role and requirement for many different technologies and approaches: horses for courses -

One size doesn't fit all, and taking an architectural and particularly a business view of SOA helps you pick the right delivery approach for each service rather than trying to shoe-horn one method onto everything.

Web2.0 is of course more than just a set of technologies - it can also be considered to be a set of business models, and a philosophical approach to developing software around participation and easy access and these are certainly challenging to the attitudes often entrenched in the enterprise.

This is the interesting discussion to be had: how the Web 2.0 technologies, philosophy and even business models can be used in the context of the Service Oriented Architecture. And it is going on in many places for instance here – but somehow, perhaps, it is all less interesting than a good fight.

Posted by rbradley in Market trends • SOA concepts • Web2.0 | Permalink | Comments (0) | TrackBacks (0)

In the current issue of International Banking Systems , SAP’s SOA strategy for banking systems is covered in some detail.

The fact that it is SOA based would of itself not be that interesting except that in this case SAP is very much engaging with the customers, focusing on industry specific problems – the first phase involves 9 banks, all well known names such as Credit Suisse and Standard Bank – and are defining generic banking services. Generic in the sense that they cover capabilities that are required by any bank – say credit card payment processing.

The win for SAP is obvious – they can re-engineer and consolidation their many banking product lines in to a single SOA based framework of plug-together components (reducing on-going maintenance costs and complexity) and the components are being designed with help from their potential customers who will presumably buy what they have themselves designed.

The win for the banks depends on how generic these services are. I should point out that SAP are quoted in the article as claiming that the services are not necessarily SAP specific and approximately one third of the current crop are outside of the scope of SAP’s applications.

This is a great example of the huge amount of work that is going on turning SOA from concept into valuable reality - valuable to vendors and customers alike.

Posted by rbradley in Financial Services | Permalink | Comments (0) | TrackBacks (0)

It has often struck me as strange that while SOA is all about loosely coupling, the principle hasn’t often been extended to user interaction which after all is almost always at either or both the start and end of business processes. Instead even in a SOA environment, most enterprise users still interact almost all of the time with one application at a time through tightly coupled clients or browsers.

This neatly brings us back to the Web2.0 technologies which are after all primarily a revolution in user interaction and user interface. To narrow the focus a little, lets consider only the following elements of Web2.0:

- The evolution of browser based user interfaces (thanks to AJAX) into something richer and closer to the familiar desktop interface. This is exciting because client side computing hasn’t really changed since the browser first appeared in the late nineties.

- The concept of the mash-up as a combination of services into a single client-built interface has spread in the consumer web.

(While user interaction can be created through wikis and blogs, I won’t delve into these primarily knowledge management tools although they have a clear potential role to play in service definition/reuse and governance and I will focus on the social and collaborative implications of Web 2.0 in the next blog item in this series.)

The implication of much better user interfaces – and hence must more functionality embedded comfortably within those interfaces - and the ability to combine connections to multiple back-end systems as mash-ups has big implications in the enterprise: the client can now evolve towards being a peer of the back-end systems. Although clearly still a peer with limitations – it is now capable of initiating actions independent of a single system and capable of interacting with many systems in a much more intelligent way.

Furthermore because it is easy to create these interfaces and mash-ups, the client can become truly decoupled from the applications as they are now capable of rapidly adapting to handle new applications, new data feeds and new services.

A further implication of the simplicity (and productivity that I previously commented on) of some of the Web2.0 tools is that it creates the potential to take central IT function out of the implementation of these clients completely and instead focus them only on the associated governance issues: allowing less IT expert, but more business focused individuals to solve their problems themselves.

You may wonder whether this can actually happen: can non-IT people create their own interfaces. I don’t think the AJAX tools are even close to being mature enough to support this just yet – they are still built by techies for techies – unlike wikis and blogs which are usable by most people. However, I think the concept is completely plausible – look at the way Microsoft Excel is “programmed” by non-technical people to carry out very sophisticated calculations. Imagine if the same could be done with could be done with a end-user focused AJAX combined with an enterprise edition of myYahoo.

Posted by rbradley in SOA Predictions • SOA concepts • Web2.0 | Permalink | Comments (1) | TrackBacks (0)

The National Association of State CIOs (NASCIO) recently published a research brief entitled “Service Oriented Architecture: An enabler of the Agile Enterprise in State Government”. Although targeting government, much of what it says applies to any large organization and as such it provides a good succinct summary of the case for SOA. The fact that the US Department of Justice in particular has chosen to sponsor NASCIO’s enterprise architecture initiative, of which this is a part, is not surprising given the fit many people see between what enterprise architecture in general and SOA in particular can deliver and what justice networks need to do in terms of integration.

As I have said before the justification for SOA in government is clear: an environment consisting of numerous complex organizations and the increasing need to integrate rapidly across and within these organizations requires a high degree of integration agility such as promised by SOA. As the report puts it:

“In today’s world, government leaders expect to be able to move quickly, using all information appropriate to make effective choices that benefit constituents. Enablement of this expectation requires an organizational culture that embraces the sharing of assets and information.”

It is also nice to see some positive deployment stories showing quick return on investment – and of course as they report rightly points out these are not solely due to SOA, improving project management also plays a role.

“This particular project [in Chicago] saw immediate return on investment in multiple departments, for example, an increase of 20% in parking permit revenues and collection of outstanding parking violation collections.”

The collection of fines is a major problem for governments world-wide. As a SOA starting-point, it has a number of benefits: the return on investment is obvious and it is at its heart an integration problem: linking between applications to complete the business process.

As I said at the beginning, the report also covers much of the usual ground – SOA is an architecture and although building on previous approaches, it is new in significant ways. It is particularly clear when it explains SOA’s relationship with enterprise architecture more generally:

“The objectives and ultimate outcome of successful SOA is already embedded in the philosophy of enterprise architecture. The concepts of sharing, and reuse within both the business and technology areas of the organization have been around for sometime. Maybe what is really going on with the “advent” of “SOA” is that there is finally an understanding of the value of architecture, reuse, and services.”

And finally, before recommending that anybody interested in SOA and government go read the report, I will pull out one last quote:

“SOA focuses IT on being business driven. The underlying assumption in SOA is that not everything in technology can be the same, so standard methods and processes must be defined to enable disparate technologies to communicate, regardless of manufacturer or language.”

Posted by rbradley in Government • Market trends | Permalink | Comments (0) | TrackBacks (0)