There have been a few articles bouncing about recently on the subject of whether SOA represents an enterprise internal view of SaaS. I picked up the thread initially on Joe McKendrick’s blog and followed it for a while across other sites. The central question was “Is SOA Software as a Service, delivered internally?“. To be honest before reading through this article I was in a similar camp to Harry Pierson, whose reply suggested that SaaS and SOA are different things with the only common element being use of the word service. My only disagreement with Harry’s assertions initially came from the fact that he concentrated on describing service-orientation as a software concept whereas I believe that there are powerful applications for the concepts of service-orientation in business architecture as well. (And if you’re in Europe, Harry, we also have the ITIL/ISO 15000 definitions of service to further confuse the issue!).
Now overall I admit to having been slightly dismissive of SaaS on occasions as I feel that we’re heading towards a market where we will increasingly be able to subscribe to BaaS (sorry – Business as a Service :-)) propositions. I believe that in the future organisations will be able to leverage specialised external capabilities in order to construct an overall value chain rather than just buy software that enables them to continue executing non-differentiating capabilities themselves (for more information see previous post here). I’ve never been in any doubt that SaaS would be a constituent element of such services but I’ve doubted the value of software delivered in isolation – I’ve seen such things very much as a stepping stone to much more complete business services delivered by increasingly specialised providers (some good examples of the kinds of capabilities that I see emerging are discussed here and here).
So, in this context I read through Joe’s article and immediately thought “No, SOA is not SaaS”. I could see that some software propositions could be offered as services but to me the wider shift that’s occurring sees organisations looking to purchase results rather than the tools they would need to struggle to gain results for themselves. In this context, therefore, the idea of different parts of an organisation becoming internal SaaS providers was kind of anathema to me as that was too application centric and placed too much of the business onus on the consuming organisation.
The idea continued to bug me for a few weeks, though, and I eventually came to some accommodation with the concept based on a number of perspectives:
- As an elevator pitch the whole idea has some merit as many people now get SaaS – “I’m buying something that I can get more quickly and more cheaply and which makes my life easier” – whereas SOA is still a pretty ephemeral concept for most. SaaS pushes the right buttons in terms of reuse, economies of scale, standardisation and cost transparency that I feel are absolute imperatives for the organisation of the future and gives us a frame of reference to start moving the conversation up the value chain to business services;
- Interestingly, as an organisation modularises – and then begins to unbundle -each business capability will offer a number of services back into the enterprise. Such services will no doubt have several interaction models to support their service provision, from simple document based interfaces through to service portals that would look pretty much like SaaS from the outside but which represent access to a complete service rather than just the software needed to support the delivery of one – think Amazon for example. As a result if we can get people used to the idea that they offer interfaces for other parts of the organisation to use by leveraging the current buzz around SaaS then that’s probably a good thing;
- Finally there’s the whole conversation about how such services get realised – I mean there has to be software and infrastructure somewhere in the equation. Interestingly I feel that as organisations increasingly grasp the concepts of service provision – driven by SOA, ITIL and SaaS – they will need to grapple with the idea that as a service provider they will need to deliver services quickly, cheaply and reliably with inclusive service management, reporting, billing etc. in order to be competitive. As a result organisations will unlikely be able to sustain expensive, bespoke and plodding enterprise infrastructures and will start to look at external utility computing platforms to take advantage of the economies of scale and repeatability that such platforms provide. At the end of the day a traditional IT department can only provide computing resources to a business capability for a certain level of amortised cost – since the resources are hardly shared at all – and they don’t have the capability to support comprehensive service support infrastructures. In addition, once an enterprise starts to consist of smaller, more specialised providers with complete control over their execution the previously high but shared costs of IT will become more apparent to capability owners and make them realise that either a) their IT support costs are unsustainable given their likely level of revenue or b) they have been subsidising the rest of the organisation and can actually gain access to IT at a far lower cost via utility platforms and providers. When this realisation is coupled with the ability for capability owners to choose alternatives I believe that this will result in a flight to utility platforms for the delivery of the software services needed to underpin the capability. Differentiation for such utility computing platforms will be based on cost, reliability, support for rapid delivery and customisation of services and comprehensive support for ‘service enablement’ i.e. reporting, auditing, billing etc. As a result a third perspective on the question of SOA as SaaS is to turn the issue on its head and say that although internal SOA is not SaaS we will increasingly need to offer businesses the ability to deliver their enabling SOA on a utility SaaS platform based on commoditised and cheap technology – the whole thrust of the industrialisation initiative that I’m working on. As a result there is a value proposition into business capabilities based on migrating away from their existing IT and delivering their services on a 3rd party utility infrastructure – much like the move away from company specific electricity generators in the last century and into the hands of infrastructure providers (I don’t really want to get into the ‘IT doesn’t matter’ argument here but I will say that such enabling platforms fall into that camp, taking care of commodity technology so that service providers can concentrate on how technology is used to support innovative business models instead). Interestingly interoperability also means that each business capability – even those contained within a single ‘enterprise’ – could potentially leverage different utility providers if each offers a proposition most suited to the needs of each capability (and thus the idea that we can – or would want to – enforce a ‘common’ enterprise technical architecture becomes defunct beyond the level of interoperability standards).
So – returning to the original question – is SOA SaaS delivered internally? The answer is no. Does a story built around SaaS concepts give us some leverage to sell the concepts of SOA? I would say yes.
So where does that leave us? Well as a CIO I would be looking for opportunities to lead the shift to utility computing by using the SaaS analogy and helping my business colleagues to understand the advantages of a shift to more modular organisational styles. If we apply SOA at the business capability level then it provides the first sustainable approach to reuse that we’ve seen – if businesses can understand reuse at the level of services that they actually need – rather than have low level object designs waved in their faces as examples – then they will much more quickly understand the benefits on offer. If we couple this with monetisation at this service level and use understandable concepts like SaaS to sell the idea then we have the potential to lead our colleagues into a more sustainable model rather than just be reactive to ever louder demands for less cost or more adaptability. Most importantly, however, such changes open the door for the significant shift – increasing and aggressive partnering to drive innovation in business models. All of these issues represent the new and changing role of the CIO – no longer just someone to manage the IT resources of an organisation but rather the leader who interprets the opportunities that technology presents to change business models and gain competitive advantage for the organisation.
From an IT provider perspective the challenge is both to evangelise the opportunities created by the current seismic shifts brought on by technology change whilst building out the utility computing – nee SaaS – infrastructures that will deliver increased repeatability in service development and delivery. Such platforms will need to be much more than just computing power – they will need to be complete, industrialised solutions to the delivery of computing services. They will require strong service discovery and development factories, bullet proof, virtualised and maximally shared hardware, comprehensive service enablement and provisioning capabilities and finally they will need to be highly automated and support self provisioning of services, billing strategies and rates, service level monitoring and access rights. In short they will need to encapsulate all of the commodity elements of the technology – both runtime and design/development time – and make it simple for service providers to create and deliver the services they need using a production line approach.
To summarise I guess that SOA isn’t SaaS delivered internally but SaaS might increasingly be the way in which organisations – whether ISVs or enterprise customers – deliver their SOAs, helping them to concentrate on their differentiation whilst delivering to an industrialised, utility computing infrastructure using strong service development factories.