The Oracle Australia and New Zealand Middleware and Technology Blog.

Friday, January 23, 2009

Are you doing SOP’s instead of SOA?

Hi and hope you all had a great Christmas and all the best for the upcoming year!

This is the second blog in our SOA Governance series around a report that came out detailing the ten common mistakes in doing SOA that cause it to fail. I am going to address the second reason which is underestimating the impact of organisational change.  I am very much in agreement with this sentiment and we at Oracle, while we sell SOA technology, believe that SOA is really more a state of mind than anything else.  I also do not think this key issue that gets as much airtime as it should. In fact the mis-application of SOA at the isolated project level rather than the enterprise level largely determines whether you are doing what might be called Service Oriented Projects (SOP’s) or whether you are changing the culture and organisational behaviour with a truly transforming Service Oriented Architecture.

Well let’s get started. In any business initiative that involves technology there are three things that need to work together – people, tools and methodology. This is true with SOA but with an extra emphasis on the people component.  Actually organisational change is really at the heart of SOA.  If we go back to the beginning, SOA is all about making the business more adaptable to change. Why is this an issue? Because the processes and systems we have to support us are so rigid, so slow, so difficult to modify that they have become obstacles to future success.  These process and system difficencies have been  caused by many years of “oiling the squeeking wheel”. What I mean by that is mostly no-one has been really willing to tackle the problems at the core of the enterprise – they just focus on their particular project and seek to deliver it cheaply and quickly.  That is not bad in itself its just that each project done this way adds just a little bit more inter-system dependency, a little more tight coupling, a few more duplicated modules… You get the idea I think.

Why do we take this approach? Well its simple – the metrics by which we are measured encourage us to do so. Fast, cheap, self contained projects are viewed as successful. If we get kudos for this we do it again. Its also simpler than trying to fit in with what other parts of the organisation are doing.

SOA is about changing all this isn’t it? We want to streamline the organisation processes. We want a common set of data definitions. We want a re-usable library of services. We want common tools and consistent methodologies.  The problem should be now self evident - we are talking about cultural change as much as Web Services and XML and UDDI. Guess what is easier to do – change the culture or learn the SOA technologies?  

The bottom line is this – if you don’t try and change the culture of the organisation you are not really doing SOA.  How can I say that?  Well if you use SOA technologies and basically build a point-to-point architecture (so you don’t have to worry about any other project in the organisation) then you are not doing SOA you are doing SOP's.  You are just using SOA tools to build the same things you did before.  Is there any real value in doing that? Apart from learning how to build and use Web Services, ESB’s, BPEL etc - not really.

Is this what is happening? Have people invested in SOA tools but not taken a true SOA approach? Have they then expected to see cross organisation value that didn’t appear and then said “SOA doesn’t deliver what the vendor promised me?”

Recently on Kate Carruthers blog she spoke about why reuse fails. Guess what – its because of the psychology – its about the people – its about resistance to change - its about wanting to keep my project isolated and independent of the other.  Reuse is just one aspect of SOA (albeit an often overhyped one). To really make SOA work we need to think in a process oriented way because that’s how our suppliers, customers and partners interact with us. We need to structure our organisation around  these processes and measure the effectiveness of the lines of business in supporting them. If we do this the all IT projects are subordinate to the process and must work to make it efficient, flexible, fast and available.  Then and only then will any technology – including SOA – really deliver measurable business value.

Wednesday, January 14, 2009

SOA is Dead. Long Live SOA.

Long term Red Room blogger Saul Cunningham along with the newbies Richard Ward and myself from the ANZ team will be looking into how SOA projects should be better run and governed to ensure they deliver their full potential back to the business over the next little while... But what a time to begin such a series of posts, amid the extensive 'SOA is dead' debate that someone described as being to the blogoshere what "Sarah Palin was to late night comedy TV!"

The king may well be dead but SOA isn't, and it's not working down the chipshop either. There is a single indestructible truth at the heart of SOA that will keep it alive and that is this; it makes sense to build something once and reuse it. This is a staggeringly good idea and it's not going away any time soon. It's just a tough sell to convince an industry that has grown up on building new stuff every time a new requirement comes along that there is another way to go.

This is not a new argument. I started work as a COBOL programmer in nineteen eighty something or other and one of the first things I learned was that my development shop had a library of pre-built services - only we didn't call them services back then, they were sub-routines or includes if I remember rightly - that you could use as a programmer to do repeatable utility tasks, date conversion for example. These were battle hardened bits of software that you gave certain inputs too and would deliver you a result. Black boxes that did exactly what they said on the tin . Sub-routines, includes, components, services, call them what you will, reusing something that has been built already delivers speed, flexibility, reliability and is cost effective.

Other industries know this. The highly unfashionable car industry for example knows that building a component once and plugging it in to a shared platform is an idea that works. The Volkswagen Beetle and Audi TT are different cars aimed at very different drivers (different applications aimed at different end users to use the language of the software industry) but are built around a common architecture and share common components. Just look at how many applications the VW group have actually built on the A4 platform.

Slightly more fashionable than any company in the traditional car industry today - no Tesla, that doesn't include you - is Apple. Does Apple understand that building a component once and reusing it in many devices is a good idea? Too right it does. The hardware side of the IT industry has figured it out too, take a bow Dell, Asus, Lenovo - or any other laptop maker you care to mention. Why has proven to be such an un-natural act for the software side of the industry when it's a no-brainer to anyone on the outside looking in?

And then there's the old chestnut about business value and aligning IT with the business. Why are we still having this conversation? I saw this perfectly summed up by David Foster, responding to a post on Nick Carr’s blog via Joe McKendrick at ZDNet. On a seminar entitled Aligning IT with the Business; "Isn’t it a bit strange that there is still a need for this kind of thing? There have been IT organizations in American business for fifty years now. No one sends out seminar invitations for ‘aligning manufacturing with the business’ or ‘aligning sales with the
business’ or ‘aligning logistics with the business.’ Clearly, there are a lot of businesspeople who believe that IT marches to a drummer other than the best overall interests of the corporation."

SOA, alive and kicking, fixes that. The fundamentals of SOA are business fundamentals not technology. One of the favourite phrases of Paul Coby, CIO of British Airways is, "There are no IT projects. Only business projects". And he's right. If you're working in IT and can't see how what you are doing on a daily basis links to the goals of your organsiation then something is wrong. IT exists to support the aims of the organisation, to serve customers better - quicker, faster, cheaper - and SOA is the best way that we in IT have yet come up with to represent what we can offer to our organisations as a set of reusable reliable components. SOA isn't dead, it's in rude health.

-Sean

Thursday, January 8, 2009

SOA is Dead, Long Live ECM?

So, you've probably read the posts from a bunch of people lamenting the passing of SOA to the next world. The thing is, SOA is a fantastic technology that provides so many benefits and capabilities to the modern organisation that it shouldn't be allowed to die without so much as a lowering of the corporate flag or the echoing sounds of a muted trumpet!

Now, most of you know I'm the Blog's content expert - so you're probably asking yourself what does SOA have to do with me.....? The thing is, you see, SOA is really, really important for ECM from the integration and goverance perspective. If I had a dollar for each time I've been told 'ECM is really difficult to implement and enforce' I would have $12 (probably). ECM can be hard to deploy and get user buy-in simply because it adds a layer of complexity to the user and we all know, users' really don't like change. But, what if leveraging an ECM solution was simplified - what if adding, changing or deletion content was easy and looked/felt/smelt no different to the tools being used by the workforce today? This is probably the easiest example of how SOA and ECM can work effectively to enable the deployment of an unstructured content solution for the enterprise.

SOA, by definition, allows services to be built in a manner that promotes reuse. They also help simplify the process for interacting with a solution - like ECM. Most people in the industry understand basic SQL - SELECT, INSERT, MODIFY and DELETE are standard functions that allow the interaction between an application and records within a relational database. But when it comes to unstructured content that is managed within an ECM solution - there are a lot more features and functions that could be called, i.e. security, version-management, rendition-management etc. The API calls required within an ECM solution go beyond the 4 basic concepts that SQL provides and this is where the power of SOA comes into play - allowing the combination of complex API calls within a simplified service layer where an application 'simply' requests that a document be stored or retrieved from the repository. Where an organisation creates (or buys) a library of ECM-related service calls - the integration of the ECM repository to a large number of difference applications (or uses) becomes more straight-forward and allows the 3rd generation ECM model of an Information Infrastructure to be truly deployed. SOA makes this, and other types of integrations, much easier to implement - maybe it is time for a rebranding of what is one of the most useful tools in the IT kitbag - SOA certainly hasn't died!

Paul

SOA is just fine…thank you

"The rumours of my death have been greatly exaggerated" Mark Twain

So…SOA is dead. Apparently it died on the 1st of January 2009, so said Anne in her Burton Blog. With the recession, SOA budgets are being slashed; countless failures abound, with the odd success peaking above the sea of despair. The doom and gloom prophets are all saying, “we told you so”. So we turn our heads to some new saviour, the SOA offspring of mashups, SAAS and clouds.

The wrong approach
So why is there so much disillusionment? Probably because a lot of people just did not get it and the approaches they took were wrong. The recession has been a great opportunity for business to bullet wayward IT strategies that were not aligned to their business. So it is not a bad thing really. In my books, it is a good reality check. On the other hand SOA is old news, and our industry is often guilty of needing something new to talk about. But lets not go and run off to something new, when all it takes is understanding what we have today.

Getting back to basics
There is nothing wrong with SOA; in fact the intention of SOA to support transformation of organisations is pretty profound, but probably overwhelming for some. I agree with Anne when she said that SOA needs to be part of something bigger, such as organisational transformation. I have seen too many instances where organisations think that they are doing SOA, but all they are doing are technology projects focusing on aspects such as service-based integration, resulting often in poor ROI, no service reusability, and achieving no business alignment.

We have been afforded the opportunity, through many years of SOA experience, to reflect on what is required and what to avoid. Anne highlighted 10 reasons why SOA fails ranging from a lack of business alignment and case, through to skills and governance. Her arguments are sound and should be read by everyone who is interested in understanding what issues you need to address. David Linthicum in his five things articles talked about bigger issues at play, such as misrepresentation by vendors, and the requirement of strong leadership. While we can’t change some of things he spoke about, it is still good to be aware of what to watch out for.

Will SOA evolve, yes it will. That is a definite. But even with new approaches, we will still have to have the ground rules in place. There is no quick fix or magical elixir.

In the coming weeks the Red Room Team will discuss some key issues that need to be addressed if SOA-assisted transformation is to be successful.

Tuesday, January 6, 2009

Your Web Presence in 2009

Hi All

There's lots of talk around the Global Financial Crisis and the impact it is going to have on businesses around the world this year. To remain successful, your business must acknowledge its web-presence and the importance it serves to your audience - your customers and partners. Knowing who your audience is and what they are interested in reading will help you publish the right content, keeping interest in your organisation and your products and maintain and grow sales. An interactive site, driven using Web2.0 technologies will go a long way to achieving this, allowing customers and partners to provide you feedback - that you hopefully take on-board rather than ignore - will help you remain relevant to your marketplace. But, equally as important is the content that you publish itself - this is what drives people to keep coming back to your site (assuming you don't just sell widgets online that-it!).

Your web-presence is your online marketing platform. Tell your audience regularly what you are doing, which direction you are taking and what you hope to achieve. Invest in getting your content right - not just on cool technologies within the website that actually provide little value - high-quality, informative and relevant. Your content should be targetted towards your audience and its effectiveness measured. Look to experiment, don't be afraid of failure and try to avoid over-personalised content where possible.

Happy online marketing

Paul

Thursday, January 1, 2009

Happy New Year and Here's to 2009

Happy New Year to all our readers around the world from the team here in Australia.

First of all, I would like to take this opportunity to update you on our contributors here at Oracle. We would like to thank Barry Matthews and Steve Williamson for their contributions during 2008. Barry and Steve have moved to different roles within the organisation. We do welcome 3 new regular contributors to the Blog though, may I introduce to you...

Sean Boiling - who joins us from the Oracle Sales Consulting team. Sean has been contributing to the Blog for some time now and now becomes a regular contributor. Sean's interests lie in Fusion Middleware covering SOA, Enterprise 2.0, BPM, and the BEA-products we acquired during 2008.
Richard Ward - who joined Oracle Asia-Pacific in November 2008. Prior to this, Richard spent time with Oracle South-Africa - his homeland - and with one of our partners here in Sydney. Richard will focus on SOA with Saul Cunningham.
Marc Caltabiano - who is Oracle's Director of Enterprise Architecture for Australia and New Zealand, based in Melbourne.

Sean, Richard and Marc will be posting to the Red Room and join Carl, Saul and myself as regular contributors for 2009. Gareth Llewellyn, our roving-reporter in waiting, will continue to contribute on a specialist basis as he did during 2008 for the Oracle OpenWorld.

We aim to deliver more content, more regularly during 2009. We reach our 1-year anniversary in February and look forward to celebrating (virtually) with you all at that time. We have some special features coming up this year - starting with a 10-part series around SOA Governance that we know so many people are interested in (from the feedback you give us). We will continue with the Enterprise 2.0 product stack - talkling about our portal offerings and giving some real-world examples of where we have deployed the products for our customers and discussing some of the benefits that have been gained. IDM and Enterprise Architecture will also get some significant coverage this year - driven by some of the current business-challenges faced by organisations trying to support Web2.0 initiatives for example.

Here at the blog, we think that 2009 is going to be an incredibly interesting year around the world considering the current economic environment and the changing drivers for businesses as a whole. There will be a lot of consolidation, merging and acquisition during 2009 and beyond as organisations are challenged to respond to changing conditions and external pressures. This activity will affect our customers, our partners and our competition - the global marketplace for our technologies, products and solutions will change - our challenge is to prepare for this and be in the best position possible when it happens.

The greatest possible challenge for organisations is going to be adapting a business model to the Web2.0 way of doing business. During 2008, it was reported on several occasions that successful businesses were addressing end-user requirements through a collaborative approach. Rather than trying to second-guess what might be popular - these organisations actually listed to what their customers ask for and address these needs. This is based upon the simple principal that if you are selling something that somebody wants at the right price - they will probably buy it!

A Web2.0 approach introduces challenges for an organisation. There is structured and unstructured information to be managed, CRM and ERP systems to be intergrated with, Portals to be deployed and mass-security to manage. Ever-increasing storage requirements need to be met and the IT department is challenged with Green issues. Organisations want a complete solution from a single vendor, it has to be able to be integrated into any legacy system or application and must support an open standards approach to enable ease of development and supportability.

During 2009, we will concentrate on how organisations address these requirements and talk in-depth around Oracle's products and solutions in a non-sales manner. We realise that many of our readers want to learn about our offerings and this approach should provide some context for you all. Of course, if you want us to concentrate on something else - let us know and we welcome any feedback on the blog you want to give us.

We wish you a prosperous New Year

Paul, Carl, Saul, Richard, Gareth, Sean, Marc.