MonumentSmallOn day 3 it was straight into the streamed conference sessions rather than plenary/spotlight sessions.

Cloud Computing

I went for the cloud computing stream in the morning, and quickly concluded that I’d made a good call. First up was Mark Skilton from Capgemini talking about the work of the Open Group cloud computing working group, specifically the ongoing work on a cloud ROI model, cloud business use cases and a “cloud buyer questionnaire”. Mark’s presentation style was to throw a wall of information at you very quickly, which I found very engaging and he raised some really interesting points. For example…he stressed the importance of separating buyer and seller perspectives/use cases in the cloud discussion (as they tend to be rather blurred together in most discussions), there was some interesting discussions about pricing models/cloud elasticity (i.e. that infinite elasticity is an illusion), and also some debate around cloud architecture models and the “is a private cloud really a cloud?” argument.

The second cloud-related presentation from Francesco Pititto covered cloud in the (Italian) telco sector, and there were a couple of key points for me from that:

  • Cloud adoption gives the tier 1 telcos a means to combat the declining revenue challenges that they have with their traditional business, by re-exploiting their expensively developed assets again, and with longer-tailed business opportunities, i.e. customers they could not have economically serviced before.
  • Telecom Italia are offering IaaS to the Italian market – this has confirmed to me something that I’ve been chewing over for a while. On the face of it, how could a largely country-specific business even hope to compete with a scale IaaS player like Amazon Web Services? Well, the point is that they have some USPs that allow them to compete – whilst they may have a much smaller data centre footprint and less economies of scale, they can guarantee that data remains in territory (and so alleviating some customer concerns re legal/regulatory issues), they are a trusted brand in their territory, and they have an existing SME customer base to sell to (I stress SMEs as this is where the early adoption is).

Later on in the day, a presentation from Enrico Boverino raised an interesting point about ITIL’s CMDB and role it has to play in providing cloud governance. The basic point was that CMDBs require enhancement to be able to adequately support the dynamically assigned and elastic assets of the cloud computing world – and that this will be a barrier to cloud adoption (or at least the adequate service management of cloud-based solutions).


In the afternoon I attended the two ArchiMate sessions which I was keen to hear, just to test my own reluctance to adopt ArchiMate within Smart421. The first one (from Harmen van den Berg) gave an overview which was a good intro to the motivations and capabilities for ArchiMate. I’ve never doubted that ArchiMate is superior to UML for enterprise architecture modelling, but my resistance to taking it any further up to now has been based on the reluctance to learn and educate staff & stakeholders about yet another notation. I don’t doubt that I can be more effective on holiday if I learn Spanish, but English is “good enough” and can get the job done. It’s rather a lazy approach, but I guess it’s all about time/energy investment vs return. So I wanted to “see the light” in this presentation, and to some extent I did. The most powerful aspect of ArchiMate that struck me was the elegance with which you can model and show the relationships through the architectural domains (business, data. application, technology) – this can be done in UML via stereotyping etc but as it was a design goal for ArchiMate, it’s really nicely done and clear in the actual usage. Of course, it’s relatively easy to pick up so the personal investment for anyone used to modelling is small.

The second ArchiMate presentation was a case study from Alexander den Hartog about the use of ArchiMate to model the EA for a global organisation. Ironically, the thing that struck me about this presentation was the enthusiasm of the presenter and the excitement he portrayed about having got the EA of his organisation under control and understood – so I think he probably had the force of will to make this happen even if he hadn’t adopted ArchiMate (though it certainly made it easier). It was quite a tour de force of an EA case study – he’d actually got the entire EA modelled, and more importantly maintained – not something you see that often. Hats off to Alexander! This led me to question what had enabled this to happen in his organisation when so many others struggle to create baseline models and “keep them alive”. The reasons seemed to be:

  • He was the modeller – he owned the model, did all the updates, did his own governance (but had external reviews), so this makes things massively easier to manage. I guess I would describe this as a relatively agile model ling approach. Of course, it’s not so scalable, but it was not a trivial organisation by any means so it was interesting that this was possible. The key observation here is…if you’ve got someone who knows what they are doing, then you don’t need an army of EAs to build and keep a model up to date. As we know, the more people you add, the more complex the communications paths become etc.
  • He had the right relationships (e.g. into the infrastructure teams) so he was alerted when things are changed, so the model does not go stale.
  • The communications strategy was right – the model was published in HTML format and shared with stakeholders and reviewed by them in this format.
  • The change management approach was clear – he published a “change document” that defined what had changed in the model and this was the key review vehicle. In their case, this output was produced and reviewed as a project initiation document which works for them.

Summary of the conference

Overall I thought it was a good event – although I personally didn’t find some of the plenary sessions on the first two days as useful as some of the more specialised afternoon sessions. I’ve met some really clever people – whilst the travel etc is a PITA, there is no substitute for just meeting people face-to-face and hearing their viewpoints, having your preconceptions about EA challenged etc. I liked the way that it didn’t feel like anything was not up for debate, so there was some great interaction in some of the sessions – not that I agreed with all of it, but that’s not the point. Vive la difference.

Out of all the presentations I’ve attended over the three days, I’m pretty sure only one of them, whilst being entertaining, must have involved some kind of hallucinogenic assistance :). So that’s not a bad percentage…

I am continuing to build business models using Archimate, but I am getting bogged down in functions, processes, services and interfaces,

To my mind Archimate seems to move the modeling of the Business (ie that which is not IT see Tom Graves ) only a little way on from the practice of Business Process Modeling.

Archimate still seems to have a “start with IT and from an IT perspective” feel to it. Or put another way “how can we answer the what & why IT is needed”.

This may be the result of my mind set, however it may also be a result of trying to make Archimate compatible with UML, so that for instance an existing class model can be referenced in an Archimate model.

What I am trying to imagine, is how to represent say a centralized hierarchical organization as opposed to a decentralized flat structured organization. Can a decision be represented as a function, process or an interaction?

I have been looking a little more deeply at Archimate 0.1 to see if it will give a useable off the shelf meta model.

The 1st reaction that I have had from a real customer, is that we EA’s may understand it – but not his business managers and executives.

How much time and effort do we expect customer’s staff to put into understanding our EA models?

I guess that it depends, but for the at a most 20 minutes attention span of busy a CxO the answer is that there is not much time and energy available for learning to understand different graphical representations. Indeed it is not just the symbols, but the juxtaposition of them and of the relative diagrams. By which I am thinking in particular of the representation of layers within Archimate.

Coming from a Zachman filled culture the layers are troublesome, and mentally I tip the diagrams on their side! I cannot get my head round the implied hierarchy within the Archimate meta model, after all if a task or activity can be performed by hand, by IT system or by machine where is the hierarchy?

I have been searching for some real life examples of Archimate, and I have not found any. My efforts so far tend to look a little between a process model and a tiered application architecture.

Does any one have any reviews / articles that will shed more light on the use of Archimate?

I think that we could start a thread or discussion on this, if there is sufficient interest. For instance what do we make of the “layering” of Business, Application and Physical dimensions?

May be this has all been discussed at length in the Archimate forum?

Is there a place for peer review of models on the web?

I have just come from the 22nd Open Group EA Conference in London and survived my 1st presentation to the Open Goup. I was on edge all conference hoping that no one would be presenting similar material to mine. They did not and the Sustainable Enterprise Architecture remained my baby.

It has been good to have affirmation that Smart421′s EA proposition matches up to other’s best practices.

This conference was the launch of TOGAF 9, thankfully for my British reserve, not launched with great pomp and ceremony, nor with any Razzmatazz. A little Raz may well have re-enforced the significance of the achievement.

Judging from the mood of the delegates there is confidence in TOGAF 9 and agreement that it is solid as far as it goes.

TOGAF 9 has put on a lot of weight since it was 8, much of it valuable and not at all a middle aged spread into verbosity. Although covering TOGAF 9 at breakneck speed in a couple of presentations on the first full day was not necessarily the best use of conference time. It did show that at least 3 people know TOGAF 9 inside out – I am sure that all of the Open Group members who worked on it are as equally as knowledgeable.

Watch out for Archimate – A meta model to describe Enterprise Architecture, you download the symbols for use in Visio, but to use it properly it needs to be installed in your modelling package. Archimate 2 is due out by the end of the year. I think that it will be a unstoppable force for architecture modelling in years to come.

There were a couple of recurring themes at the conference:-

The hole that is Business Architecture – As Tom Graves pointed out IT is less than 10% of expenditure and so by inference we are missing out on 90% of the market for architecture. The hole is being addressed by the Business Architecture working group, but progress is slow and there is much to be done. I want to encourage all knowledgeable practicing Business Architects to dive in a get it sorted.

Enterprise Architects need more soft skills. We are too geeky! This may also explain the lack of progress on the Business Architecture front as the non IT people can not bear to talk to us or as Paul Homan put it – they do not understand what we say and anyway we take too long to say it!

Some people were trying to extend the scope of TOGAF, Jason Uppal talking about the 10 year lifecycle of architecture artefacts. Amit Bhagwat explored the relationship between leadership, time span of control and Enterprise Architecture.

On balance there was a lot of talk about theory and not much on the practical application of TOGAF. There is great need to publish practical EA experiences and best practices.
We have contributed to the EA community, thrown in our hat with a model for Sustainable EA. It was interesting that some other presentations were near to our offering of Sustainable Enterprise Architecture – Danny Greefhorst talked about Just in Time architecture as part of his Pragmatic Architecture, Paul Homan talked about architecture becoming out of date quickly, if it if not used (Use it or Loose it I suppose) – Martin van den Berg talked of the “Project Start Architecture” the EA Killer Application – But I think that the EA Killer app is the Project Start-up in the our Sustainable EA.


Get every new post delivered to your Inbox.

Join 1,122 other followers