Earlier this week I was at one of my favourite enterprise-architecture conferences, the annual Integrated-EA conference in London. It’s oriented towards the Defence context, but does cover the broader scope as well: for example, this year’s programme included presentations on EA in logistics, cyber-security, air-traffic control and the Olympics.
One of the reasons I like this conference so much is that it isn’t solely about IT, but instead usually covers the entire spectrum of the respective context – including the people-issues. There’s also a nice amount of that blunt realism and wry humour that Defence folks bring so well to the table.
Here’s a quick summary from my 30-odd pages of notes from the conference:
Keynote from Air Vice Marshal Jon Rigby. A nice note that a couple of years ago it was ‘career-suicide’ to talk about architecture at senior levels in Defence; now it’s much more the recognised norm. He talked about cyber-security, and commented that “the financial-crisis is the strategic even of our generation – and is yet to play out”. He explored the challenges of uncertainty in the Defence space: “the enduring certainty is MOD’s inability to predict events”. (I was a bit disappointed to hear no mention of either peacekeeping or disaster-management, both of which (to my futurist’s eyes) seem more likely – or at least more common – in the future than warfighting as such.) And he also emphasised the need for a whole-scope – the EA “needs to be the integration of everything, not just information”. A practical example was management of limited-bandwidth: “a small amount of training [‘people-stuff’] can manage the bandwidth [‘technical-stuff’] much more effectively” – and appropriate architectures can help a lot with these real-time trade-offs.
Next up, Mark Foden, of FodenGrealy, on the practical realities of business-change in government contexts. As he put it, “my work is sleeves-rolled-up, with interaction workers”. A large part of this is the shift from ‘builders’ – complicated, big-change, planned, constructed – to ‘growers’ – complex, iterative, nurtured: “a lot of my work is helping folks in government understand how to work in a ‘grow’ way” – “think big, start small, move fast”, ‘working out loud‘. Practical concerns such as overall cost, date of delivery and step-by-step processes are all fundamentally different in a ‘grow’ model – for example, ‘grow’ doesn’t have a Gantt-chart! He pointed to the work of UK Government Digital Service: as Mark put it, “I love their floors” – segments covered with ideas on bits of paper – “I love their walls” – whiteboards and sticky-notes – “I love their windows” – one papered over, with a small torn hole, looking down on the busy street below, with the handwritten caption ‘USERS’. When one of the attendees asked “How do we get the bean-counters to tolerate uncertainty?”, the answer was that support from senior levels is essential – and sometimes we do just have to move on elsewhere, because we can’t make people change.
Then Edwin Swidenbank, from Atkins Defence. The main points I picked up were about the ‘Why’ side of EA: on vision, mission, etcetera “I’ve recently been converted – it does give people a central reference-point”, and “really important to establish values – I’ve seen the value of that”. There was also a very good point about the need to maintain a ‘line of sight’ between ways of working, common capabilities, client-service and strategic goals. The only bit I didn’t like was that he seemed to assert that information-entities were the only items of real significance to the architecture – taking us back to the TOGAF trap again. Oh well.
After that, Sharm Manwani, from Henley Business School – enterprise-architecture in general, rather than specific to Defence. He talked about strategic conversations, and the importance of surfacing conflicts; a brief mention of TOGAF, but added “business tends to be the lightest part of the framework” (to which I would certainly concur). A section on the value-disciplines framework: “product leadership, operational excellence, customer intimacy – choose one of these as your main focus”. Much of it still felt like ‘inside-out’ push-marketing: beginning to look at trust and ‘outside-in’, but perhaps not enough awareness of ‘outside-out’? “The first step in change is creating a sense of urgency”, he said: but how do we create a sense of urgency for action we have to do now to prepare for something that’ll happen maybe twenty years’ time? More usefully, he noted that the key for Agile is to set the time and goal, and allow most else to be uncertain: “I like to distinguish between risk and uncertainty – there’s usually not analysis of degree of uncertainty”.
Then one of the conference regulars, Group Captain Stu Jack, CIO at MOD (UK Ministry of Defence), on the current state-of-play with MoDAF (Wikipedia overview) and MODEM (MoDAF Ontological Data Exchange Mechanism). The main updates there are that MODEM is now complete and in use, and MODAF itself is likely to be merged into and replaced by NAF (NATO Architecture Framework), as part of the work on a more worldwide UAF (Unified Architecture Framework) for the defence context. Stu also focussed on the ‘so what?’ question for enterprise-architecture: “it’s just to make things better”, he said; “it’s not about trying to ‘solve the world’s problems’ – if we try to do that, we will fail”. Coherence was another major theme: “what am I making coherent, and why?” – coherence between business needs, not just technical-level coherence. Much the same too with compliance: “compliance with policy is two-way” – we need to be clear why compliant. And realism: “I can tell them to do it, and they will ignore me; but if give them something useful to them, they will use it” – an important warning, perhaps, for those who cling too much to the role of ‘architecture-police’.
Next, a kind of tag-team presentation between Steve Winter of Raytheon and Pedar Blomqvist and Niklas Häggström of the SESAR (Single European Sky) project, on air-traffic control. The section by Pedar and Niklas was more about the technical side, and (to me) mostly straightforward data/information mapping and exchange: “it needs to be global”, they said, “not national or regional, and central to that is sharing of information”. Steve focussed more on the broader context, the broader realities: “we need to think in terms of ‘kerb to kerb’ – it’s not just about the aviation component of travel”. He looked at the deeper complexities of long-term change-projects: “part of what [the systems] have to address is that when we get there, it’ll all have changed”. He also reminded us that “technology is only one of the issues – social and political issues must be addressed”. (I’ll admit I was very pleased to hear this). Likewise that “technology is merely the enabler – that’s why this is enterprise-architecture, not systems-architecture”. And also that “enterprise-architecture is probably the only tool that can handle this level of complexity”. Worryingly, though he noted that “EA has been booted out of [US] FAA” – but perhaps because it’s still viewed as technology-only there?
The final presentation for Day 1 was from not-their-real-names ‘Gavin’ and ‘Dave’ from GCHQ (part of the UK security-services, together with MI5 and MI6). In the main, it was a (very useful) lessons-learned about how to create adoption of enterprise-architecture. They’d started by developing a bespoke framework and metamodel which, they said, was a mistake, because it was too context-specific; the effort almost fizzled out at one point, with more people maintaining the framework than actually using it… So they restarted, almost from scratch, using MODAF as their base, together with a simpler metamodel, resulting in a much stronger take-up: in fact they complain now that “to some extent we’re victims of our own success”. They too acknowledged the challenges of uncertainty: “we are starting programmes where we have not a clue where they’re going to end up – we just know we need to be on that journey”. And also the need to focus on the non-technical aspects, even in a highly-technical environment: “a lot more soft-systems – we need to do a lot more of engaging the stakeholders”. A consistent enterprise-architecture also has important side-benefits: it “brings new joiners up to speed much faster”, and “also enables what-if review with stakeholders”. Interesting stuff, anyway.
(The last part of the day was a panel-session, for which I wasn’t really able to take notes because I was on the panel itself. All I remember was that someone asked about the difference between business-architecture and enterprise-architecture, to which I, uh, did my usual rant about EA as the literal ‘architecture of the enterprise’, with BA as a domain within that overall space.)
Day 2 began with a session by Dr Jonathan Cook on SOSA, the MOD’s ‘systems-of-systems approach’. In essence, I’d summarise it as a proper whole-of-enterprise view melded into a classic systems-architecture, giving a framework and methodology that’s still somewhat information-oriented but, as he put it, does “operate beyond the information space – all the disciplines in MOD”. (Hooray!) Getting politicians to understand that uncertainty is uncertain does remain hard, but they’re certainly getting somewhere with it: a definite ‘Watch This Space’ item, anyway.
That was followed by an ‘international updates’ session, with Regis Dumond and Stephen Challinor respectively reporting for the French and Canadian equivalents of MOD.
Regis Dumond focussed on the challenges of developing architectures where there are “many stakeholders and roles with high turnover”, and the need to build long-term relationships as people churn through roles – much the same as in the UK defence-context. A theme he returned to again and again was responsibility: “describe the systems landscape… who is responsible?” and “who is responsible for connecting between projects?” He also commented on the need to connect with international standards: France has now abandoned its AGATE architecture-framework – “a total failure”, he said – and is now adopting NAF, the NATO framework, for common language across the international space. The challenge there is that “there is no defined methodology for NAF” – most people seem to use some variant of the TOGAF ADM for this – and also “no data format for architecture interoperability”, a problem that at present is crippling all EA frameworks, not just NAF.
Steve Challinor described what, to me, is the right way to go: “enterprise-architecture is [about] everything in the Canadian Forces”, and “the whole enterprise, not just the technology”. Linked with this is an important shift in role and placement: EA is moving from the CTO/CIO space to the office of Vice-Chief of Defence Staff, with an emphasis on resource-management for the whole of CF (Canadian Forces) – again, exactly the right move. EA is viewed as decision-support, with a theme of “encourage / train / mentor”; communication is key to understanding, and to help manage expectations, though it must be adapted or tailored for the different stakeholders. A nice comment at the end about frameworks, that “people don’t care about what framework we use, they do care that there’s some rigour to it”. All in all, very encouraging!
Next was Michael Fuller, from SCC Global, on the architecture and operation of the Security Coordination Centre for the London Olympics. Again the same necessary awareness of the larger picture” it’s about “the process of business strategy to real practice”, he said, and that it was important for people involved to understand that “technology is not the be-all-and-end-all – it’s an enabler“. Also a focus on enterprise as enterprise: “very important: we had always to remember it was a sporting event with some security concerns, not a security event with some occasional sporting themes”. It was also good to see some solid realism: the greatest security-threat to the Games was not terrorism so much as the English weather… The other key was a recognition that the security involved a very wide range of agencies with very different cultures and concerns: the prime role for the Centre was to act as a bridge between them, across all of the varying timescales and technologies. Overall, remarkably low cost – only $13m (according to my notes, though it may have been £13m); but perhaps the hardest challenge here, as with much enterprise-architecture, is that the real measure of success is to the extent that things don’t happen that otherwise could – which makes it difficult to ‘prove’ return-on-effort and suchlike.
At this stage I went to the ‘unconference’ stream, which is well described in Mark Foden’s post ‘Put a bit of ‘un’ in your conference‘. Between us we had parallel mini-sessions on human in the system; uncertainty and EA itself; and semantics, tools and techniques. It went well, anyway.
Unfortunately, being in the unconference stream meant that I missed several presentations in the main hall:
- by Steve Harris of Garlik Experian, on cybersecurity and cybercrime
- by Andy Ellis of ATLAS (part of HP?), on big-data, MOD’s Defence Information Infrastructure and the challenges of uncertainty
- by Adrian Apthorp and Larry Sears of DHL Express, on application-rationalisation and capability-planning for a large multinational logistics-organisation
- by Kevin Wallis of UK MOD, on the increasing role and challenges of mobile computing in the defence context
The session by Australian Rory Cleland, working with Department of Community Safety in Queensland, was almost brutal in its honesty about the problems of working with middle-management in government: “the middle-managers, they’re all b*stards – sitting in their fiefdoms”. “As soon as it hits the lawyers and accountants it turns to sludge”, he said; “lawyers and accountants think of only one thing: money”. In their case, MODAF did provide a way to go from value (intent) to actual capability. But the challenge was that “emergency services get all of the blame and only a small part of the funding”; and because of this, “it’s a losing battle trying to explain the value of architecture to an accountant – they just don’t care”. He kept returning to this theme again and again: “when it comes to cost reduction – and it always comes to cost-reduction in government”, he said; “do not fight the accountants and lawyers – they will always win!” The only way to win through is to frame everything in their terms: “talk about capability; talk about money”. Overall, perhaps the best tactic here is that we can reframe the need for architecture in terms of managers’ personal liabilities in law: “enterprise-architecture will stop you going to jail” is a point that definitely does get their attention!
Which segues rather nicely into the theme for my own presentation, on the role of the ‘business-anarchist’ in enterprise-architectures. The conference organisers, Ian Bailey and Penny Creed, had put me on as a kind of ‘counter-keynote’ (or perhaps closer to a comedy-act to end the show with a bang? ) – they even explicitly described it in the programme as “The (Tom) Grave(s)yard slot”. Because of that, I did it as more of a performance-piece than a straight formal-presentation, but I think it went well: there was a lot of good discussion afterward, anyway. Here’s the slidedeck, anyway:
Overall, a great conference: I’m already looking forward to next year!
That’s it, anyway: over to you for any comments, anyone?