Thursday, 21 March 2013

TOGAF Good or Bad? - Definitely Ugly!


I'm struggling with the value of TOGAF on my current assignment.

Background:


·        We have a need to develop architectural thinking up the 'food-chain' to the C-level to help inform business strategy.


·        The IT Leadership want to focus on: agility & resilience (fast response and ability to thrive under constant change), security, value-for-money and continuously improving User Experience.

·        Unusually for a private company, we are not currently focused on competition (we are a monopoly) nor inefficiency (labour cost - we rarely 'let people go').


·        We have established business-engaged governance mechanisms around Cyber Threat, however, there's cultural resistance to Western-style Governance practices and methods in other areas.


·        The IT department is mostly seen as a Cost Centre, although we are making some progress here.


·        English is the second language for the vast majority of our employees.


·        Experienced Enterprise Architects are rarer than hen's teeth in Hong Kong.


·        A few on my team have been trained and 'Certified' in TOGAF (mostly before I joined) but they have not actually practiced EA (for a wide variety of reasons) since attending the course.


Observations:


·        TOGAF is hard to comprehend for those whose 1st language is English, so it's an even greater challenge for my team.


·        TOGAF just tries too hard and ends up failing on a few counts; it's too comprehensive to be a usable framework and not specific enough to be a methodology. It's almost a philosophy, but a very incomplete and, at times, dangerously misleading one. This is very hard for my team to make sense of and I find myself having very long conversations with them where we end up agreeing that we reduce or focus on one or two of the TOGAF concepts (usually around a deliverable).

·        TOGAF doesn't seem to help very much when it comes to the challenges we face around Consumer-led IT.



·        TOGAF does encourage SOA, that would help our agility & resilience goal eventually, but we're quite a way from developing a genuine component-based, shared-services architecture due to the business organisation, culture and funding mechanisms. And we can't wait for those to change to be able meet our agility & resilience needs.


·        It’s fair to say, TOGAF training does help introduce newbies to some important EA perspectives and does help with common terms and concepts, but it requires a lot of additional buddy-work with an experienced architect to become useful and, what's worse, it conveys the wrong message; 'Enterprise Architecture is complicated and requires a high intellectual capacity to understand'. The latter being the absolute opposite of what I need to convey across IT and the business; 'Enterprise Architecture is about joining-the-dots and making things simpler to understand'.


I do continue to put my team through TOGAF Certification. Why? It's the only credible option here getting newbies up-to-speed on the basics of EA/SA and it helps my staff build their CVs (which is important). I just wish there was a better option that was both less and more:



and


  • more - on the basic mentality required of of an EA (e.g ability to abstract) and why all organizations take a different approach to architecture based on culture, maturity and business priorities.


I sometimes wonder if the authors of TOGAF are motivated to make it more understandable, or, as seems to be the case, keep it obscure and arcane. It certainly felt that way when I was exposed to IAF (Capgemini's Framework) and its zealots for the first time.

An aside here: in a recent chat with the Chief Architect in a well-known travel company, I said I needed the 80 page version of TOGAF - he laughed and responded that he'd implemented an 8 page version! BTW: all his architects are native-English speakers hired in from abroad.

Sunday, 10 March 2013

Enterprise Architecture: Charting the Journey to Business Value

Enterprise Architecture: Charting the Journey to Business Value by Nigel Green

VPEC-T Pesentation to SCiO

VPEC-T A thinking framework as presented to ScIO by Nigel Green

2013: The year the Internet-of-Things takes-off?

I’ve been reading a lot about M2M/’The Internet of Things’, many pundits believe 2013 will be the year the concept finally goes mainstream – it’s been a while since its inception in the late ‘90s!

I have to say I’m among those believers, but I can see a lot of dust in the air before we get to anything that might resemble a ubiquitous eco-system for all-things-Smart.  Here are a few reasons why:
  1. Open standards for connectivity and data interchange will take a while to agree. Having said that, I don’t think Businesses and Consumers will want to see proprietary platforms (the nightmarish vision of an ‘ITunes-for-M2M’)!
  2. Object Identity standards: Everyone seems to be talking about IP v6 for this purpose, very few in the ‘new wave’ of M2M seem to be aware of the Electronic Product Code(EPC)  from GS1/EPCGlobal. I’ve not been close to EPC developments since 2004, but a lot of good thinking was done that tackled many of the issues yet to be resolved in the M2M world – not the least of which, delineation between the identity of the object and the identity of an object’s interface(s) - a debate that continues around the use of IPv6 for identity.
  3. Who will lead the M2M market? Or should it be markets? Will it be consumers leading with ‘Home & Personal’ gadgets , as Alex Hawkinson, founder/CEO of http://smartthings.com/  believes. Or could it be led by large Energy providers with their Smart Grid projects – often subsidized and encouraged by governments? Or will the Telco’s and Network equipment guys to fight back?  Many telcos (mostly outside the US) have been dabbling in this space for ten years or more – I worked with BT on an early Auto ID service back in 2003. Some Telcos have continued to invest, for example; Telefonica recently announced their proprietary ‘Smart M2M’ solution and clearly have global ambitions in M2M services. Meanwhile, the likes of Alcatel-Lucent, Ericsson, HP, Juniper Networks, Motorola Mobility, Qualcomm, Samsung and Texas Instruments are rallying around the OneM2M movement (http://www.onem2m.org). Not to forget Cisco’s long standing ambitions in this space.
  4. Who will lead Enterprise-quality data integration? This would probably include correlation, aggregation and other, signal-based & enterprise app generated data. Think multi-source, data ‘mash-up’ services. This feels to me that this could be the sweet-spot for the more ‘application-and-data-as-a-service’ focused Cloud vendors. Or could be a SI or large software vendor play (SAP or Oracle spring to mind).
  5. Which roles will the Complex Event Processing (CEP) platform vendors adopt in the M2M eco-system? How far can we expect the likes of TIBCO, Oracle and IBM to push M2M/CEP/Big Data combinations within the Enterprise?
But despite the above challenges and battles yet to be won, I do believe we will see far greater deployment of ‘Smart’ things over the next 12-24 months than at any time since those early days of Auto-ID. My bet is that both Smart Grid and consumer-led lifestyle solutions will lead the early adopters.  Don’t, however, hold your breath for pan-industry standards at anything above low-level communications protocols!