Is there a need for a next gen EA Framework?

 

There are interesting debates all around the globe about whether there is the need for a next gen EA framework.  James Lapalme recently published an excellent article: Three Schools of Thought on Enterprise Architecture explaining the reasons of such debates.   

In this article James identifies three schools of thoughts for EA, each with their own scope and purpose:

  • "Enterprise IT architecting" which addresses enterprise-wide IT, and the alignment of IT with business.
  • "Enterprise integrating" which addresses the coherency of the enterprise as a system with IT is only one component of the enterprise.
  • "Enterprise Ecological Adaptation" which addresses the enterprise in its larger environment

For each of these 3 school of thoughts, James describes not only the scope but also the differences in objectives, principles and assumptions, skills, challenges, insights and limitations. 

The first two approaches are most familiar to EA practitioners. We can roughly associate a generation of EA methodology for each of these schools of thoughts: TOGAF for the "enterprise IT architecting", and the most common Business Architecture methods for "Enterprise Integrating". But there isn’t today a cohesive body of thought for a next gen EA framework to address the "Enterprise Ecological Adaptation". I think this 3rd school of thoughts is useful for enterprises which care more broadly about values and in particular "customers values". The outcomes of the enterprises are not only financial but also more about corporate social responsibilities, for example.  

James extensively end-notes his article and identifies some authors and works that he sees address this 3rd approach:

·          J. Gharajedaghi, Systems Thinking: Managing Chaos and Complexity: A Platform for Designing Business Architecture (2nd ed.), Elsevier, 2006.

·           T. Graves, Real Enterprise-Architecture: Beyond IT to the Whole Enterprise, Tetradian Books, 2008.1

·           J.A.P. Hoogervorst, Enterprise Governance and Enterprise Engineering, Springer, 2009.

·           J. Martin, The Great Transition: Using the Seven Disciplines of Enterprise Engineering to Align People, Technology, and Strategy, Amacon, 1995.

·          K. Smith and T. Graves, An Introduction to Peaf: Pragmatic Enterprise Architecture Framework, Pragmatic EC, 2011.

·          J. Lapalme and D.W. de Guerre, "Can a Re-Discovery of Open Socio-Technical Systems Strengthen EA?" J. Enterprise Architecture, vol. 8, no. 1, 2012, pp. 55–61.

I would add some names like Dominique Vauquier with Praxeme, Alain Wegmann with SEAM and Forrester with Business-centered EA.

 What are the barriers to this 3rd school of thought for EA? Do you agree we need a new framework for the new challenges companies are facing?  Or are today’s adequate? Any frameworks you might suggest?  

Comments

TOGAF: too big "too" fail

You already know my point of view. TOGAF was already outdated 5 years ago. It's good for teaching, like the 7 OSI layers in networking. No more.

Today, with the rise of API, massive distributed systems, cloud, and ecosystems of services from different suppliers EA is not able to cope.
Why?
1/ There are not good frameworks there and worse, there are no good tools to apply them.
2/ cost of implementing EA (hyper expensive proprietary tools and framework and team) is not sustainable today anymore except for companies that SHOULD due to regulatory constraints.
3/ when it's done, it's not accurate anymore or you must freeze the system and the bsuiness for months to let EA team to do its job.
4/ continuous development is killing the notion of AS IS and To Be model. I know there is still a debate between Big design Upfront or not. But the pace of innovation is just too high to predict anything. For me the new Internet Giants like Twitter, Facebook, Google, amazon, ebay showed the way. The re-imagined nearly everything (from datacenter, to DBMS, programming language, API, etc.). Of course, if you are running still SAP, MVS, or old generation packaged applications, then it's not urgent for you to quit old way of doing EA.
5/ TOGAF is not open source ... And it's a way for consulting/teaching firms to earn money, not to deliver value.

There is a need for an automated, near real time vision of the dynamic architecture of the business: systems blocks, and interfaces, and SLA and security control.

EA needs automation, openness, and open source.
EA needs social collaborative tools, so that each architect, dev, business analyst could contribute to the global knowledge. Trying to put information by hand in framework is far too long. Let people write down what they know and force systems to have introspection API everywhere so that data collection is easy. And that's it.

And finally, instead of doing beautiful drawing during hours, let's create solutions that will compute automatically the cost of running your business. How much do I spend now to run my business with my system (easier with Cloud)?

EA needs a revolution, EA practitioners need to quit their Ivory Tower to become part of the business and understand that the best technical solution is the one that does the job AND does not kill the ROI (too costly).

Chief Marketing technology is gonna replace sooner or later the chief EA. Infrastructure will stay, but development and Architects will become part of the business. Agile is a disruptive force.

Of course there will still have some "regulatory compliant" frameworks to use in certain business domains enabling the old EA framework and habits to stay ...

So no need for EA framework if there are no EA anymore

Thank you William for your input... and explaining your extremist position which I am sure will provoke debate.
Of course there is no need of EA framework if there is no anymore enterprise architect. Despite what you said William there is still some need like planning, rationalization, optimization, alignment (even if I think this is not anymore the right word), etc.

Who should be in charge of tooling these actions? What type of new or not so new tools these actions require?

Henry Why not organising a

Henry

Why not organising a workshop? Can Forrester do that?
I will play the role of the "bad guy" and let's discuss this subject and propose solutions!
Today with devops tools and cloud solutions (private or public) and a good social tool you will be amazed to see what can be done.
What i said is : everybody is an architect, and it is nearly impossible to reconcile all point of views. Forget about Togaf, invest in people!
Is google, amazon, linkedin not able to plan and rationalize? This is a myth!

Forrester Business-centered EA school of thoughts?

Henry,

totally agree: James has really well nailed down 3 primary school of thoughts I see in our consulting practice.

To which school of thoughts would you position Forrester's Business-Centered EA ?

BCEA establishes a bridge between 2nd School and 3rd School

Thanks François for the question.
I think that BCEA is not fully in the 3rd school of thoughts. I have not read all contributions that James referenced in his article but I suspect that no one can seriously claim having THE solution for the 3rd school of thoughts. But surely all these show the directions. As William said everyone should be his own architect. The problem is to organize the participation and collaboration between all these supposed to be architects. I feel this is one of the challenge for the next gen EA Framework. BCEA proposes 4 principles for an "agile governance" which applies not only to architects but to any IT and/or business governance. This is a new culture within enterprises in which everyone can be entitled but still requires controls, guardrails, guidelines, sharing same values, etc.

Collaborative architecture

It's hard to explain my vision on a blog, but, the idea is to use:
- automation: gather everything about the architecture automatically when possible and push them in a repository
- collaborative architecture: use people to build the "global" views by chunk and ensure the "global coherence" is achieved.

The idea is also to enable lean culture to disseminate in the organisation. Less discussions and drawings of boxes, more testing, more small improvements. Documentation of all whanges (Rationale of a change is as important as the chnage itself), and involvement of everybody!

Decision are made based on testing and real feeback, not politics and statu quo preservation.

More here: http://blog.resilient-it.com/2013/02/continuous-change-and-innovation.html - especially in the 3rd section Cultivating Outraged.

Frameworks are contextual

Great summary, Henry - thanks!

I'd strongly agree with you on "I suspect that no one can seriously claim having THE solution for the 3rd school of thoughts", because it's just as true for the 'other two' schools.

If we look across the 'first school', we have a wide variety of different frameworks, each with different purposes and emphases, and tackling different parts of the architecture need: TOGAF8 is mainly about IT-infrastructure rationalisation, DoDAF around building the same overall picture for a Defense capability or program, FEAF around IT-consistency in a government-specific context (i.e. neither for-profit nor not-for-profit), SCOR for an end-to-end supply-chain, and so on.

For the 'second school', TOGAF9 is perhaps the archetypal framework (though still quite strongly IT-oriented). However, most Open Group folks would emphasise that (quote) "no-one uses TOGAF out-of-the-box" - it always needs to be customised to the specific context of that business or organisation.

Between 'second school' and 'third school' sit a wide range of industry-specific and/or context-specific frameworks: TRAK (for rail), Frameworx (for telecoms/media), EMMMv (mining) and PACE (defense/aviation) are some that come to mind.

So perhaps the key assertion in the 'third school' is that "the only framework is that there is no framework" - or no "_the_ framework", rather. Instead, there seems to be much more agreement that frameworks are useful rather than 'true', and that the emphasis is more on 'metaframeworks' - the disciplines needed both to create appropriate context-specific frameworks, and to link between them so as to create a stronger sense of the whole-as-whole.

I'd be very interested to be involved in the kind of workshop that William suggests, though for practical reasons - diversity, IP and suchlike - it would probably need to be broader than just under the auspices of Forrester alone?

EA Revolution workshop

Henry, William, Tom: these thoughts and exchanges are refreshing. As I went through Forrester's Business-centered EA, I started to recognize some of the tools (agile, cap maps, balanced scorecards) , strategies and tactics I've been applying throughout our EA organisation deployment mandates. James brought an interesting perspective with his article and William's focus on people call to action could not be more true.

I do believe a workshop to dig further these thoughts, exchange on related practices and lessons from the trenches would be quite interesting and formative.

I personally don't mind if it's under Forrester's auspices: BCEA indeed shows Forrester desire to lead the revolution!

Would love to participate in

Would love to participate in a workshop.

Let's talk of ecosystem rather than architecture

Under extreme pressure by William (!), I will try to summarze my feeling in few words.

The business and IT business keep talking of ecosystem to qualify the next level of an enterprise extended environment... which makes perfectly sense to me as a business an enterprise whatever large it is is just a tiny entity/organism/agent of the big unlimited digital and economy world. Most of interacting bodies that can matter are just ignored or outside enterprise control.

Well the problem is that people keep using a (buzz) word but forget what the original meaning is.
Architecture is about engineering, predefined models, known components, seeking at robustness and stability.
Ecosystem is very different, it's about interaction, dynamism, permanent changes and adpatations, unstable equilibrium.....

Thus I don't think an architectural framework will ever be able to help managing the extreme complexity of an ecosystem (even the wonderful TOGAF :-))

Second, I have a strong feeling that instead of trying to defined an all in 1 framework, it"d would rather be easier and more pragmatic to adress each segment of the big picture with the best appropriate model.
Traditional framework for stable organisms (business, functions, processes, solutions, technologies....)
Integration architecture (or whatever system/network based model) to handle the interactions/relations between those organisms
An eco-based model to depict the surrounding environment including highly verstatil and unknown components

As eco-based something... I suggest to have a look at the complex adaptative systems type approach that have been for sometimes associated to the concept of business ecosystem..... and by the way they are being used to address biological ecosystems.....

There is a need... but for what???

Thanks William, Tom, Daniel and François. Seeing the numerous debate there is definitely a need to address the new issues that companies are facing digital disruption. But should it be a framework? Kevin Smith with PEFF and PETF is proposing for example a meta framework. Meaning guidelines for assembling your framework. Despite Kevin's recent launch, it is too early to judge (too theoritical for the moment). I think though there is an interesting idea in the meta framework : there is definitely a need to change the dimensions. What does that mean? We can use James Lapalme list as requirements for the 3rd school of thoughts and see how existing EA practices should evolve to address these requirements. For examples :
- EA principles <-> Enterprise philosophy (EA philosophy) see Tom Graves' last blogs posts which can form example of such EA philosophy. He wrote also some blogs around enthropy that I think can be interesting. Daniel thinking in complex adaptative systems. And there are others I am sure.
- EA metrics/value <->business outcomes including values+Architecture business value
- EA Governance <-> agile governance
- EA process <->???
- EA Framework <->???
etc
I have not all the answers of all the things which should be treated at another dimension but I do think that as William said we could altogether brainstorm on these ideas and describe what would be the complete but simple (the is is the challenge) next tooling for EA. Is it still a framework and we will create confusion the same way that TOGAF created confusion as people thought TOGAF was replacing Zachman while it was a complement. And I think that the next whatever you call it should complement the same way TOGAF did not try to reinvent what is already good but complement it to address new business challenges. The next step? let me talk to my boss to see if Forrester can organize that... We had already in mind a co development approach with customers but (except William) they are generally not so advanced. Co developping with some advanced people like James Lapalme, Tom, William and all the ones who want to join is a good idea!!!

I have a student that

I have a student that assessing current EA frameworks in order to evaluated their capacity to foster innovation and learning. When I have the results, I will make them available.

Ecosystem Capabilities and Services, not Enterprise Applications

I am very much with Daniel B. We need to manage and coordinate the interactions NOT the implementations.

And we have the mechanism to do this. The essence of a service architecture is encapsulation and we need to make the final, admittedly big, step to embrace business services as the external manifestation of capabilities.

The architecture framework must therefore be federated, enabling plug and play of services from multiple sources.

In my work I advise the very nature of (what we used to call Enterprise) architecture must change to be a) heavyweight and finely detailed in matters governing connections, b) exceedingly lightweight in all other areas.

I expand on this topic in a recent blog:
http://tinyurl.com/bjsu7tv

Eco system, services, shared business outcomes/values,

Thank you David for noticing your excellent blog post which is summarizing our thinking.
For example we developped the term "operational business services" within BCEA with a "business view", an "implementation view" but also an "operational view" for example to complete your business service use as complementary to Business Capabilities. And you are right decoupling implementation from the business view (sometimes called requirements, but I do not like to reuse that wording because the business view can change and generate variants).

We also think David we need to share the same values/business outcomes in an eco system to make it work which is rarely the case today. So the enterprise should not only be the center of overlapping ecosystems but should also work on the business outcomes...
We also think that "agile methodologies" and extending them to the enterprise will not be sufficient to address other agilities than "time to market". And there are many other type of agilities which require other practices like "agile governance" for example.

Thanks for the insight.

Ecosystem, shared Bus, etc.

Agree on the description of the issues and impacts. I'm still not convinced by the solutions proposed.

The main issues of frameworks are :
- accuracy of what is described, or TTL (time to live for the information)
- business value and ecosystems are dynamic by nature. It is impossible to predict their evolution.
- the subjective vision of each stakeholder. For sure, some people do not show all the truth, so you need several ones to get the real vision or to measure.

Thanks William, I think there

Thanks William,
I think there is another level of EA framework so it is not the same framework than what we have seen previously. To address your points :
- a framework and artefacts directly connected to the execution
- dynamicity does not mean "not manageable" or "not controlled". Trading markets are showing they are trying to manage changes.
- offering transparent viewpoints
I am optimistic that we can have a next gen EA framework dealing with these challenges that you described. But you are right should we call it a framework? probably not anymore!!!

Hello, Do you know what

Hello,

Do you know what TOGAF is ?

If i believe what you've written : You don't understand TOGAF at all.

TOGAF is about adapting a set of best pratices : if some point of the method does not fit in your context, don't use it.

This phase or process does not fit ?

Don't use it ...

My company is ultra agile and need a very light governance board ? Or no governance at all because it's breaking innovation?

Use only TOGAF's item which fits ...

If you don't understand that and apply TOGAF like a donkey, you don't do TOGAF, you are doing donkey stuff.

Best regards.