ie8 fix
madison

SOA, Enterprise Architecture, BPM all the same underneath

By | February 5, 2008, 7:50pm PST

Summary: At the end of the day, SOA, Enterprise Architecture, and Business Process Management all address the same problems

Is it time to fold SOA into Enterprise Architecture (EA)? Is there even really any fundamental difference between SOA, EA, and for that matter, BPM (business process management)?

That’s the question Kyle Gabhart asks in his latest post. (Dave Linthicum also predicted last year that SOA would eventually fold into EA).

At the end of the day, SOA, EA, BPM (and let me add Enterprise 2.0 and enterprise data management) call for similar methodologies, and address the same problems. Are they essentially the same thing? Kyle Gabhart-Is it time to fold SOA into Enterprise Architecture

Kyle notes that at a governance panel at last week’s Open Group conference, members of the audience asked panelists to delineate between governing SOA and governing EA. This proved to be somewhat difficult, as “the line between these disciplines is quickly blurring within many enterprises,” he said.

“We need to look past the labels that are applied and take a more business-focused and goal-oriented approach to education, mentoring, and ultimately solution development. We need to probe more intently with our clients to discover their strategic direction, business drivers, and objectives for one-year, two-year, and five-year timeframes. We may have the perfect service offering for them, but because it is labeled as SOA, BPM, or EA, it may not jive.”

Agreed. Business leaders don’t care if a project is called “SOA,” or “EA,” or “DOA.” They want an initiative that achieves faster time to market, streamlines an outmoded process, or improves front office productivity. Period.

Kyle backs up his assertions with some examples. For example, at one workshop he conducted for the Department of Defense, participants observed that the discussion on SOA principals, methodologies, and best practices “all flowed very nicely” into a broader scope of EA. “The ‘SOA’ label that we had begun with was immaterial to the objective — better governance of enterprise assets and business processes.”

He also observes that his company’s clients often associate SOA very closely with BPM, and visa versa. “It was the same standards, same tools, and the training was about 90% the same,” he said. “The difference was found in the labels attached and the emphasis upon certain key concepts.”

At this point, BPM and EA seem to be in separate camps within organizations, and, significantly, the BPM and EA folks, by heritage, may have tighter relationships with the business, versus the IT-centric heritage of SOA. This may be the main divide at this time. But as Kyle points out, it makes a lot of sense to bring these all together in pursuit of better business performance.

Kick off your day with ZDNet's daily e-mail newsletter. It's the freshest tech news and opinion, served hot. Get it.

Joe McKendrick is an author, consultant and speaker specializing in trends and developments shaping the technology industry.

Disclosure

Joe McKendrick

Joe McKendrick is an independent consultant, editor and speaker.

Joe has performed project work (white papers, articles, blogs, research and presentations) for the following companies in the IT marketspace:

  • CBS Interactive/CNET/ZDNet (this blog)
  • ebizQ
  • Evans Data
  • Gartner
  • IBM
  • Informatica
  • IDC
  • Microsoft
  • Systinet/HP
  • Teradata
  • Unisphere Reseach, a division of Information Today, Inc.
  • WebLayers

Joe has also performed research work for the following sponsoring organizations in partnership with Unisphere Research, a division of Information Today, Inc.

  • IBM
  • Luminex
  • Noetix
  • Oracle Corp.
  • Teradata
  • Informatica
  • International Oracle Users Group
  • Oracle Applications Users Group
  • Professional Association for SQL Server
  • International DB2 Users Group
  • International Sybase Users Group
  • SHARE (IBM large systems users group)

Biography

Joe McKendrick

Joe McKendrick is an author and independent analyst who tracks the impact of information technology on management and markets. Joe is co-author, along with 16 leading industry leaders and thinkers, of the SOA Manifesto, which outlines the values and guiding principles of service orientation. He also speaks frequently on Enterprise 2.0 and SOA topics at industry events and Webcasts, and serves on the program committee for this year's SOA & Cloud Symposium in London. As an independent analyst, he has also authored numerous research reports in partnership with Unisphere Research, a division of Information Today, Inc. for user groups such as SHARE, Oracle Applications Users Group, and International DB2 Users Group. In a previous life, Joe served as director of the Administrative Management Society (AMS), an international professional association dedicated to advancing knowledge within the IT and business management fields. He is a graduate of Temple University.

13
Comments

Join the conversation!

Just In

pigeon's parchment?
seabird20 18th Mar 2008
I had a hard time parsing that! I could not figure out what a pigeon was doing in possession of parchment. And then I realized - pigeons, parchment! I focussed on the apostrophe and not the comma!

Anyway that is rather beside the point (but a pet peeve - spurious apostrophes).

I love the general points, we are in danger of focussing on the wrong bits. SOA is buzz wordy - like relational was buzzwordy in the 80s. I remember that DBASE was a relational database, "Because it supported relationships." Try telling that to Dr. Codd!

We have this goal of not wanting to repeat ourselves. Every time we see ourselves doing the same thing a second time, we think, "Damn, I have done this before, so I will just reuse what I did last time." From the earliest subroutines to complex dynamically discoverable services. The point seems to be, how do I find "interesting" functionality, how do I bind to it, and how do I execute it?"

If you will permit an old person's story, in 1967 I was hanging out at an English Polytechnic school in the computer area. We were taking a 3 week programming class - a rare privilege for 18 year olds back then. I needed a square root function, and was told that it was "Hanging on the coat hook labeled square root." Sure enough there was a coil of paper tape (specially laminated so it would not tear). There were also the interface instructions for it. Which values to have in which registers, memory locations, etc. So, I wrote the code to set up the interface and then copied the paper tape inline into my program. When the copy was finished, I coiled the tape back up, put its paper clip onto it and rehung it on the peg. Now that is early binding! The analogies work.

We have been looking for ever more flexible ways of creating components. We have been looking for as little coupling as possible between our components because that makes it easier to localize changes and reduce unwanted side effects. We have been looking to increase cohesion because keeping "things" that belong together together gives us an easier way to manage the thing. We create deployment mechanisms that allow us to scale our components by adding new instances. We create deployment mechanisms that keep the functional behavior away from the operational details. But still we have trouble with the right level of abstraction to get value out of the components.

That's partially because we have relative complexity in our businesses, we are encouraged to design for the "pretty case" i.e. the case where stuff doesn't break and we don't deal with timing and ambiguity well. The complex business processes are those that deal with the expected but out of tolerance. The retry to get information, the arrival of information in the wrong sequence, the "almost always 1:m constraints". The kinds of things that the various rules working groups have been associated with.

So the hope of having "process state-less" services is a longshot. Nice to strive towards, but not easily achievable. Of course getting the process state out of the base data objects would help - butthat's a discussion for another day.

Chris
0 Votes
+ -
There never was a line
reamon@... 6th Feb 2008
I recall jumping on to the SOA Yahoo group months ago, hoping to more fully understand the prevailing definition and practice of what is referred to as SOA. I came to the conclusion that SOA as stand-alone entity never existed. Oh, there was SOA this and SOA that (and a parody titled exactly that featuring Greg the Architect) but fundamentally, architecture is architecture, whatever the guiding principles may be.

Service orientation is just one way of creating an enterprise architecture (or an application architecture or an integration architecture or ....).

To say SOA is folding into EA is amiss in my opinion. It was never its own thing to begin with.
To further clarify my point, consider what an architecture is. It is a description of components and their relationships. For EA, these components are at the enterprise level. For AA, they are at the application level.

The components in an architecture may be based on the notion of services. Thus, that architecture is service oriented or a service oriented architecture (SOA).

But without the further refinement of scoping, we have no idea what an SOA applies to. Anne Thomas-Manes says SOA is always an enterprise thing. Steve Jones says its a business level thing only.

Well, we already had terms for those--EA and BA. SOA by itself didn't create a new level of analysis/design. It modifies how we approach existing architecture levels.
If we look at Zachman and have practised it, there is no doubt. However the business compnentization and resulting services for business layer and security are two different ways of looking at Zachman framework for EA otherwise no difference in concepts or execution BPEL, UML, MDM all are in EA BPM AND SOA,
Dr. Ravi Sharma (Professional Opinion).
Given the following link http://www.zapthink.com/news.html?id=2518 is this really something to be encouraged?

My experience of EA is exactly as described above, SOA especially at the composite level, and BPM is better focused for delivery at the 'Coal Face'.
0 Votes
+ -
EA is the thing to focus on, SOA isn't
reamon@... 8th Feb 2008
Where EA has failed, due to the factors Linthicum identifies, so too will SOA where SOA is practiced at the enterprise level. I don't think the ZT article is intended to show that enterprise architects will fail at SOA because they lack the skill. Rather, the relative attention SOA has can be used as a lever to move EA forward and make the guidance they provide more useful.

SOA isn't an independent thing. Applied at the application level, service orientation will have minimal, local impact--and suffer all the same bad characteristics Linthicum listed in the article--complex, static, and fragile

Service orientation at the EA level, where the benefits are really useful, faces all the same challenges EA generally faces regardless of the architectural approach--relevancy and scope. Get too abstract and you're ignored. Get too detailed and you get bogged down.
0 Votes
+ -
SOA is the focus EA isnt.
rpegamorrg@... 11th Feb 2008
Intesting point, well made. I think I was a little rushed in the note I put down. It would be wrong to say EA should not have an interest in service delivery, they are one of the key four protagonists.
To say EA and SOA and indeed BPM are all the same thing is really missing the point. EA should focus on the architecural purity and direction of an orgasnisation - the more abstract elements of day to day life. BPM or workflow is really about the day to day running of the business, quick and dirty is the rule of the day. SOA is the missing bit in the middle. The right SOA solution comes from give and take from a collection of interested parties not just one. SOA is something the whole organisation have to be in and understand and isolating this to what is seen as the 'Ivory Tower' of EA in my experience is not a good move.
I would point out that I am generalising here, the result of discussing this in many organisations. The dynamics of a company should always be considered, if an effective efficient EA group are in place, use them extensively. It just seems this is the exception not the norm.
0 Votes
+ -
I'm not sure I agree with your point about SOA being between EA and BPM. SOA isn't in the middle of anything. It's a way to approach an architecture, be it an EA or other common level.

I'll reiterate my earlier point--where EA has failed or is ineffective for some reason, so too will SOA if SOA is the attempt to "fix" EA.
0 Votes
+ -
I think we are getting there...
rpegamorrg@... 13th Feb 2008
I think we are talking about the same thing, but I'm not sure. It depends on the definition of EA.

EA - An entewrprise initiative - the roadmap that everyone is working to
EA - A group of people sitting in isolation as described in the zapthink article described earlier.

The SOA issue is a perspective thing, and I didnt describe it very well. SOA interfaces, in the real world will be a compromise between the need of 'Now' (BPM) and the need of the future (usually prescribed by an EA 'group'). EA can't define services in isolation from other groups in the enterprise, essentially as a dictatorship.
I think you're right .. BPM and SOA definitely fo together. Most companies are now talking of BPM and SOA in the same breath. Check Fujitsu with their Interstage offering at fujitsu.com/interstage or IBM .. where if you search for BPM you get an article on BPM and SOA happy. There is a definite synergy betn BPM and SOA .. and organizations are now realizing it.
0 Votes
+ -
It's ARCHITECTURE, period.
richard.lendvai@... 12th Feb 2008
Kyle and Dave have chosen the right perspective towards all blurring buzzwords in the architecture arena.

In SOA, the A counts, not the S.
In BPM, how can this ever be achieved without Architecture?

So, nothing new for me, I even would like to take it one step further: Architecture in general is about the enterprise's capability to work in an orderly manner, be specific in designs end to deliver results to the business. It has little to do with modelling, paradigms and other hype-stuff. See also excellent recent MIT research and view the blogs on Atos Origin Netherlands (sorry, in Dutch) to get inspired. www.atosoriginblog.nl

Richard.Lendvai@atosorigin.com
0 Votes
+ -
Right you are!
reamon@... 12th Feb 2008
+1.

An architecture will incorporate more styles/approaches than just service orientation. SOA has been a distraction for far too long. Service orientation is great in a broader context. But SOA must disappear--it's an overloaded, ill-defined, and distracting term.
0 Votes
+ -
Cant disagree at all!
rpegamorrg@... 13th Feb 2008
Interesting start to thread; my summary - "EA is about delivery not modelling, paradigms and other hype-stuff". A different message to the usual, TOGAF blah, Zachman blah, governance blah...

EA as a roadmap everyone is following, for the good of the business with an eye on delivery, sounds excellent. I cannot do anything but agree. This does pose a question then - Why does the zapthink article described earlier ring true with every enterprise architect i talk to except you guys?

SOA = Seriously Over-used Acronym! I agree, its just a sales point for companies to sell product to companies that participate in 'Magazine Management'. The original concepts of loosely coupled, business focused, process parts are still invaluable, especially when you think about Web 3.0, SaaS and virtualisation etc. Although you can implement a SOA using pigeon's, parchment and string.
0 Votes
+ -
pigeon's parchment?
seabird20 18th Mar 2008
I had a hard time parsing that! I could not figure out what a pigeon was doing in possession of parchment. And then I realized - pigeons, parchment! I focussed on the apostrophe and not the comma!

Anyway that is rather beside the point (but a pet peeve - spurious apostrophes).

I love the general points, we are in danger of focussing on the wrong bits. SOA is buzz wordy - like relational was buzzwordy in the 80s. I remember that DBASE was a relational database, "Because it supported relationships." Try telling that to Dr. Codd!

We have this goal of not wanting to repeat ourselves. Every time we see ourselves doing the same thing a second time, we think, "Damn, I have done this before, so I will just reuse what I did last time." From the earliest subroutines to complex dynamically discoverable services. The point seems to be, how do I find "interesting" functionality, how do I bind to it, and how do I execute it?"

If you will permit an old person's story, in 1967 I was hanging out at an English Polytechnic school in the computer area. We were taking a 3 week programming class - a rare privilege for 18 year olds back then. I needed a square root function, and was told that it was "Hanging on the coat hook labeled square root." Sure enough there was a coil of paper tape (specially laminated so it would not tear). There were also the interface instructions for it. Which values to have in which registers, memory locations, etc. So, I wrote the code to set up the interface and then copied the paper tape inline into my program. When the copy was finished, I coiled the tape back up, put its paper clip onto it and rehung it on the peg. Now that is early binding! The analogies work.

We have been looking for ever more flexible ways of creating components. We have been looking for as little coupling as possible between our components because that makes it easier to localize changes and reduce unwanted side effects. We have been looking to increase cohesion because keeping "things" that belong together together gives us an easier way to manage the thing. We create deployment mechanisms that allow us to scale our components by adding new instances. We create deployment mechanisms that keep the functional behavior away from the operational details. But still we have trouble with the right level of abstraction to get value out of the components.

That's partially because we have relative complexity in our businesses, we are encouraged to design for the "pretty case" i.e. the case where stuff doesn't break and we don't deal with timing and ambiguity well. The complex business processes are those that deal with the expected but out of tolerance. The retry to get information, the arrival of information in the wrong sequence, the "almost always 1:m constraints". The kinds of things that the various rules working groups have been associated with.

So the hope of having "process state-less" services is a longshot. Nice to strive towards, but not easily achievable. Of course getting the process state out of the base data objects would help - butthat's a discussion for another day.

Chris

Join the conversation!

Formatting +
BB Codes - Note: HTML is not supported in forums
  • [b] Bold [/b]
  • [i] Italic [/i]
  • [u] Underline [/u]
  • [s] Strikethrough [/s]
  • [q] "Quote" [/q]
  • [ol][*] 1. Ordered List [/ol]
  • [ul][*] · Unordered List [/ul]
  • [pre] Preformat [/pre]
  • [quote] "Blockquote" [/quote]
ie8 fix
Click Here
ie8 fix

The best of ZDNet, delivered

ZDNet Newsletters

Get the best of ZDNet delivered straight to your inbox

Facebook Activity

White Papers, Webcasts, & Resources
ie8 fix
ie8 fix