| CIO Blogs
IT Blog Awards |
Subscribe
RSS news feed
Add to Technorati FavoritesRecent Posts
- Web 2.0, Enterprise 2.0, Agile Enterprise, and Business Technology means what?
- User Expectations are being formed outside the Enterprise
- Innovation Brief
- Olympics 2.0, Miss Bikini and the end of Operating Systems
- Gartner and the Hype Cycle of Emerging Technologies
- Model the power to influence. The new generation of CIOs: ‘Walk in the Shoes’ of the Leaders that Drive Change and Business Growth!
- Foraging for Information versus Marketing
- The Fallacy of separating Web development from IPv6
- Cuil bursts onto the semantic search scene
- iPhone 3G, and Mozilla show user adoption to new levels of Power
Navigate
Search the blog
« The ‘memory reset’ feature | Main | Unavailable Soon or no two are ever the same! »
7 Ways to scare EAI experts (part I)
Enterprise Application Integration. Sometimes, you just can’t help associating it with a absurdist theatre play. Two EAI experts are sitting on a bench, somewhere in a deserted place. They are awaiting the arrival of someone named Godot. But Godot never shows up and the EAI experts seem to have forgotten why they are waiting in the first place. At the end, nothing has happened and nothing has changed. One expert says to the other “shall we go?“ and the other one responds “yes, let’s go”.
Nobody moves.
It used to be an art, or at least a scarce design skill, integrating all these different, heterogeneous systems, let alone knowing the ins and outs of an abundance of EAI tools and platforms. Professionals used the title ‘EAI Expert’ with pride on their business cards and established a firm street credibility among IT peers. But every now and then, some of them seem to think that their very mission in life has become integration, forgetting about both the original business rationale and the evolution of technology, systems and strategy.
They desperately need to unfreeze. And in order to achieve that, they may need to be scared a bit first. Let me introduce you to 7 opening lines to achieve that, based on many engagements with EAI experts (and suppliers of EAI tools for that matter) across many different organizations:
1. Integrating is bad
2. It’s all Open Source now
3. You forgot about the data
4. It comes with SAP
5. Just put it in the server rack
6. Amazon does it for you
7. Only vertical matters
There is sort of an ascending order to it, possibly increasing the shock effect. It all depends on how long one is already waiting for Godot. Let’s see if the first three already work for you:
1. Integrating is bad
There is one real way of being more effective in producing code: write less code (generate it instead, using models, domain specific languages and frameworks). Likewise, a real effective EAI expert should try to decrease the number of heterogeneous systems that need to be integrated, before even considering the first integration project. Consolidating multiple versions of applications, getting rid of systems that really do the same in just a slightly different way or replacing systems by standard packages, can all do the job. Reducing complexity is not about implementing one single, multi-talented integration broker, it’s about avoiding the need for one in the first place. Yes: integrating is bad for you.
2. It’s all Open Source now
There are probably not many organisations any longer that maintain their own, homebrewed operating system or database. But definitely more than one are still struggling with custom-made, outdated and proprietary integration middleware. Have a look at what is available in Open Source: there are many excellent integration brokers available, all based on public standards, using the latest in technology and of a mission-critical robustness. Or simply replace your own middleware by a commercial, industry-strength product. It will at least save you valuable times and resources. Indeed, integrating is bad but making integrating possible is even worse.
3. You forgot about the data
Before you start to link up all sorts of different systems, you first may want to consider what data will be travelling on that fancy Enterprise Service Bus. The immense interest in service-oriented systems has even further downplayed the importance of a well-structured, solid data household. Which is really a fatal mistake: there is no point in connecting services to each other if they don’t have a common understanding of what information they are exchanging. This is why Master Data Management (MDM) is so high on the development priorities of major suppliers such as IBM, SAP and Oracle. Think about it, you might even have ‘MDM expert’ on your business card and be credible for it.
I hope these first three opening lines will already help you to unfreeze some of your own EAI experts. If not, it’s time for further measures. More on that in my next item. In the meantime, don’t wait for anybody: if you have suggestions of your own, we are eager to learn.
Posted by Ron Tolido on September 5, 2006 in
Architecture
| 
Digg this | 
Seed Newsvine
TrackBacks
TrackBack URL for this entry: http://www.capgemini.com/cgi-bin/blog/mt-tb.cgi/60
Comments
# on September 17, 2006 11:18 AM, ronen said:
so you don't like EAI experts... but i didn't get anything from this article.... just bombastic declarations.... the claim "don't integrate" and that will solve my problem is like putting my head in the sand in the middle of the road waiting for a trailer to come.
# on September 18, 2006 10:45 AM, Ron Tolido said:
@ronen: Actually, I truly respect my fellow EAI experts, many of them being near colleagues and even good friends. It’s just that they tend to forget why they are into EAI in the first place. Trying to integrate complex, heterogeneous systems without ever asking ‘why’ or ‘can we first simplify before we start’ is – as far as I’m concerned – the real escapism.
# on September 18, 2006 5:27 PM, reamon said:
Throwing in my 2 bits for each of the items:
1. Integrating is bad.
Agreed. It almost always preferrable to avoid integrating. Integrating adds complexity. It tends to "hide" the business process. It requires additional administration. It can increase flexibility in some cases but it can also *decrease* flexibility. It is better (usually) to have a cohesive, robust application that does all that one needs--but of course that's rarely the reality, is it? Integrating is bad. But it's quite often necessary.
2. It's all Open Source now
This point focuses on "custom-made, outdated and proprietary integration middleware." I'm not sure I'd title this point "It's all Open Source now" but the point is one I definitely agree with--if one is using home-grown integration software one has to ask, "Why the heck are you still doing that?"
3. You forgot about the data
This might be a little bit high on this list. I'm not aware of too many places that "forget" about their data semantics. In most places, that's the first thing they look at--with the intent of doing DB sync. But the overall point is spot on--a data dictionary along with data semantics is key to success. And it should be a supplement to the business process definition.
I'm quite interested in the rest of the points that are to come in future blogs.
# on October 14, 2006 8:42 AM, Movies said:
Movies
# on February 22, 2008 4:55 PM, Digitürk said:
Trying to integrate complex, heterogeneous systems without ever asking ‘why’ or ‘can we first simplify before we start’ is – as far as I’m concerned – the real escapism.
# on March 26, 2008 11:00 AM, Senthil kumar said:
Integrating is bad:
Now Most of the companies are in global market and focusing on automation so everywhere they r giving priority to EAI experts i dont understand why you write this blog
# on March 28, 2008 1:40 PM, Ron Tolido said:
@Senthil Kumar: they should be focusing on their (collaborative) business, not on creating the integration infrastructure to enable it. That's the point I really try to put across. It's all a matter of focus and priorities.
# on July 7, 2008 3:05 PM, Enterprise Application Integration said:
I totally agree with you Ron...trying to integrate complex, heterogeneous systems without without knowing why is absolutely ridiculous.