Are Enterprise Architects up to the Challenge of Integrating Cloud Resources?

Loraine Lawson

My friends sometimes jokingly call me Cassandra, after the famous soothsayer of Troy. You may recall that Apollo cursed her so that no one would believe her predictions.


It's because I often share little tidbits-gathered from my years as a reader and reporter-that they consider pessimistic, but I just consider factual. What they forget-and what I frequently point out-is that Cassandra was right. No one listened to her, they hated her prediction, but in the end, she was right: The Trojans really shouldn't have pulled that stupid horse into the city walls. This observation is usually followed by an uncomfortable silence.


I thought of Cassandra after reading this blog post on David Linthicum's presentation at the SynsCon Cloud Computing Expo in New York city. It's written by Brenda Michelson, who posted notes on a ton of presentations from the Expo. Her post on Linthicum's talk includes this warning:


"You can't replace enterprise architecture with cloud computing. You can't replace SOA with cloud computing. You always need an architectural strategy."


It's good advice but, like Cassandra's predictions, I suspect it will go unheeded. After all, as this Cloud Computing Journal article points out, companies don't have a good track record with the architecture-first approach, even with traditional solutions:


"Frequently, this difficult situation is exacerbated by non-technical managers who are not aware of all the effort necessary to integrate data properly. As a result, a potentially dangerous and costly domino effect ensues: these executives overlook or ignore the difficulties associated with integration; underlying architecture is sacrificed for speed, yet still many projects are deployed too quickly; users' first experiences are negative; and ultimately, the development team is left in a permanent state of catch-up."


The article is written by David Bressler, the SOA evangelist at Progress Software, and is a recap of his presentation at Cloud Expo. Overall, he takes a much sunnier view of cloud computing and integration:


"Fortunately, cloud computing - essentially an offering where infrastructure is provided as a service - promises to help businesses overcome these challenges. Briefly, integration initiatives based upon cloud technologies see more immediate results, as they do not mandate a time-consuming infrastructure build-up process."


Maybe, but only if companies follow Bressler's and Linthicum's advice to map out an architectural strategy for success.


Bressler provides an outline of best practices to integrate cloud computing with existing IT resources:


  1. Use a mediation layer. "It's the single most critical architecture enhancement a company can make when using the cloud, because it enables the enterprise to change on their own terms and not be dictated to by the external provider," writes Bressler.
  2. Manage your service level agreements, which he notes can be done via the mediation layer.
  3. Understand that security will be managed differently and plan accordingly.
  4. Don't fear mistakes. To me, this seems more like advice than a best practice, particularly since he predicates it on the belief that cloud mistakes are less costly than traditional infrastructure missteps and that you can easily switch vendors if there are problems.
  5. Change your IT culture to support application integration, rather than infrastructure support.
  6. Map out a strategy for success, which includes envisioning "cloud-based computing as enabling the 'network' to be a single application-delivery platform delineated by service interfaces between components."


You can access Bressler's presentation slides-which actually work much better than most online slides - and a related paper from Progress Software's blog, or you can read Michelson's notes.


One interesting note is that both Linthicum and Bressler geared their presentations toward enterprise architects, which raised a question for me: Are enterprise architects actually up to the task of delivering this kind of leadership? Enterprise architects have a reputation of being too ivory tower. When you start to talk about cultural changes in IT, that's going to require more than vague discussions. Maybe, as I discussed during a recent podcast with Linthicum, enterprise architects should step up their game and provide more practical, even tactical, guidance about how to create their architectural goals.


Gartner anticipates cloud computing will surpass $56.3 billion this year to jump to $150 billion by 2013. They're not alone in predicting a huge growth spurt for cloud computing, according to this article. If those predictions are correct, cloud computing could certainly provide ample opportunity for enterprise architects to step up their role.


Provided, of course, the business and CIO don't treat them the same way the Trojans treated Cassandra.

Add Comment      Leave a comment on this blog post
Apr 1, 2009 4:52 AM David Bressler David Bressler  says:


Frankly, I'm shooting for that uncomfortable silence. It means what you've said strikes to the root of things, whether we like to look there or not.

You're right that architecture approaches often fail, and personally, I think it needs to be a tight loop with a "take what works, and turn it into an architecture so we can repeat it" kinda strategy. I believe that's what AWS (amazon web services is) and I know that's how Cisco's e-commerce portal emerged back in the day.

There was a little subtlety in my presentation/article that might have been lost at 7PM in my delivery, but... I think with the way "cloud" computing eases enterprise integration (at various levels - cloud's a couple of different things), it will make it easier to "do architecture better" and spend more time with it. By better, I mean with more discipline and better communication, so that the challenges to "data integration" become more clear, and non technical managers can understand it better (not because they get smarter/more-technical, but because we, the architects, communicate better in a language they can understand because we have more discipline). Whew.

When it comes to leadership, the short answer is no. Organizations need to find a way to reward the leadership you're talking about, so that it's encouraged and embraced.

Look at the relationship between "control" and "influence" and ask how organizations reward each, and where the power lies. Too often, power lies in "control" (and good behavior is rewarded with more control), where it should like with the influencers. This is a tough problem, but a systemic one. Until people learn to let go, and find safe harbor in influence instead of control, an architecture driven approach will never be respected, and value delivered by cloud computing will be nowhere near what the tea-leaves show us it could be.


PS Thanks for the nice comments, I appreciate it!!!


Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.