Yes, Clemens, there is a Service Oriented Architecture

Clemens Vaster doesn't think SO Architecture is real. Roy Osherove agrees. Udi Dahan disagrees. Rich Turner and Sam Gentile seem to be squarely in "let's drop the A in SOA" camp. This all reminds me of a conversation I had with Ted Neward on his blog back in October. Before we get to the fundamental reasons for keeping the "A" in SOA, let's dispense with a couple of particularly silly arguments against SOA:
 
  • SOA is bogus because there's a lot of hype around it.
  • SOA is bogus because it's nothing new
 
The amount of hype around a particular technology is completely orthogonal to the value of that technology. I don't mind if folks want to criticize the hype surrounding SOA, but if you turn that into a criticism of SOA itself, you are part of the problem. If can't separate the hype from the technology, you're not thinking clearly.
 
The fact that SOA is not new is perhaps its greatest strength. Unlike most folks in this business, I'm suspicious of anything new. I want to build my applications on proven concepts, not the latest fads. It's more important for software architects  to understand Aristotle than it is to understand Don Box. You'll learn more about SOA by reading about Max Weber's work on modern bureaucracy than you will by reading anybody's blog.
 

Service Oriented Architecture and Bureaucracy

 
Seriously. Go read this. Now, read it again and where it says "bureau", think service. Where it says "sending and receiving messages", think, hmm, sending and receiving messages. SOA is just another way of describing bureaucracy. Go find Microsoft's four tenets of service orientation. Those tenets are just a rehash of Weber's analysis of bureaus (of course, Don uses a lot of fancy computer science terminology that Weber would not recognize). The important thing here is that Weber realized that bureaus only made sense within larger system of bureaucracy. It was the whole system of autonomous units working together that made the system so efficient. If you want to achieve the larger benefits of service orientation, you need to develop an architecture that allows services to flourish.  Next up, I'll discuss how you can grow an SOA for an organization.

Posted May 02 2005, 11:58 PM by john-cavnar-johnson
Filed under:

Comments

Udi Dahan - The Software Simplist wrote re: Yes, Clemens, there is a Service Oriented Architecture
on 05-03-2005 12:11 AM
Clemens' post on my blog entry states that SOA does not necessarily include "autonomous computing", what he might define as services working totally in parallel. I think that that may be quibbling the definition, but from an architecture point of view, distributed systems don't make sense without this kind of parallelism.
John Cavnar-Johnson wrote re: Yes, Clemens, there is a Service Oriented Architecture
on 05-03-2005 3:14 AM
I wonder when Clemens changed his mind. Last year his view[1] was much like yours and mine.

[1] http://staff.newtelligence.net/clemensv/CommentView,guid,0339fe35-2328-44be-93e6-f004bd869a84.aspx
Girish Bharadwaj wrote What is SOA? Huh?
on 05-03-2005 8:45 PM
YeYan wrote re: Yes, Clemens, there is a Service Oriented Architecture
on 06-08-2005 6:47 AM
RePost:
http://www.yeyan.cn/Programming/ServiceOrientedArchitecture.aspx

Add a Comment

(required)  
(optional)
(required)  
Remember Me?