LOGIN | REGISTER NOW | SUBSCRIBE
 
 
News on Monday
more>>
SharePoint Tech Report
more>>


   

 
 
Download Current Issue
ISSUE 3/15/2010 PDF

Need Back Issues?
DOWNLOAD HERE

Receive the print Edition?


 
blogs tab
ASP.NET MVC 2 Ships
ASP.NET MVC 2 has shipped.
03/12/2010 10:26 AM EST

Microsoft plans 'open' Silverlight analytics framework
Microsoft is going to announce a multipurpose analytics framework for Silverlight at MIX.
03/11/2010 09:51 AM EST

About CSS processing
Two sites that lead to a startling CSS conclusion.
03/10/2010 02:29 AM EST

 

Events calendar tab
SHARE
3/14/2010 to 3/18/2010
Seattle, Wa.
SHARE

Cloud Connect
3/15/2010 to 3/18/2010
Santa Clara, Calif.
TechWeb

Microsoft MIX10
3/15/2010 to 3/17/2010
Las Vegas
Microsoft

SharePointPro Summit & Expo
3/16/2010 to 3/19/2010
Las Vegas
Penton Media

TheServerSide Java Symposium
3/17/2010 to 3/19/2010
Las Vegas
TechTarget


 
print Printable version 
Most Read Latest News Blog Resources
Even with its success, .NET causes some consternation
Despite its popularity, the developer community believes that .NET isn't being fully serve...

Appcelerator dev environment targets Android, iPhone
This unified development environment compiles JavaScript into Objective-C or Java, and wil...

Eclipse creates new sub-projects for Mylyn
The sub-projects are designed to help organizations adopt agile ALM, and they will also br...

Digg!  Digg
Reddit  Reddit
Slashdot  Slashdot
Share on facebook  Facebook
Share on Twitter  Twitter

Services Rise to the Surface


By Larry O'Brien
June 15, 2006 — 
For 40 years, we’ve known that successful software design features self-contained modules that do one thing well. In the 1960s, such systems were described as highly cohesive and loosely coupled; in the 2000s, we call them service-oriented. Service-oriented architectures don’t increase the abstraction of how we approach systems; they reiterate what we already know: It’s good to compose systems from fully functional (and debugged) subsystems.

Discussions of SOAs pay great attention to the presence of the network and its implications for data transport. This is understandable, since spending a few milliseconds transmitting a network message rather than accessing RAM is the same millionfold decrease in performance that sending a piece of paper across the Pacific by container ship bears to a face-to-face conversation. Crippling differences in access time were not unknown in the 1960s either, as anyone with a copy of Knuth’s volume on sorting and searching and a memory of tapes, whether magnetic or paper, can attest.

We are at the dawn of the multicore and manycore era (a friend just went to four cores, and I’ve been spec’ing out a similar system, because my life just doesn’t have enough drama without adding water-cooled electronics to the mix). As I discussed in a previous column (“Under Concurrence,” Feb. 15, page 32), we are poorly prepared. Service-oriented architecture is one of the few current memes that map into the concerns that will soon begin dominating the technical aspects of software development. Patrick Logan, blogging at patricklogan.blogspot.com, puts it succinctly when he says, “SOA in its best possible interpretation is just reality for tomorrow’s programmers.”

By the time you read this, broad access to a feature-complete beta of Windows Vista is likely to be available. I hedge the statement because, at this point, Windows Vista release dates are about as reliable as those for “Duke Nukem Forever.” Even though by all accounts the Vista death march is ending and the operating system is garnering positive early notices, its schedule delays and “resets” are all the testimony needed to say that the challenges for Microsoft’s flagship product are only going to increase in the years ahead.

Even if, as PC Magazine said, “not since the summer of ’95 has the Windows interface taken such a giant leap forward,” (“Why You Want Vista Now!” by John Clyman, May 9) the disappointing truth of Windows Vista is that the underlying architecture has not advanced similarly. Running “link-dump” in the Vista system directories shows all the usual suspects. Refactoring the world’s most widely used operating system is a Herculean task (“the Augean stables,” the wags will quickly say), but let me venture that “SetTapePosition” could be moved from Kernel32.dll.

I’ll take it on faith that backup applications use the Tape API. I’ll take it on faith that a transactional backup requires kernel-mode interaction. But, surely, the Tape API need not be part of the surface area of the kernel of the operating system. Conversely, whatever elaborate disk-access juggling is necessary to make backups transactional is surely complex enough to justify dedicated resources.

What transport performance is to network SOAs, resource contention is to concurrent SOAs: both justification and guiding principle. To run quickly in the manycore era, a system, whether OS or application, must be factored so that access to scarce resources (such as, say, tape drives and file volumes) is highly cohesive and loosely coupled.

Evolving Windows toward a concurrent service-oriented architecture suitable to the manycore era is an awesomely large, risk-filled project. To the extent that Longhorn was that project, Windows Vista is a failure, no matter how nice its GUI. I don’t begrudge a building its internal trusses and beams, and that performance dictated the thickness of the fa?ade in places was not surprising. What is surprising, though, is how little the core of Vista, supposedly the results of a “bet the company” venture, differs from previous Microsoft operating systems. That C and C++ is used? Not an issue. That the operating system seems as monolithic as ever? Big problem.

SD Times editorial director Alan Zeichick has advocated (or at least mused on the benefits of) a non-backward-compatible Windows: For Microsoft to take a great leap forward, it needs to put everything, including the wrath of some existing customers, on the table. It’s too late for Windows Vista to gain anything from such a decision; at this point, any changes in compatibility would just be gratuitous.

The Vista architecture will carry us into the next decade and the threshold of the manycore era. At that point, the operating system will have to change or it will be dramatically ill-equipped to move forward. “Bet the company” was always too dramatic to describe Longhorn/WinFX/Windows Vista. It won’t be next time.

Larry O’Brien is a technology consultant, analyst and writer. Read his blog at www.knowing.net.


Share this link: http://www.sdtimes.com/link/29378
 

Add comment


Name*
Email*  
Country     


  • Comment
  • Preview
Loading