The Roman general Pompey added a grain of salt to his food, believing it would counteract poisons added by would-be assassins. Thinking himself safe, he was eventually stabbed to death.
So take what you read about Web services … a technology that’s solidly in the “hype” phase of the technology life cycle (the other phases being “disillusionment” and “application”) … not with a grain of salt, but with a spoonful of salsa. Neither will prevent assassination, but at least the salsa will spice things up enough to keep you awake at night. And although Web services might turn out to be a useful architecture, it’s doubtful it will turn out as useful as it is now being promoted, so you need to stay alert to maintain the skepticism you need.
Last week we explored some of Web services’ problems, among them production scheduling. When you distribute a batch job over a bunch of computers running in other companies’ data centers and communicating at unpredictable latencies, run times become unpredictable and long.
The most straightforward technical solution is caching, which gives the user control over both computing power and network performance.
But the vendors that create the components you link together to build a Web services application will want to profit from their efforts. My guess is they’ll prefer an ASP (application service provider) model, where customers pay based on usage. Otherwise vendors would just be renting what amounts to a subroutine library, a pretty uninteresting business and one that’s never generated the kind of wealth that attracts venture capital.
Since caching makes ASP-style billing awfully difficult, it’s probably out. What probably will work, though, is similar: You’ll build or license and install web services applications pretty much as you do with existing architectures — on your own hardware in your own data centers. Once all the ASP-modeled Web services vendors run out of venture capital, they’ll rent you their components as libraries, too. They won’t get rich, but they might stay in business that way, eventually becoming something that looks a lot like a traditional software vendor.
If it doesn’t happen this way, Web services probably won’t emerge from the disillusionment stage.
Pompey, who took his food with a grain of salt, was killed in Egypt, where years later an asp killed Cleopatra, according to legend the most beautiful woman of her time. Is it just coincidence the ASP mentality might just be what kills the beautiful architecture of Web services?