Web Services Strategies
Beyond the technology, IT strategies for implementation of Web services by Doug Kaye.
|
Read Any Good Books Lately? With IT Conversations on hiatus for a few weeks, I'm going to spend some time revamping its web site and planning new shows. In particular, I want to interview authors of the most interesting books for IT professionals.
If you've recently read (or written) a book whose author you think I should consider for a future IT Conversation drop me a line with the details.
Posted Monday, September 01, 2003 3:10:33 PM 
|
|
|
Web Services are not Distributed Objects. This is an excellent article by Werner Vogels to be published in IEEE Internet Computing. Really...one of the best. His six misconceptions:
- Web services are just like distributed objects.
- Web services is RPC for the Internet.
- Web Services need HTTP.
- Web services need web servers.
- Web services are reliable because they use TCP.
- Web services debugging is impossible.
Regarding REST, he writes, "It is becoming quickly irrelevant for the bigger picture of web services, given that transport independence is surpasing the importance of the ‘web’ part of web services. The REST principles are relevant for the HTTP binding, and for the web server parsing of resource names, but are useless in the context of TCP or message queue bindings where the HTTP verbs do not apply."
Werner writes well, too. I recommend his weblog, and I'm looking forward to his new book.
Posted Monday, September 01, 2003 3:05:20 PM
|
|
|
The Madness? Jeff Schneider of Momentum Software is on a rant about an article in Web Services Journal written by two IBMers, John Medicke and Thomas Pack. Jeff sees this as another in a stream of articles telling the world that coarse-grained services are the only way to go. He wants us to acknowledge that fine-grained services are okay when latency is not an issue, specifically for local systems (i.e., within a box, cluster, or LAN) or even when widely distributed systems are linked in such a manner that latency isn't an issue. He points out recent Microsoft WSE and IBM WebSphere enhancements that optimize local service calls by bypassing sockets (the transport, essentially) and even XML.
Jeff makes some good points, but not to the extent that I think Medicke and Pack are wrong. First of all, can we really use the label "web services" for local interfaces that neither use XML nor talk over a wire? I think not. They're degenerate cases. Second, Jeff refers to latency and usability as the criteria for selecting the coarseness of an interface. I accept that, so long as "usability" includes maintainability and what Medicke and Pack refer to as an "outside in" approach. I agree with the IBMers that a well-designed SOA should, for example, minimize the need for client applications to help maintain the state of the service. And the best way to do that is to move entire documents rather than create and manage a session to exchange little bits of information.
I used fine-grained interfaces all the time--probably 100x more frequently than I use those that are coarse-grained. But I don't consider the former to be web services or part of an SOA.
Posted Monday, September 01, 2003 1:56:05 PM
|
|
|
|
|
|
