My latest IEEE Internet Computing "Toward Integration" column, entitled "Web Services Notifications," which Werner recently alluded to in his blog entry on the same topic, has now been published. You can either get the PDF from the IC website, or get my version of the PDF from my home page.
This column compares, contrasts, and discusses WS-Addressing, WS-Events, and WS-Eventing. Despite the title of the column, it was written before WS-Notification was announced (maybe I'll cover that one in a future column). Overall, I think they're decent specs, but that shouldn't be too surprising given that we should be well past the stage of needing to reinvent the wheel in the events/notification space.
As always, comments and feedback welcome.
Comments (2)
Steve, great article.
We are currently in the process of developing a product which is very messaging - more specifically pub/sub - based. The entire product architecture is being built around web services, and a key competetive feature is ease of extensibility and the fact that we provide so many "hooks" into the guts of our product. Eventing is, we are finding, an effective way to accomplish this.
A stumbling block thus far, however, has been deciding how our interfaces and messages manifest on the wire. Obviously, referencing a widely adopted specification is more customer friendly than providing our own cooked up implementation complete with its own documentation. It's not obvious what the right choice is (nor if there even is a "right choice" at this point, which is the gist I get after reading your article... :) ).
I'm at the point right now, however, where I am thinking about implementing both WS-Events and WS-Eventing, and killing off the one which seems less likely to survive once we're ready to ship. The eventing subsystem is absolutely decoupled from the external interface, so this is certainly possible. WS-Eventing is much more elegant, but I just can't get past the lack of event discovery. (Sure, I can specify a nice XPath expression for filtering, but I need to know a priori the schema of the XML an event will generate! Not very pretty!) Aarghh...
I've not read the WS-Notification specification yet, but am heading off to do so right now. Maybe I have three specifications to implement? ;)
Posted by Joe Duffy | March 11, 2004 12:47 AM
Posted on March 11, 2004 00:47
Thanks for this comment, Joe.
Given that HP is a co-author of WS-Notification, I assume WS-Events will be going away as it's effectively been merged into WS-Notification. Also, note that there's a brand new OASIS technical committee for Web Services Notification co-chaired by an HP person and an IBM person, and my guess is that they'll use WS-Notification as a starting point but hope to attract the WS-Eventing crowd as well.
http://www.oasis-open.org/committees/workgroup.php?wg_abbrev=wsn
Thus, if you wait long enough, you might just get away with implementing only a single spec. ;-)
Posted by Steve Vinoski | March 11, 2004 9:09 AM
Posted on March 11, 2004 09:09