Blogarithms
Doug Kaye's thoughts on web services, web hosting and managed services.
|
Web-Services Standards Status. Roger Cutler posted this status report of most of the web-services protocols on the W3C's Web-Services Architecture mailing list. In case you're curious where we stand as of 1 October 2003.
Posted Wednesday, October 01, 2003 9:51:05 PM 
|
|
|
The Metadata Debate. Lots of bloggers (e.g., Joi Ito and Tim Oren are discussing metadata this week. I went back to my archives and found my postings (1, 2, 3, 4) from two years ago. Bottom line: metadata doesn't work. I've been involved with far too man projects in which we tried to use user-supplied metadata as the basis for the organization of information. All ended in failure. And it's not just me! :-)
Posted Wednesday, October 01, 2003 7:44:37 PM
|
|
|
Baker: SOAs are more loosely coupled than REST? Mark didn't let me down. I was confident I'd hear from him regarding my Continuum of Services post.
Mark and I have had an ongoing discussion about the looseness of REST. I say it's somewhat tightly coupled--certainly more tightly coupled than SOAs. Mark believes the opposite. When last we left this debate, Mark referred to DNS as (one of) the mechanisms by which REST uses delayed binding to keep coupling loose. In response, I explain (perhaps not very clearly) how the very idea that information is tied to a single location--even if that location can change--is already a tightly coupled concept. In a loosely coupled SOA, information passes from one service location to another. The information is contained within documents, which may not have a permanent home. REST requires that information be located at one unambiguous location. The physical location may change, but the address (specified by URI) may not. I believe that's too inflexible and tightly coupled a model to support loosely coupled processes.
Posted Wednesday, October 01, 2003 3:25:02 PM
|
|
|
Web Server Performance Myths. Don Park wrote, "Here is a recent semi-public paper on web server performance mentioned in a message to Tomcat developer mailing list. Download article.zip and read the PDF file inside the zip file. It has some interesting discussion about web server performance myths. Here is an choice excerpt:
[...] yahoo gets 1.5 billion pageviews a day. [...]
Yahoo uses 4,500 server to serve up 1.5 billion pageviews each day. If we divide that by the number of seconds in a day, we get 17,361 pageviews per second. Assuming the load is distributed evenly across the servers, each server handles 3-4 pageviews per second per system.
"One of the key points the paper stresses is the performance/value offered by hardware XML accelerators for XML-happy web applications. There are other choice bits in the paper, so check it out before the authors take it offline." [Source: Don Park's Daily Habit]
Posted Wednesday, October 01, 2003 8:43:22 AM
|
|
|
RSS-DATA Jeremy Allaire has posted his thoughts for expanding the role of RSS into data-oriented applications. I particularly like this idea because of its support of unintended consequences. Just like Amazon.com's and Google's web-services interfaces, I can imagine publishing all sorts of data-based RSS feeds for others to take and run with.
However as folks deploy feeds of a similar nature, it will highlight the need for standardized semantic models. Right now, RSS works in part because of its relatively consistent semantics. An RSS-DATA spec leaves that issue open, but that's okay for now. Others will address application-specific semantics.
Posted Wednesday, October 01, 2003 8:36:28 AM
|
|
|
|
| October 2003 |
| Sun |
Mon |
Tue |
Wed |
Thu |
Fri |
Sat |
| |
|
|
1 |
2 |
3 |
4 |
| 5 |
6 |
7 |
8 |
9 |
10 |
11 |
| 12 |
13 |
14 |
15 |
16 |
17 |
18 |
| 19 |
20 |
21 |
22 |
23 |
24 |
25 |
| 26 |
27 |
28 |
29 |
30 |
31 |
|
| Sep Nov |
|
|
