"Lily Y. Fu" wrote:
> Hi,
>
> I wonder what your opnion is on a good combination for
> serving servlets based on your experience?
Hi ,
This article maight be of interest to the list, Although I have no
experence of the products under test I will cirtainly be looking into iPlanet.
Mindcraft (and independant test lab) have released results of it's "SPICE"
benchmark
which compares Sun/Netscape Alliance's iPlanet Web Server 4.0 (formerly
Netscape Enterprise Server)
against a COMMERCIAL version of Apache namely C2Net's Stronghold 2.4.2.
The iPlanet server was seen to easily outperform the Apache server in
static and dynamic content delivery. The emphasis of the SPICE tests were
on the server-side programming interface.
The testing was requested by Sun but C2Net were consulted throughtout to
ensure
fairness.
The news article follows. For the whitepaper published which includes full
details including the platform used follow the link in the article.
October 12, 1999, 04:58 AM
New Benchmark of Dynamic Content Shows iPlanet Web Server, Enterprise Edition
4.0 Outperforms Apache-Based Stronghold 2.4.2
ScreamingMedia, PR NEWSWIRE SAN FRANCISCO LOS GATOS, Calif., Oct. 12
/PRNewswire/ -- Today, Mindcraft released the first results produced from its
new Server-side Programming Interface Call Efficiency (SPICE) benchmark. This
comparison shows that the Sun/Netscape Alliance iPlanet Web Server, Enterprise
Edition 4.0 (formerly the Netscape Enterprise Server) outperformed C2Net's
Stronghold 2.4.2 (a commercial version of the Apache Web server) delivering
static and dynamic content from 20% to more than 200% faster.
The full comparison study, which includes a complete description of the
SPICE benchmark
and the programs used to generate dynamic content, is on Mindcraft's Web site
at:-
http://www.mindcraft.com/whitepapers/iws/iwsee4-sh242.html .
Mindcraft developed the SPICE benchmark:
1. To help users choose the most efficient way of generating dynamic content
on
a Web server.
2. To facilitate a fair comparison of different Web servers when using dynamic
content generation.
"Two significant problems with other Web server benchmarks using dynamic
content
are that they hide the performance of the programming interface used to create
the dynamic content and that they do not give you an idea of the upper bound
of
the dynamic performance you can expect. The SPICE benchmark solves these
problems," claimed Bruce Weiner, Mindcraft's President.
For the iPlanet Web Server, the SPICE benchmark tests show:
* Its Java servlet interface is 2.2 times faster than its CGI interface.
* Its Java Server Pages interface, which allows a developer to add Java code
directly in an HTML file, is 1.9 times faster than its
CGI interface.
These benchmark results make it clear that Java servlets or Java Server Pages
are much more efficient programming interfaces than CGI on the iPlanet Web
Server.
For Stronghold, the study shows its CGI interface is 1.03 times faster than
its
Java servlet interface.
Mindcraft's comparison shows that the iPlanet Web Server provides more static
and dynamic content per second under all load conditions than Stronghold.
Also,
the comparison shows that the iPlanet Web Server delivers a complete response
to
client requests faster than Stronghold.
"The SPICE benchmark is based on a highly portable specification and is easy
to
implement using almost any Web server benchmarking tool. It's clearly a case
of
less is more," Weiner stated.
Doing a fair Web server performance comparison for a mix of static and dynamic
content without developing or porting either a pseudo- or a real application
is
easy with the SPICE benchmark. The key is using the simplest possible
applications that return the same number of bytes as an average static
request.
This constraint isolates any performance difference between static and dynamic
content requests to one of three areas: the overhead of the Web server
starting
the program that produces the dynamic content, the overhead of the program
itself, and the overhead of the Web server stopping the program. In addition,
the simplicity constraint yields performance measurements that show the upper
bound of real application performance.
Mindcraft certified the accuracy of the results reported in the comparison
study. The Sun/Netscape Alliance commissioned the study. Mindcraft consulted
C2Net, the makers of Stronghold, on several occasions to verify Mindcraft had
achieved the highest possible performance for Stronghold and to resolve a
technical issue with their SSL implementation.
About Mindcraft
Mindcraft is a service-oriented, independent test lab. The company was
founded
in 1985 to provide high quality services and products to vendors and end users
who want to test software, system, and network products. Mindcraft is
committed
to work to promote standards in our industry. Bruce Weiner is a member of the
Standard Performance Evaluation Corporation (SPEC) Board of Directors.
NOTE: Mindcraft is a registered trademark of Mindcraft, Inc. in the United
States. Other product and company names herein may be trademarks of their
respective owners.
SOURCE Mindcraft, Inc.
CONTACT: Bruce Weiner of Mindcraft, Inc., 408-364-2866, or
[EMAIL PROTECTED]
Web site: http://www.mindcraft.com
Copyright 1998-99 PRNewswire. All rights reserved. Republication or
redistribution of PRNewswire content is expressly prohibited without prior
written consent of PRNewswire. PRNewswire shall not be liable for any errors
or
delays in the content, or for any actions taken in reliance thereon.
Karl
===========================================================================
To unsubscribe: mailto [EMAIL PROTECTED] with body: "signoff JSP-INTEREST".
FAQs on JSP can be found at:
http://java.sun.com/products/jsp/faq.html
http://www.esperanto.org.nz/jsp/jspfaq.html
