This site focuses on architectural, sizing, and design issues around enterprise web projects.
 

I need a list of your sales/deployment successes/failures with the Sizing Guide. Let me know how it helped you or hurt you :-(

I will send a special gift for the first 10 contributions that says
 

Mine's Bigger, Faster, and Stays Up All Night
http://sizing.org       "How does yours measure up? "
e  n  t  e  r  p  r  i  s  e     a  r  c  h  i  t  e  c  t  u  r  a  l     r  e  s  o  u  r  c  e  s

"Ernie's Infamous Sizing Guide"

Also, I am not much of an HTML designer. Any volunteers for the layout and look and feel should e-mail me at erniep@netscape.com.

The hope of this site is to remove the FUD from the facts and give you fairly accurate information about a few important things:
 

• Hardware/Software Sizing - As the name of this site implies, the primary focus of this site is to deliver aggregated, accurate information regarding sizing of enterprise applications designed to run on web servers and applications servers. Additional information will be provided for peripheral services like messaging, calendaring, LDAP, proxy, portal, and anything else that comes up.
• Architecture - Proper architectural information will be made available about each platform. Discussed will be platform limitations, fault tolerance issues, and design considerations.
• Test results - Actual test results performed by independent consultants either generated directly for this site, or generated as part of an engagement at a client will be made available. Each test will be complete with a test summary, the test profile, the code and related components used to generated the test, and an interpretation from either the contributor of the data of an independent consultant. This site hopes to validate assertions and assumptions with valid information collected and submitted in good faith by our peers. All submissions will be published with the contributor's name, e-mail and quick profile as provided by the contributor.

I will post all valid information as I get it, but I need help.
 

To validate my results, I want third party core architectural information/knowledge about the following:

iWS - running serlets, jsp's
Apache - same
IIS - running proprietary stuff
Etc . . .
 

For bigger stuff .  .  .
Enhydra - Free, somewhat fault tolerant, and soon to be J2EE, wireless support, portal and personalization . . .
WebSphere - All versions (standard, enterprise)
BEA Weblogic - As above
iAS - usual stuff, fault tolerant, load balancing, scalable with limitations, etc.

Additional platforms
Portal

Finally, I want to outline architectural limitations, and how to build fault tolerance into each product set.
 

What other products should I include on the list?
 

I take into account number of VM's, threads, threads/cpu, fault tolerance, etc, with each configuration. Keep in mind that my methodology is very specific and uses my presumption that thread based sizing is accurate enough (within a few percent, since I don't take into account machine efficiencies, etc, at different load levels, and so on). My focus on sizing should also be a favorite among hardware vendors since I only size for presumed worst case scenarios, like absolute worst possible load, and overall system/network/request latency at that ridiculous load level.As a result, a properly sized app server or servlet execution based architecture will be idle for the majority of the time, except for those few moments in the day that the load goes very high, and during that time, properly sized, the network could still survive a failure of components, or entire machines. In other words, a trading site at market close, with the world on the site trading heavily, could survive through a machine failure in a multimachine iAS cluster without impacting their clients, and while still able to finish the day.

Please respond directly to me with any expertise in the listed products. I will send you directly an architecture questionaire.
 

Additionally, if anyone has been involved in controlled testing against a competitor, I'd like to publish the test summary, or at least discuss it.
 
 

Thanks all,
 
 
 

Ernie