Re: Microsoft destroys TPC-C records!

I bought DCOM / MTS hook, line and sinker. Then I found that even Microsoft has to eliminate the D and simply use COM in order to get MTS to scale with IIS (and running as an in-process Library Package, no less).

And to top it off, I recently learned that the July 11th, 1999 Web Application Server Shootout article in PC Week used a VC++ implementation relying on an ISAPI extension instead of MTS as the article insinuated. This DLL preallocated connections, memory buffers, etc. up front, and used the ODBC API with implicit commits - where are the transactions?

Based on my knowledge of the VB / MTS environment, I can unequivocally state that it is not possible for the VB test to perform at 2/3 the throughput of the VC++ / ODBC API test if the VB / MTS case were using transactions.

I have yet to see a case where MTS / COM+ transactions are measured head-to-head with a competitive product.

I'm not trying to bash Microsoft, as I make my living using its tools to develop applications. My irritation with the BS being published stems from implementing a large project for a Fortune 100 based on smoke-and-mirrors technology.

Microsoft needs to publish less marketing hype and more accurate technical content on its technologies. Also, releasing some of the issues known by the Microsoft staff as MSKB articles could keep clients from wasting money on fruitless technologies.

Mike

DNP wrote in message <38CD9303.22C5_at_btinternet.com>...
>If Com+ helps you keep the right data in memory at the right time then
>yes. There is no other significant issue.
>
>David P.
>
>
>----------------------------------------------------------------------
>
>Norris wrote:
>>
>> Do you mean the fast performance of Windows 2000/SQLServer 8.0 is not
 related to the speed of COM+ ?
>>
>>
Received on Sun Mar 19 2000 - 00:00:00 CST