Re: "Transactions are bad, real bad" - discuss

From: Pablo Sanchez <pablo_at_dev.null>
Date: Thu, 08 May 2003 11:47:39 -0500
Message-ID: <Xns93756DCE5F72pingottpingottbah_at_216.166.71.233>


"Paul Vernon" <paul.vernon_at_ukk.ibmm.comm> wrote in news:b9dtck$481e$1_at_gazette.almaden.ibm.com:

>> Cost.  The more burden you place on the DBMS, the more hardware you're
>> going to need to run it.  The more hardware, the more power.  The more
>> power, the better the power conditioner, more AC, probably more
>> operation room floor space, etc.

>
> Fair enough. Not a very theoretical reason, but a reason nonetheless.

... which is why I was steering clear of the thread :) ... I realize that the NG is 'theory' ... although in the past, it hasn't stopped me from sprinkling some practical ... :) ... anyway, being that I'm kinda slammed, I haven't read your thread carefully so I decided to opt out.

> However, I might suggest that removing unnecessary layers such as
> 'application severs' would reduce the total amount of hardware
> needed.

FWIW, I try to balance the workload between the app server and the DBMS during development.

A simple example is I try to avoid using ORDER BY's if at all possible. I try to push that out to the middleware.

Or how about, how can a web application efficiently page across sets of data without caching the entire set? My biz partner came up with an elegant solution to solve this problem and when we get a chance, we'll slap a whitepaper and plop it on our website. The solution may be used across all RDBMS'

> I'm a believer in "More's law": For 99% of business applications, we

                     ^^^^^^^^^^ hah hah!

> probably had more than enough computing power a few years ago,

... then the bubble burst! :) I believe there are a lot of small companies that don't have the budget to spend on over cap'd machines. At least that's my experience - statistically, not significant.

Thx!

-- 
Pablo Sanchez, blueoak Database Engineering
http://www.blueoakdb.com
Received on Thu May 08 2003 - 18:47:39 CEST

Original text of this message