Re: Slightly-OT: Throw HW at a SW/DB problem

From: Tim Gorman <>
Date: Mon, 27 Jun 2011 15:18:02 +0000
Message-ID: <W865646541139041309187882_at_webmail6>


The opening thesis of the tree falling in the forest and not making any noise is sound (pardon the pun) -- it is the basis for the triaged reactive tuning that many of us perform. There is always a threshold under which we ignore problems, and that threshold could be a system-wide value or it is complaints from users.

Hasn't this guy heard of Moore's Law? It's only been around for 40+ years, which may explain why he forgot it. I remember similar articles written in the late-1980s with the onset of multi-CPU systems, and later (to a lesser extent) with multi-core CPUs. Such advances have not reduced the need for tuning, but merely reduced the hardware footprint and/or increased user expectations. When has *anyone* ever overbought hardware capacity, as opposed to just saving money and buying less, or sharing that higher capacity amongst more workloads? I mean, seriously, c'mon... :-)

Efficiency and tuning never goes out of "style", and an inefficient application can crush any system, no matter how huge.

Just because someone got away with KIWI (kill it with iron) yet again does not make a seismic shift make.

Thanks Rich!


-----Original Message-----

From: Rich Jesse [] Sent: Monday, June 27, 2011 08:45 AM
Subject: Slightly-OT: Throw HW at a SW/DB problem

I think I'm gonna be sick: mentality **in a database professional** completely escapes me. Itbrings to mind that I just "upgraded" my Blackberry to one that has 2x theCPU and 5x the RAM -- and still works just as bad as the one it replaced (ifnot worse).Anyone from the Oracle camp advocating/predicting hardware-based "tuning"?Rant over.Rich--

-- Received on Mon Jun 27 2011 - 10:18:02 CDT

Original text of this message