I think you could call this the google effect....
Google is "hip", and they have a succesful businessmodel based on many many commodity servers, very scalable, based on their own linux version and technologies.
Because they are so successful, and "cool", many in IT seem to believe that this is the way to go. Just put some commodity servers and software together and bam there you have it, scalable, cheap (not), etc etc...
But this works well for google, not for 99,9% of companies. Google has one massive application, search, with relatively simple functionality. Google also does not want to be dependent on IBM, MS or whatever for obvious reasons. For them this model works very well. You can say it's their core business to manage this huge server park.
Bur for most companies, with ever evolving complex functionality, which needs to scale but not endlessly (like google's), this model does not work or is simply very expensive. Simply put an "i" in that room and it works. But this is not "cool". IT seems to be very driven by hype and mind-share. Google has lots of it and therefor, especially the young generation, seem to think that this model is best for all businesses. But of course it's not.
Subject: Food for Thought
Date: Mon, 17 Nov 2008 16:12:01 -0600
In the last five years, I have been adding on to my language skills,
learning the Java., C, C# and all the other web languages from my local
During one of our class "discussions" one of the younger set (under 25)
was discussing how you need a new server for every different thing you
want to do.
So for a standard web app that required data you had about 6 different
servers involved. The SQL server was different than the app server which
was different from the exchange server, so on and so forth. During
this long drawn, it became clear that in his mind it require at least 4
people to keep the thing running and would be available about 80% of the
time. I just sat there and laughed at him. Of course, (probably
rightly so) he became infuriated at me and asked me to prove him wrong.
I informed him that I had all those servers consolidated onto one server
(my System I) that never failed and had a consistent 99.95% up time. I
described how the system could manage all the functions at once and keep
them all running with merely one system and one support person.
Of course he tried to say the cost was the problem. I quickly pointed
out the error of his statement. Between the cost of licensing all those
smaller M$ servers, hardware costs, personnel costs, and lost business
costs, I proved the System I was a much better investment. (This was
before the Power systems existed.)
During this entire "discussion" the professor just sat there with a grin
on his face. He told me later that I was the first person he had met in
a long time that actually believed in the one-server-does-it-all
mentality. He was quite happy to here that there are some of us out
there. He constantly fights the smaller is better even if it needs more
What I have discovered is if you discuss the overall strategy without
mentioning the hardware platform until after the idea/concept is
understood, it works better. Then the discussion about platform is easy
for most business people to handle. Starting the conversation with "we
have a Power system for you!" does not work.
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
Jouw nieuws en entertainment, vind je op MSN.nl!