Re: Sorry, but...

From: ddf <>
Date: Fri, 6 Jan 2012 10:10:17 -0800 (PST)
Message-ID: <>

On Jan 5, 3:40 pm, TheBoss <> wrote:
> Noons <> wrote
> > On Jan 4, 3:21 pm, onedbguru <> wrote:
> >> You are kidding, right?!?!?!...   I would have to categorically
> >> disagree with you here... Maybe you haven't worked in a real shop
> >> before... :)
> > No, I haven't.  Please tell me what it is like?  Because in 40 years
> > of IT and 25 with Oracle, I certainly have never seen any of those
> > "real" shops and I need you to tell me what they are like....
> >> data center with as many targets as they have, SAVING 9.5M is a BIG
> >> deal in licensing cost and even personnel cost. They have more than
> > What the hell has saving 9.5M - unqualified - have to do with EM?
> > Do you know how to read English?
> >> Operations cost in a data center of that size is pretty amazing. 8000
> >> targets with > 2000 databases takes a LOT of human resources.
> > Unless they are a provider of db services, there is simply NO NEED
> > whatsoever for "8000 targets with 2000 databases".  And if they are,
> > then 9.5M is way too low for that kind of work at that dimension.
> >> windows server.  Have you paid for a HPUX or AIX license recently ?
> >> Multiplied by 8000?  And the hardware maintenance costs?.. 9.5M is
> >> just a drop in that bucket.
> > And pray tell in what deranged univers is EM going to help reduce the
> > number of Aix/HPUX licences?
> > Unless the number of databases and nodes is reduced, which ha NOTHING
> > to do with EM.
> > Can you stay on the subject for longer than a sentence?
> >> 9.5M doesn't seem so great now does it?   With savings like that,
> >> they can employ more people (sys admin's, DBA's, operations support,
> >> application support, power substation support) , buy more servers,
> >> incur more licensing costs, take on more customers etc etc etc.  So,
> >> I wouldn't be so hasty in your disparaging remarks there...
> > I stand by what I said.  Unlike you, who changed tack at every
> > sentence.
> >> I have also worked for a company that was in the top 10 corporate
> >> customers for Oracle licensing.  Believe me, data centers of this
> >> size, you will spend much more in licensing cost than you can
> >> imagine.
> > Once again: what the heck has that got to do with EM?
> > Wince when has EM helped anyone REDUCE the licenssing costs?
> > Are you on this universe yet?
> >> I would add that if you paid for EM or the tuning/diag pack, you are
> >> an idiot. It is something you cannot not install and it is not
> >> something that can be removed and it uses your resources whether or
> >> not you "pay" for it.  These tools are an integral part of the
> >> database engine that cannot be separated.  IMHO, Anything that cannot
> >> be removed should be included in your licensing. Period.
> > Like I said: get a clue, learn to speak and read English.
> > Then, and only then, comment.
> > <plonk>
> <tongue-in-cheek>
> As you said, it can't be a reduction in licensing costs, so if the story
> told by Cerner is true (remember this is Marketing...), the savings must
> come from somewhere else. Actually the very last sentence spoken by the
> Cerner string pupp.., ehm sorry: spokesperson, gives a clue what this
> could be. He said something like this:
> <q>
> "As our client base has continued to grow by 12% annually, we have been
> able to increase our number of databases managed by the DBA by 30%."
> </q>
> As the annual growth of the number of databases (12%) is less than the
> net average productivity growth of a DBA (30%), they probably have laid
> off a number of DBA's...
> Let's do some rough calculations:
> First let's assume Total Cost for a DBA is a quarter of a million USD
> and that for a DBA Team Manager is a third of a million USD.
> To get an overall cost reduction of 9.5M USD, we should lay off 34 DBA's
> (= 8.5M USD) and 3 Team Managers (= 1M USD).
> Now we can calculate DBA-productivity at Cerner expressed as the number
> of DB's managed per DBA (#DB) as follows:
> (1) no. of DBA's before using EM = 2000 / #DB
> (2) no. of DBA's after introduction of EM = (2000 * 1.12) / (#DB * 1.30)
> We've estimated before that 34 DBA's are laid off, so (1) - (2) = 34 or:
> (3) (2000 / #DB) - ((2000 * 1.12) / (#DB * 1.30)) = 34
> If you do the math correctly from here, you'll get at an initial
> DBA-productivity of a little over 8 databases per DBA (growing to about
> 10.5 after introduction of EM), which doesn't seem very high assuming
> fairly "average" databases...
> </tongue-in-cheek>
> --
> Jeroen- Hide quoted text -
> - Show quoted text -

The talking head from Cerner stated they 'reduced capital expenditures by $9.5 million ...' and capital expenditures do not include labor costs (according to the definition) so the presumption that they laid off DBAs and/or DBA managers is *probably* incorrect (unless their definition of capital expenditure differs from the accepted definition). So what could those capital expenditures be? Facilities? Equipment? Parts for the company Lear jet? The talking head doesn't go into detail but he was probably reciting a script from marketing made to sound convincing. We may never know how they saved such a bundle (or if they actually did).

For the record Cerner does what we do -- offer hosted data services (being a remote data center, as an example). As a remote data center servers, databases, switches, routers, etc. all reside in one of several locations owned by Cerner for the benefit of their clients. Each client gets its own set of databases/servers so it's very possible to have 8000 targets and 2000+ databases (spread out among prod, dev, test [or pre-prod] and qa), where some of those databases may be RAC or Data Guard configurations. We use OEM and it's quite useful in managing instances/databases for the data center. I don't know if we've saved trillions of dollars because we use it, but it is a fairly reliable tool.

My two cents.

David Fitzjarrell Received on Fri Jan 06 2012 - 12:10:17 CST

Original text of this message