RE: [SEMI-OT] Oracle to cloud or not to cloud?

From: Goulet, Richard <>
Date: Tue, 15 Dec 2009 14:50:05 -0500
Message-ID: <>


    Kindly explain what the difference is between buying a certain amount of compute power (I'll group that as CPU's, memory, and it's associated interfacing), an amount of disk space, and other resources, as you put it, in a "cloud" vs. actually purchasing hardware that I'm having put into someone's data center and hosted there?  

    The actual difference, that I've had to deal with, is that response time of our cloud is dependent on that of everyone else's clouds in the farm that we were sharing. I'm having a pile of trouble with these "virtual" environments because of the fact that we are now asking our app to compete with other applications and software layers upon layers to get the work done, and you as a customer are completely blind to what's going up outside of the cloud  

    Now I will admit that some vendor's virtualization software is better than others as well as some vendors software does run better on certain hardware than others, but the end result is the same, we're trying to make use of every cpu cycle that we can in getting actual work done instead of wasting them. The premise of this is that the laptop sitting in front of you is only consuming 5 to 10 percent of it's cycles working for you and the rest either handling background tasks or in the system idle task. And if your going to do so on the micro scale of a corporation then why not on the macro scale of the internet. That scales up rather well into the world of servers and was what virtualization was suppose to fix, scale that up to a "cloud" system and there are real economies of scale to be made, assuming that it all hangs together properly. And hence the "beauty" of cloud computing, you buy what you need and someone else has to purchase the 12 ton monster that hosts it, but makes a pile of cash because you can split the power as needed. Remember Oracle's statements about RAC, "run out of power, simply add another server into the cluster". In the cloud would you get the advantage that you can exceed your purchased CPU to maintain a service level. Wonderful, I get the big monster when I need it and only have to pay for the lesser amount.  

    Sounds absolutely great on paper and marketing glossies. The reality I'm afraid does not match the promise. The same could also have been said about client-server computing that was so totally popular a couple of decades ago. Again the promise did not match reality do to the complexity of keeping every client's software in sync. The good part is that people keep thinking up these things and the underlying software, and sooner or later someone will unlock that holy grail that has been sought after for all of these years. Who and when I surely don't know and I don't expect it in my lifetime. Till then, not withstanding the statements and pronunciations of the Oak Table panel, I will remain a skeptic.  

Dick Goulet
Senior Oracle DBA/NA Team Lead
PAREXEL International  

From: Jeremiah Wilton [] Sent: Tuesday, December 15, 2009 1:48 PM To: Goulet, Richard
Cc: Oracle L
Subject: Re: [SEMI-OT] Oracle to cloud or not to cloud?

I really have to take exception with the multiple assertions on this list that "cloud computing" is just a marketing rebrand of something we already have.

Some companies have co-opted the term and are using it to refer to any and all manner of technology and even imaginary technology. But Google and Amazon, the leaders in cloud computing, are building something that greatly exceeds the previous technologies. Yes, you could host applications on the internet in the past, but Amazon's model allows instant provisioning of virtual servers, and virtual resources, like disk, memory and compute power, using a standardized web API. And this is all very cheap.

This is very different from calling your hosting provider and asking them to add a virtual for you or using Gmail online. It is not just hosted apps, and it is not the same thing Oracle has provided for a decade. I understand how the whole thing reeks of marketing nonsense, but suggest those feel it is comprised of no actual new technology take some time to educate themselves.

(Plug for Oak Table book...) I wrote a chapter that quotes Larry Ellison in the video liked below for the forthcoming Oak Table book:


Jeremiah Wilton
Blue Gecko, Inc.

On Dec 15, 2009, at 8:57 AM, Goulet, Richard wrote:

        "Cloud Computing" is just another marketing pile of POO so that simple minded people, like your CIO/CEO, will think that their behind the times and need to spend more money on yet another pile of technology vaporware so they can add the buzz word to their résumé's and company advertising. Cloud computing = hosted apps, period. Been around for a long time making money for the likes of Capgemini and others who have hosting centers that are getting empty due to smaller hardware, lack of usage, etc..., namely the economic downturn. One of these days we'll probably learn how to distinguish that old Ford in the new paint job from the really new model. Though if marketers have their way, we won't.         

        If you can't impress them with brilliance, smother then in .         

	-----Original Message-----
	From: Rich Jesse
	Sent: Tuesday, December 15, 2009 10:40 AM
	Dateline September 21, 2009, Larry Ellison rants against cloud computing:

Received on Tue Dec 15 2009 - 13:50:05 CST

Original text of this message