May 5, 2005 3:11 PM PDT

Provocateur predicts 'end of corporate computing'

Related Stories

Barrett still has some fight in him

October 19, 2004

Tech bigwigs dispute guru's pessimism

November 17, 2003
Nicholas Carr, the former Harvard Business Review editor who agitated the information technology industry with his article "IT Doesn't Matter," has published a sequel that predicts another, even more disruptive change.

"The history of the commercial application of IT has been characterized by astounding leaps, but nothing that has come before--not even the introduction of the personal computer or the opening of the Internet--will match the upheaval that lies just over the horizon," Carr predicts in a summary of his next work, "The End of Corporate Computing." The article appears in the spring 2005 issue of the MIT Sloan Management Review.

Carr's previous work made the case not that computing technology was unimportant, but that it's no longer a route for one company to gain competitive advantages over others. Carr riled many in the computing industry; Intel Chief Executive Craig Barrett was among those to deride the position.

This time around, Carr argues most companies will stop messing with information technology altogether, instead tapping into the resources of gigantic centralized computing utilities.

"Information technology is undergoing an inexorable shift from being an asset that companies own--in the form of computers, software and myriad related components--to being a service that they purchase from utility providers," Carr argues. "IT's shift from an in-house capital asset to a centralized utility service will overturn strategic and operating assumptions, alter industrial economics, upset markets and pose daunting challenges to every user and vendor."

Carr's latest position jibes better with prevailing computing industry thinking.

Many computing companies are embracing the idea of utility computing in varying degrees. In particular, Sun Microsystems rents out the use of its own grid of computers for calculation tasks; in the future, Sun expects chiefly to supply plumbing to business partners that actually sell the service to the ultimate customers.

Sun Chief Executive Scott McNealy said the shift is slow in coming, though.

"They don't seem to have any problem buying electricity on that basis, but when it comes to computers, they freak," McNealy said this week at a product launch. "It's more of an anthropological issue than a technological or business model issue."


Join the conversation!
Add your comment
Did Scott McNealy really say that? I find it hard to believe that he would openly imply that getting computing done on outsourced hardware is not unlike getting electricity from a power company. I mean, when's the last time anyone had to worry about someone breaking into a power company and stealing the sensitive data of their utility customers?
Posted by (1 comment )
Reply Link Flag
Yes, it has been Sun's strategy for several years now. They
believe, and I suspect they are right, that people will be
purchasing computing power instead of computers. It makes a
great deal of sense. McNeally calls it the "great webtone switch."

You say that it can't be done because of security, you are wrong
for two reasons: 1) Companies can't secure the resources they
have inside their own networks today. 2) Decentralized services
make much more sense for security. How can a hacker take
down your computer resources when they are distributed all
over the country, or even the planet? Built in redundancy,
hardened networks, etc. This IS the future I think.
Posted by (38 comments )
Link Flag
Sounds just like....
... what MS had in mind for the personal computer. All the apps on
big servers, all your data on other big servers, and what you do is
rent access time via the internet ia your Passport. And MS would
guarantee the security of the system, .... and of your data, ..... and
there would be no peeking .....

yeah, sure, and pigs can fly!

Thank goodness this MS conceept didn't fly.
Posted by Earl Benser (4310 comments )
Link Flag
Spot On
Carr is dead-on. It's not about "computing" anything these days. Like "desktop," this is an outdated 80's metaphor. Sure, computing is going to be there, as part of the substrate, which is why the utility computing movement will continue to grow: just pay for cycles.

The notion of the computer as computing device has been obsoleted by the Internet. All of the real action these days is in using the computer as a *communications* device.
Posted by tparisi (10 comments )
Reply Link Flag
Yes everyone is backwards
Everyone's business notions are WRONG, in fact as you say people are stuck in a frame of mind that is a quarter of a century old. It's Mr. Carr that's right and everyone else is wrong. Mind you Mr. Carr doesn't manage an actual IT business, but he does make money selling his book in which he claims to have great insight into this field.

Oh wait, that can't be since companies that are in any sort of business flourish. "Utility computing" companies manage to suck. Have you heard of Citrix lately?

Carr's position doesn't jibe with "traditional thinking" it jibes with "thinking." Utility computing will continue to exist, but only in instances where it makes sense, such as the uses it is put to currently - mainly large scale applications.

If something is so great companies would adopt it - they don't care about "notions" of ownership versus leasing a service. All they care about is profit and if utility computing provided that companies would be going into it en masse.

I wonder who owns Carr's computer at his office/provocateur den.
Posted by sanenazok (3449 comments )
Link Flag
those that can, do...those that can't, teach...
Posted by tlite722 (160 comments )
Reply Link Flag
Ms. Adbul helped him write this article...
Yeah, he has tapes of her leaving messages on his answering machine:

Ms. Abdul: "Write a book where you say that things will change dramatically sometime in the future. You picked on IT infastructure before, do more of that!"
Posted by sanenazok (3449 comments )
Link Flag
that's the best response to a provocateur
Posted by sanenazok (3449 comments )
Reply Link Flag
Users, and more importantly companies - will not trust computing that they down control and own. PERIOD. This myth that you just purchase and rent apps and run it all virtually over the network - I don't believe it and/or see it happening and I work for a very large multistate healthcare organization. We tried moving all of our servers to a central location and found out there were a few apps that didn't LIKE being run across a 400 mile distance. Couple this with the fact that PCs 3Ghz PCs are relatively cheap and..... it's a no brainer to keep - and use - that kind of distributed computing. A third party vendor does not have as much interest in your business as an in house IT shop.
Posted by Bob Nixnorf (2 comments )
Reply Link Flag
ADDENUM from Bob Bodega
P.S. We have and run Citrix at my work. It's great for it's niche and what it does. It took years, of research, experimenting, vendor support, heavy back end hardware and * MONEY * to get it working rock solid right. Yes..we have Citrix but WE own the servers, network and integrate everything about our infrastructure ourselves. We find it a heck of a lot more reliable that way. When fingers need to point, we KNOW who can and WILL fix things.
Posted by Bob Nixnorf (2 comments )
Link Flag
This guy knows nothing
Has this individual ever worked within an IT department or for that matter a large corporation. Probably not, he sits in his ivory tower spouting ideas that he believes are "cool" if you will.

I have been working in technology consulting for nearly 10 years now. Most of my clients are the CIO/CTO/CEO's of fortune 100 companies and I can say unequivocally that this "prediction" is so far off that it is laughable.

I have worked with several large clients that were interested in the "utility" computing. But the security requirements (we all know how sensative companies are about security, how do you expect this to work in a shared environment)and need for control of the IT apperatus that not a single client utilized utility computing.

It just doesn't work, plane and simple. The concept sounds great and in the world of academia may prove to be a good idea, but in practice it falls flat on its face.

The simple reason is, IT is complex enough when you have control of the systems, even more so when someone else has control of them. I have never met a client that could utilize something "out of the box" with out needing to change it. You really don't have that much flexibility with utility computing. Systems and business processes are just far to complex (though many times they don't need to be) to work in a utility environment.

This is just another case of Analyst and Academics not knowing what they speak of....I would like to know just how many years Mr. Carr has worked in a large corporation or delt with an IT organization. I can only assume that his "real world" experience is limited by way of the theories he puts forward.


Philip Grossman
Senior Information Architect
Posted by grossph (172 comments )
Reply Link Flag
He is right! Commentors have no vision.
You talk of security worries? Yet financial orgs OFFSHORE processing your accounts. Shared environments? Large corps use security systems all time for shared computing! Just move it up a notch. It's all about MONEY, NOT IT. It will take a few more years to perfect, but it will happen. The big companys see it, why can't you!
Posted by gfsdfge (130 comments )
Reply Link Flag
isn't it...
approximately the concept behind Application Service Providers? OK this is taken a bit further than that, but I doubt it will become as global as predicted. It might become like that to some extent, and is moving in that direction already, I don't think though every company is ready to give up on its in-house IT infrastructure.
Posted by Metasoft (3 comments )
Link Flag
It is not about vision
Oracle tries this with "give us your informaation and will store it for you" for some time and I don't see any result. If there will be some centralized something, it will be owned by the company, not rented from some firm. I'll wait and see.
Posted by orfeu_niko (104 comments )
Link Flag
Trend stretched to it's ridiculous edge
As noted in the comments above "utility computing" is just another name for "outsourcing". If you add in the common utility applications for day to day operations that so many companies use (Peoplesoft, Oracle, SAP) and outsource the hardware and support for them to another company then you've definitely using "utility compnuting". However, this scenario falls down because once a company has divested itself of utility needs it can then focus it's remaining IT resources on the competitive aspects where IT can make a difference.
Posted by aabcdefghij987654321 (1721 comments )
Reply Link Flag
Re: The only logic is open source computing
First there was mainframe computing, then client server computing, then internet computing and now, utility computing....meanwhile we are all arguing about the method of computing, comes along the popularity of open source. And the war of cultures began...between open source and proprietary systems, embody in the form of Windows/Unix vs Linux.

So what does all this means? Well, I agreed with some of the readers here, it is all about money and control. Businesses are only interested in saving money in deploying IT and control over their data. In the end, more and more open source technology will be part of corporate IT; buying commodity hardware and software to run their business, bottom line: saves money, in house control and in customization without paying license royalties.

My humble opinion, open source is the future because it address the key issues of businesses; low cost, free to customize and in house control. Utility computing will work to some extend but it will not be like electricity. Sorry. I think Nicholas Carr maybe right in some things but not this time.
Posted by penguinista3 (8 comments )
Link Flag
Who really benefits from Utility Computing
The answer is that companies who can't seem to adapt to a new business model. Cable companies understand monthly billing for hardware adds up over a year. How do they bill Cable internet service? With a flat fee for content, and by leasing you the cable modem. Do telephone/DSL providers do this? No, because it's not part of their normal business model.

I'd suspect that IBM's old mainframe philosophy is behind this drive to Utility computing. It's very easy to bill by the month and provide premium services by the hour. Anyone remember mainframes that let you speed up the CPU to meet demand (for an extra fee, of course).

The market is not driven by technology, hype, or the Next Big Thing. It is driven by customer need. If technology meets a need, then demand will appear and the technology can flourish. I'm not aware of any pressing need that can only be met by Utility Computing.
Posted by (7 comments )
Reply Link Flag
Who really benefits from Utility Computing
The answer is, those companies who can't seem to adapt to a new business model. Cable companies understand monthly billing for hardware adds up over a year. How do they bill Cable internet service? With a flat fee for content, and by leasing you the cable modem. Do telephone/DSL providers do this? No, because it's not part of their normal business model.

I'd suspect that IBM's old mainframe philosophy is behind this drive to Utility computing. It's very easy to bill by the month and provide premium services by the hour. Anyone remember mainframes that let you speed up the CPU to meet demand (for an extra fee, of course).

The market is not driven by technology, hype, or the Next Big Thing. It is driven by customer need. If technology meets a need, then demand will appear and the technology can flourish. I'm not aware of any pressing need that can only be met by Utility Computing.
Posted by (7 comments )
Reply Link Flag
The need is to cut costs!
Outsourcing or utility computing reduces costs. Real and perceived. IT staff reductions are real. Perceived hardware costs will also be reduced. Remember how hard it was to put a 486-dx2 onto a novel network? Needed a giant NIC card, jumpers, software, etc&
Now its a breeze! Give it a couple of years and so will utility computing. If they can move whole business processes overseas, why cant they get rid of that annoying computer room?
Posted by gfsdfge (130 comments )
Link Flag
The Real Deal!
You can ignore what Mr. Carr writes because it's obvious he has never worked one day in the IT industry. He is a policy-wonk and corporate executive shill. But, the 'utility computing' concept needs further explanation; it only applies to computing power (hardware) and NOT software. It is possible to 'rent' computing power ONLY if a corporation has rigid security and well-defined requirements, but the success stories will be limited since most corporations have very little control and instead only emphasize short-term costs. But, it is impossible to commoditize software because, as those of us who work in IT know, ALL SOFTWARE IS CUSTOM! Otherwise, there would not be the need for all these software support costs if it was such a commodity. Every corporation wants something different in their purchased software, so a utility software model will never work. Until clueless corporate executives 'wake up' and realize that software CANNOT be commoditized and will always be a critical part of their operations and strategy, we will be stuck with bean-counters and so-called IT consultants who can't see past the short-term noses on their face...
Posted by lazura (23 comments )
Reply Link Flag
You are Totally wrong. Goto
They are doing it quite well. Others will follow. You will be off shored. If you are half as smart as you think you are, you'll start looking for a new career because Inofrmation Technology WILL be assimilated. Business people will NOT be told what to do by a bunch of geeks.
Posted by gfsdfge (130 comments )
Link Flag
what UC should be
Utility computing (UC) needs to do more than outsourcing the desktop via a Cytrix-like technology as has been proposed by Larry Ellison (NetPC) and others until now. The terminal-server model does not scale well for resource-intensive GUI's, it's restrictive if certain users have specific needs and tends to become more expensive than simply buying a bunch of commodity boxes.

The main point of UC is to minimize support and licensing costs, but doing this right will require resource consumption to be dynamically shifted from the server to the client or the other way, so it shall be independent of hardware capabilities, but use them effectively, when available.

A feasible UC will need to mix thin and thick client models, taking the strong points of both.
The UC-ready desktop shall be the same powerful box under your desk, but (pre)configured in a complete different way. For example:

- Application/file server mixture. All software (OS, Office appl, in-house appl, etc) should be installed, configured and runned from a central server by default, but it should exist the option to run any appl from the desktop PC instead, if hardware/user-rights meets certain requirements (e.g. good for graphic intensive appl).

- Remote applications should be able to get dynamically cached in their most used pieces of appl & data into the local PC.

- Users should be able to always control their own data. An (in-sync) mirror of all user's data and configuration options shall exist in a hot-swappable drive.

- The server shall be a preconfigured piece of commodity with default software installations and able to support only a determined number of users. More users means more server boxes. If a client or server box goes down, the support guy only needs to add a preconfigured new one; the user-specific data is all retrieved from the user's hot-swappable drive.

- A finer-grained model of master server and secondary servers could be deployed; secondary servers shall be a kind of in-house caching proxies commanded by the master server located offshore (even overseas).

A feasible UC solution will come as a software *and* a hardware deal.
Posted by alx359 (40 comments )
Reply Link Flag
Cytrix is GARBAGE
Cytrix is an overpriced and unstable platform. Have you used it ever? I set it up at a public area in a library. These were effectively dumb web surfing terminals. They sucked! This was in 1998-1999. The costs were incredible: server side and for the low quality clients - which FROZE for crying outloud. These things didn't support sound and more than 256 colors. This is in 1999, not 1989! They were replaced three years later with low cost P4 desktops that standardized on the same "image." These desktops cost less and were about 100 times more capable.

The library managed to unload the license (several thousand dollars) on some poor sod.
Posted by sanenazok (3449 comments )
Link Flag
So if every application every company uses for its particular tasks is the same.
Then how can we continue to innovate ourselves?

This just doesnt make any sense. Innovation is still alive and well in the IT/Business world.

Every company would need to be exactly the same to use the same solutions from top to bottom. There are no 2 companies exactly the same. Also, IT is way 2 embedded in business operations.

You telling me IT infrastructure someday is going to be like moving your TV to another room and plugging it in to work the same way is just ludicrous.

Or completely stupid.
Posted by pcpimpster (31 comments )
Reply Link Flag
INNOVATE, I should learn to spell
One more note:

So if this could be true an you can take something as complicated as a company's IT infrastructure (even a small company) and create a conformed utility service from it, then why cant you do it for the entire judicial system, or all the different police forces around the country?

I believe computing power someday may be a utility by way of the grid, but to take it one step further and say applications and business processes have to file in line, is again just not knowing what youre talking about.

So i guess things "are" going to be like that movie 1984.
Posted by pcpimpster (31 comments )
Link Flag
Think of it, you not only pay a premium for the hardware you run, but a premium for the OS per processor, staff to maintain it, staff to design to fit your investment in a particular platform, increasing storage needs, applications, shifting needs in computer processing power, remote access, backup solutions, virus scanning.

To create an analogy, look at cable not electricity. If it were set up more tiered than it already is, it would be the PERFECT analogy.

All you need is the set top box, which the cable company is more than happy to lease to you for a minimal amount (thin client). Instead of paying for packages you get an a la carte list of each offering separately.

So, you break it down to the broadband connection, processing power, storage, 3rd party or in house developed applications, and backup/archival systems each with levels of service. In fact, with open standards, they could be offered by different vendors.

In each scenario, all upgrades are passed on to you seemlessly. In a few situations your "set top box" (thin client) may need to be upgraded as well, but typically for free and at the same rate you're paying. When new technology comes along, the current offerings should move down the tiers and become cheaper, and the providers could allow you to choose to upgrade to the price point you're used to paying for better technology.

All in the same you use cable right now.
Posted by jamie.p.walsh (288 comments )
Reply Link Flag

Join the conversation

Add your comment

The posting of advertisements, profanity, or personal attacks is prohibited. Click here to review our Terms of Use.

What's Hot



RSS Feeds

Add headlines from CNET News to your homepage or feedreader.