McKinsey: Adopt the cloud, lose money The logic - as ever with crowd-based movements - is self-evident: cloud computing will save you money because you will no longer need to own your own servers. As ever with the wisdom of crowds, the truth isn't quite what it seems. Sure, you no longer need to buy or run those servers, but along with the cloud comes a whole new …
AC,
I think Gavin called it right that time, and it was no typo.
McKinsey have totally misinterpreted the Cloud/cloud computing for it has got absolutely nothing to do with saving money, but everything to do with making money. Although one cannot fault them for trying to retain/project/create a relevance and present a incredible profile to their existing customer base about such a Novel Future Environment which they do not understand, but which is rendering existing business models as parasitic and geriatric when up against the Dynamic Virility of Cyber Services IntelAIgently Designed for Vitality.
And although it would be nice to see Sun do well ...."Sun Microsystems in particular, now talking up cloud computing as a last-ditch effort for making money on open-source, should pay special attention, given that its chief executive is a McKinsey alumnus." ..... I'm not convinced that they realise how to Control the Cloud Crowd with their Computer is the Network Model, which is just a missing piece of algorithm in the Intellectual Portfolio of the chief executive, I suppose, otherwise they would not be hooking themselves to Big Blue for cheap thrills and a Fix.
Controls in the Cloud, on the other hand, would have Big Blue as the hooked Western Punter visiting the House of the Rising Sun Madame for all Manner of Exotic and Erotic Eastern Fare for Common Mutual Pleasures Mysteriously Administered from an Alien Phorm ........ for the Future is no longer about Search with Engines, for that Tech is Mature and Mainstream, IT is all about Sublime Placement of Honey Sought so that Dreams are Fulfilled and Reality Altered by IntelAIgent Design and Shared FundaMental Thought ........ and that simply requires that All Thoughts are Constructive and Benign/Sublime so that they cannot be rejected and in the case of Stealth items, detected.
Which you might like to consider is what Conficker is all about?
First - consider why SMBs use cloud : to save on initial cost and stay away from day to day IT operations.
Now think what large corporations are: typically they have many independent business units housed inside one big umbrella. They do share some stuff: like IT. So presumably that's one advantage that large corps don't have - as they anyway can't get rid of day to day IT operations.
Now consider a small unit within a large corporation - it's toying with a completely new idea but it won't get the funds it needs. What can it do to present a reduced cost proposal ? The manager of that proposal would use cloud instead.
And don't worry about integration with the rest of the system. In a large corp it's impossible for the left hand to know what right hand is doing anyway.
Paris, coz she might need both hands with some of us.
I only read this article and not the report it was based on. However, at least in this article, no mention was made of the type of software, or the nature of the workload (e.g. 2 servers needed for most of the year and 50 required just before Christmas).
One size does not fit all for software / customers. Factors change which programming languages, development teams, methodologies, etc are most suitable. Why would it suddenly be so simple for the deployment environment?
Cheers,
Adrian
Somewhere, someone at McKinsey has probably done a similar exercise and proved that cloudification makes money. It is just a matter of who is requesting/paying for the exercise what determines the result.
I'm tired of those consultants making those bold all encompassing statements. Some business will benefit from virtualizing their own datacenters, others will be better using a rented cloud. Business that need to rely on consultants to decide which option is better for them deserve what they get: consultants telling them what they want to hear.
And yes, this could be unfair to McKinsey as I'm not familiar with them specifically and they've advocated virtualization year on year since VMWare was born. But I'm familiar with other big names and all of them are glad of telling you wathever truth you want them to tell. Is McKinsey any different?
I would be expecting at least some level of research from the Reg crew related to past history from McKinsey on cost effectiveness of alternative strategies. And a Playmobil/Transformer reconstruction, please.
and my skepticism about the quality of Mckinsey's data analysis.
It's hard to believe that providing redundancy in the cloud costs more than providing it in your own physical hardware (if they'd said that the cloud created single points of failure which just can't be solved full stop I'd be much less surprised at their results).
In the end all cloud computing will mean is a regression back to the 70s where companies such as IBM would "rent" you a mainframe which you had to pay through the nose for. The 2 differences being that instead of communicating with said machine down an RS232 cable you'll be doing it down an ethernet cable and it won't be located on your property. Oh yes, cloud computing - the way forward!
Fscking NOT!
As most IT folk have been cautious about cloud computing, generally in my view and experience taking a balanced view that it can be of use, but its not the magic money saver that people think.
For some reason these days alot of IT decisions are made without any input from an area expert (an IT guy).... just MDs and Bean counters, who swallow the marketing rubbish hook line and sinker... like they did with outsourcing, the dot com boom v1, and the dot com boom v2, social networking having a business use and of course MS products being good value.
Folks,
clouds might seem hip today, but guess what? We all are using them daily! There are multiple kinds of clouds, internal, external, and there are multiple roles: Provider, user, etc.
Google uses a massive cloud internally FOR US, so that we can search the internet (this is one example of an internal cloud).
Amazon provides an external cloud for us to consume.
So, I guess, reducing the measurement of the usage of clouds to "dollars spend" or "dollars saved" for a single application, does not provide the right thinking.
An aspect, many do not see, when talking about clouds is: You do not need to spend TIME in the ordering and set-up process of the IT equipment, it's already there.
You do not waste CPU cycles, as with all big numbers, they do measure better.
So, if you quickly need an IT-piece, but only for a short time, then a publicly available cloud is your thing. So, no need to worry about the "removal" of the equipment, once the job is done. That will be used by someone else...
Yes, it might not be the right answer to all your problems, but same as with the power industry, we all at home do no longer produce our own power, we BUY it from the "grid" (to use an older term), and the producers of power can scale according to needs, and be more efficient overall. If you would only need power from 8:00 pm to 11:00 pm, then you might be better off, buying your own "power-plant" and running it yourself, but you need to remember, that "taking care" of that also adds to the overall equation. Many people tend to forget the "time" and "laziness" factor, which we pay premiums for, for NOT BEING forced to take care ourselves. Same applöies to insurances: You pay for the security, but you never know, if it really pays off (Ever done a ROI calculation of your health-care-insurance? I doubt, you like the numbers in many of your individual cases!).
So, as stated, you need to think of more then only shifting usage from 10% to 17% by doing virtualization, there's more to the overall equation. (and, did I mention, that many of the newly created public clouds will be located CLOSER to the power plants, so that CO2-emissions and LOSS of power during transportation can be reduced?)
Matthias
What you have to get your heads around is that the Cloud is not like anything else in Business and it has no defined Rules of Engagement,Only Fair and Generous Use. And if you Think of IT as the High Class Luxury Brothotel you always promised yourself to visit but never had the wherewithall or guts to enter because of fears of being recognised, or found wanting and a fraud, you will get an accurate enough idea about those who would be Creating Virtual Controlling Empires by Invisible Intangible Proxy in its Strata, for such Petty Concerns would have no Place in their Free Thinking.
But fear the Space not, for although IT will embrace the Fool and the Folly and accept its Money, IT will Suffer Neither and always Deliver ....... well, I believe the In Vogue Catch All Phrase of this Present Age, which says Nothing at All is just Whatever it Takes, although that will always be Mentored and Monitored to Ensure also, Whatever is Required for Sustained High Class Luxury Cloud Controls and Controllers.
And for all those that would doubt such a Space, it is as well to realise that the Pleasure you are Missing is all Yours, and your Participation is not needed and your Doubts are Unheeded.
* For Some Sterling Stalwarts is that One Major Vital and QuITe Addictive Attraction, the One that Controls All Possible Others.
Paris knows, you know.
I.e. "do this new stuff and you'll lose your shirt'... sharpens the attention, doesn't it? Amazon pipes up, McKinsey slaps 'em down.
McKinsey appears to pointing their witch-doctor bone universally at Saas, outsource *and* cloud, not just having a pop at Amazons view on cloud.
Horses for courses - it will make sense for some companies but not for others. I don't believe for a second that companies would dive into the cloud idea on the back of a a few figures (or am I wrong)... if you throw anything to an outsource (cloud, SaaS, whatever), without due diligence: where, who, what, how and when, you will definitely lose money.
Looking to leverage cost-reduction technology inhouse is a good first move rather than immediately throw the problem outside your organisation.
So, you liked the Cool-Aid so much you posted the same thing twice ?
Must taste real good then !
But tell me, don't you think Google is a bit of an exception as far as cloud computing is concerned ? I mean, is there any other company that has anything even close to what Google has as far as data-crunching is concerned ? I don't think so.
As for the rest of your arguments, you sidestep nicely the question of confidentiality. Banks are not going to use the cloud for anything and you know that. I'll wager that most companies will keep the mission-critical apps in-house as well because, you know, there's that liability thing that keeps cropping up now and then. And there's already enough trouble with loss of customer data as it is, no need to add cloud hacking to the mix.
In short, I really don't see how this new fad is going to go anywhere farther than it already is. Remote hosting is an old idea, and there are domains where it can be used advantageously - I believe those domains are already well-known and the services already available.
This cloud thingy strikes me as just another train for the Web 2.0 crowd to milk as far as they can, nothing more.
I imagine that the reason that clouds that are internal work so well for Google is the way that they use them...
say they have a hundred servers, during the day time you'd imagine peak web connections, and they'd serve almost exclusively as web servers.
during periods of lesser usage the web crawling process might receive a higher priority and thus they are indexing pages ready to be searched the next day...
And yes, cloud computing is perfect if you;re an on-line business like and you need 2 servers to receive your web traffic for most of the year, and yet 50 around Christmas,
but the real question is, would you save money? what is the cost of your package that would allow for 50 of the clouds computers to be used practically exclusively by yourself?
and is it cheaper to buy 50 servers and only turn 48 of them on in the run up to Christmas?
that way you have the same provisioning, yet you make cost savings by switching off when you don't actually need the servers rather than paying all year round for capacity that you may not need.
Also, how robust are these cloud services, it's all very well to say that you;re a small business who only need 2 servers for 48 weeks of the year, but how many others will need 50 at Christmas? are there hundreds or thousands of servers sitting on standby for 48 weeks in the run up to Christmas? do you see your service degraded when everyone else is seeing a peak in demand even if you're not?
It has to be one or the other, either the cloud companies DO have enough servers that they can serve all companies demands that are using them at Christmas,
or when it gets to that time of year, you find out your supplier oversold their capacity and everyone experiences the negative effects. (much like broadband suppliers!).
and if they haven't oversold their capacity, then surely they have too much capacity for 48 weeks of the year, and who pays for that? the customer of course.
so you have the choice I guess.
use a cloud, pay for 5 servers all year when you only need 2, then at the time when Christmas comes, you need 50 servers, but you still only pay for 5, but you don't really get the full capacity of 50 servers.
still cheaper to have your own equipment that you just leave off until you need it.
(we call that provisioning!)
@ boltar 16 Apr 1029 -
More & thicker fiber ropes (with separate media/strata flows?). More laser stations. Faster, fail-safe processing. Faster, bigger memory. Additional energy supply. Not possible to be delivered by a fingerclick. Maybe that's the only reason for the cloud economics of computations yet looking like, mostly, a theoretical model. What can be next door? oh, the qubit...
Also, the word "cloud" is so widely used in the UK top IT magazine that I'm already to believe they will make the thing from it great.
73
The reason the report concludes that Cloud costs more is because
Amazon are massively overcharging for their resources! We (Memset) are as little as HALF the price for comparable virtual machines and dedicated servers to the EC2 service.
S3 is even worse! Amazon charge $0.18/GB-month ) which is about £0.13 /GB-month
A 22TByte storage array from Memset costs £499/mo, which is £0.02
/GB-month.
McKinsey need to realise that all "Cloud Computing" is, is what we
dedicated hosts have been doing for years but on smaller time frames.
If you want a decent definition of Cloud, see the BCS Data Centre
Specialist Group Web site : http://dcsg.bcs.org/
Cloud/Utility computing of the likes of Amazon only makes sense for
people with dynamic load profiles. If you just need to outsource a
bunch of servers, then go to a normal IT host like Rackspace or Memset.
Kate.
Nice advert, Kate. How'd you squeak it past the censors^Wmods?
Regardless, the "cloud" business model is still not valid, even if your company is cheaper than everyone else trying to leverage the lack of computing savvy of the GreatUnwashed[tm].
I'll believe "Cloud computing" is valid when Google, Amazon, Yahoo & MS all use other people's data centers as a major portion of their corporate computer power. It won't happen, because it doesn't scale ... and history has shown us that things that don't scale fail, at least in computing.
I had to remove the links back to my site before they let it through (PS. Thanks El Reg mods ;)
The bottom line is that "Cloud Computing" is a rubbish term; what people really mean when they say it is "Buying IT resource on-demand from someone else's bigger data centre" in most cases ("utility computing" is a better term), and that busines model most certainly does work.
Companies like mine (Memset) are much more efficient at running a data centre and lots of servers than, say, a non-IT company's IT dept thanks to economies of scale and being specialists. Thus we can do it for half the price while still being profitable enough to a) grow fast without debt and b) pay me enough to have a ridiculously fast toys (I am a simply creature at heart ;)
The other view is that "Cloud" is just the new name for the same-old approx. 7 year cycle of IT insourcing/outsourcing.
Kate.
The biggest problem with cloud computing is that it seems that no two people have the same understanding of what the term means. The comments above bear this out.
Of course every vendor will claim that they offer the only true cloud computing solution and all their competitors are offering psuedo cloud computing. That almost goes without saying. However what's really interesting about the cloud computing buzz is that no two potential customers seem to agree on what constitutes cloud computing, hell you probably can't even find two people in the same company who agree about what constitutes cloud computing and what they hope to gain from it.
Oh and @amanfromMars.
How can you be so sure that "McKinsey have totally misinterpreted the Cloud/cloud computing" and that it isn't you that's doing the misiterpreting? If you believe that "it has got absolutely nothing to do with saving money, but everything to do with making money" then you have misinterpreted money. Saving money is for most organisations the best way of making money. You have revenue streams and they can be hard to influence, however your expenditure is much easier to control. Net result of reducing expenditure? Making money. How on earth can you possibly believe that saving money has nothing to do with making money?
"Companies like mine (kate's gratuitous advert elided) are much more efficient at running a data centre and lots of servers than, say, a non-IT company's IT dept thanks to economies of scale and being specialists. Thus we can do it for half the price while still being profitable enough"
Whatever. When I set up a non-IT company's system(s), they pay me once and rarely need me again outside of a short training period. One payment, no muss, no fuss. But hey, if you can separate fools from their money on a long-term monthly basis in this economic mess we're in, more power to you. Personally, I can't stomach it.
"to a) grow fast without debt"
Enjoy it while you can. I'll be picking up the pieces when your business model crashes & burns. Again. I've been doing it for a third of a century (Thanks, IBM!). People who forget history ...
"and b) pay me enough to have a ridiculously fast toys"
Well, I have a decent collection of '60s Fords and about a dozen motorcycles of varying vintage, a couple small boats & a couple 40+ foot boats. I also have a line of good USEA horses, three (fairly) decent race horses, and my wife's couple dozen hunters & jumpers, plus the 400+ acres to house them & the kit to maintain it ... I think I'm doing alright :-)
"(I am a simply creature at heart ;)"
I'm not. I do as little IT consulting as I possibly can to maintain my lifestyle.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
