Our thanks to the readers who recently took the poll which completed the Register's workshop on "cloud computing". Since the term is a bit of a marketing catch-all, we parsed it into software as a service, utility computing and online platform provision. We further broke down SaaS into generic, vertical and core applications. We …
Some providers are better than others. If you host your stuff with a single-homed provider in one data centre, you deserve what you get. Ask them about it before you sign up, and ask how they can prove it.
None of the icons fit, so it's the Great Satan, ahead of the long awaited knifings on 15th Jan.
Cloud Computing, the IT version of letting the government "take care" of me and my data.
Send my valuable data somewhere I can't possibly verify to a bunch of people I'll never get to know to be stored on hardware I can never check?????
This is a nightmare, right?
Unreliable by nature, but kinda cheap. Perfect for data I don't care about, and applications I don't rely on. The other (even more unexpensive) solution is to channel all that to /dev/null. At least you know for sure. Using cloud computing for business? You are kidding, right? Free "cloud" e-mail have been available for decades, I don't see any large business relying on Google Mail or Yahoo Mail for core business. There's a very good reason for that. Cloud computing? Maybe, for home users with an overdevelopped hypster gene and no IT skill. That's it.
*aaS needs to go away as a meme. Soon.
We are wasting human-centuries on this drivel ... Consider these two paragraphs:
"Splitting the figures by company size, larger companies (over 250 employees) are already using the various flavours of SaaS and utility computing more than smaller organisations. But, to keep this in perspective, we're still only talking about a 32 per cent take-up of the most popular category - SaaS for generic applications - among these larger companies.
"The figures contradict what many insiders have been predicting, that cloud services are likely to be very attractive to the smaller organisation. To get from here to there, a lot of work needs to be done to remove the hype and to make the services and their benefits more understandable and accessible to these prospects."
In other words, larger companies are more likely to have management that is incapable of understanding IT issues, and are more likely to drink the "flash" that marketing types are trying to feed them.
This whole *AAS thing reminds me of the early 80s, when I was trying to beat networking concepts into the heads of kids at Berkeley, Stanford, Foothill, De Anza & Cañada. The PTB wanted me to teach arcnet, token ring, and the OSI Model ... I wanted to teach TCP/IP and Ethernet. It seems that whoever was in charge of decisions on teaching technology had taken a class, once. Or maybe heard about OSI and token ring somewhere ... It wasn't until the late '90s or early '00s that I stopped hearing so-called "engineers" babbling about the OSI model in meetings ... I did what I could within the guidelines, but what a waste of education time. Poor bastards.
Frankly, in this day and age, anyone who decides to hire a second party to host applications and store data needs to have their entire management group taken out behind the barn and severely thrashed. If the IT department signed off on it, the folks in charge of the IT department should be removed from the gene pool, preferably thru' castration without anesthetic.
If the same group of management and IT allows a THIRD party to make the decision for them (instead of making an informed choice of *aas provider for themselves, which would hopefully lead to a firm "no" on *aas from the Boss), the company should be dissolved and the idiots in charge should be shot without trial. They are that useless, in my mind. We don't need 'em.
*aas is a time and bandwidth wasting security hole, just waiting for lawsuits. It also costs a LOT more over the long-haul than simply adding infrastructure to an existing network.
Not that I have an opinion, mind.
... this sounds much like the old mainframe service bureau concept recycled.
is that the way businesses really want to go?
Much more succinct than mine.
Thank you :-)
Hands up anyone here who has a bank account. Anyone? How's about anyone who's happy to work for a company which uses a bank account, rather than storing all of its money as actual cash in a safe at Head Office.
Is that most of us? Really? So I guess we're all pretty comfortable about allowing specialists and experts in their field to look after our most valuable assets for us, rather than insisting on rolling our own solution every time. Benefiting from the economies of scale (easier for one business to look after the assets of millions than those millions to look after their own as effectively, sort of thing) and, frankly, the much improved security (because, again, the economies of scale make it easier to be properly secure once for millions than millions of times for one each.)
So next question: hands up if you'd reject the offer of working for (or with) a company or client for the sole reason that they didn't own their own office building, but rented an office, or a suite, or a floor, or a wing, or even a building from a managed services company (the prime example would be Regus, but I'm sure we can all name half-a-dozen more.) Anyone?
So it seems we're all reasonably happy about renting out services rather than owning and managing our own infrastructure from time to time, and again benefiting from the flexibility this offers ("You've taken on another hundred staff? You could take over the sixth floor, if you like...") and, again, the economy of scale that comes from using the same infrastructure to support many organisations.
I can't see that any of this is particularly controversial. Thing is, because I've framed it in terms of money and space, with analogies with which we've all been utterly familiar for centuries, it isn't particularly controversial. It's all very straightforward. Sensible, even.
So why do the very same concepts make Jake and John's blood boil when applied to something new*, like business software?
Well, would it be cheeky of me to suggest that, possibly, Jake and John might have some form of vested interest in the status quo? Actually, yes it probably would. I know neither of them personally or professionally, and would hate to be guilty of making assumptions based on little or no evidence. There must be a better reason.
Can it be security? I don't think so. It's pretty obvious that the major SaaS vendors take security very seriously. Follow me on this one: Look at the home page of a significant SaaS vendor. I'll be using Salesforce.com as an example, because they're the primary example with which I work**. Check out their case studies. Look at some of the clients that they have on their books. Now, if it were even a little bit easy for some hacker or industrial espion to break in and make off with their data, don't you think it would have happened by now? And if some hacker had managed it, don't you think we'd have read about it on the Reg? Every day for a month or so?
Nothing's 100% secure. Black holes excepted***. But I know I'd rather store my valuable money in a bank than in my own back pocket, and I'd rather entrust my valuable data to serious, experienced professionals than some of the bumbling (but good-natured) folks I've met running on-premise data centres down the years.
What about flexibility? Scalability? Actually, on both of those counts *aaS will win hands-down. Platform-as-a-Service will allow you to implement pretty much any form of business software in a fraction of the time required to develop the same functionality in-house and.or on-premise. It's the economies of scale again; most of the work's been done already, by some people who seriously know what they're about. And the major difference between a five-user system and a five-thousand-user system is four thousand, nine hundred and ninety-five user licenses. (Oh okay, and a more complex user profile heirarchy. Sure.) - no extra servers, or extra backups, or extra spare servers in case the first ones break, or...
And here's a scenario: You're opening a new office in Swindon. Or Sweden. Or South-east Asia. Where / how do you start setting up the infrastructure, data synchronisation, multilingual/multi-currency requirements for that with your on-premise solution? With SaaS, as long as they have an internet connection (and they will) then all you need is the extra licences for the new users. The rest has been done for you already.
Mobile users? No problem. Download the client. Log on. There you go. With all of your customisations, the relevant data, synchronised seamlessly whenever there's a mobile signal.
And there can be a significant advantage to putting power like this into the hands of the business user rather than their IT support team. The biggest advantage being the significantly reduced turnaround from business need to implementation. Where seriously useful, business-enhancing systems can be pushed out in weeks rather than months or years, there's more likelihood that they will reflect the current business processes (rather than the ones that were current when the system was first thought of) and that a ROI will be seen much sooner.
Okay, I realise that I'm probably a bit like a satan worshipper trying to preach to a choir here, but I figured a little balance wouldn't go amiss. It's not about undermining the IT department, or abdicating responsibility, or hiring "a second party to host applications and store data", but using the best tools available for the job. Frankly, if anyone needs re-educating round the back of the office bike sheds it's the die-hard stick-in-the-muds who'd rather mouth off about something and look good than spend a little time and effort seriously investigating a technology that could make a significant difference to their business's agility.
For those of you who are interested, a developer licence may be obtained for free from salesforce.com, and will allow you to play with the cloud to your heart's content. You'll also have access to any amount of help and documentation to assist you in your investigation. Be warned, your mind may be broadened by the experience.
For the rest, well, thanks for reading this far. Sorry you're unconvinced. It isn't for everyone, and there are certainly applications where "The Cloud" isn't the best solution, but they're pretty few and far between these days. You may be right. in your case. Might I take this opportunity to wish you a happy 1999.
*for a given value of "new" - as has been pointed out before, this concept's been around for a while. Only recently with this level of potential, though...
** Oh come on. Yes, I DO work in SaaS. I've drunk deeply of the cloud kool-aid, and with good reason. Keep reading.
*** No he doesn't. Hawking says black holes can evaporate over time, sure, but while they may lose their mass, any structure that falls in is lost. Nice try, though. If you really, really want to lose something forever, a black hole's still about your best bet.
Ah, but you miss my point. I (that's me, personally) have been weening companies away from second party solutions for around thirty years now. My blood isn't boiling. Far from it. The more companies buy into *aas today, the more work I'll have getting them out of that mess in my retirement years. I'm actually quite looking forward to it.
Regardless of what The Fine Folks in Marketing might try to convince you of, it's always cheaper and much more secure, to roll your own. At least over the long haul. Every time. For companies large and small. I've crunched the numbers, made the contracts, and then monitored the results in the real world.
Don't have the internal expertise to roll your own? I humbly submit that you have under-hired in the IT department. In a global corporation in this day and age, this is just asking for trouble.
In a smaller company with no overhead for a proper IT department, there are any number of independent consultants who can help (do your homework before hiring one ... there are a LOT of paper engineers with no street-smarts out there).
Global corporations don't "all of a sudden" hire a large percentage of new employees; such hires are carefully planned. So is their need for new office space, paper, pencils, toilet and rest facilities, parking and ::gasp:: computing power (who'd a thunk it?). At least in a proper company ... Note that I'm not talking about seasonal hires, who are typically hired for "dumb heavy work" and rarely need much in the way of computing power more complicated than a cash register/till (I'm not going to go into the waste of CPU cycles in today's POS systems).
In a smaller "mom'n'pop" operation, or even a couple dozen hardware stores owned by an individual, a modern PC (or Mac) has more than enough processing power and I/O to handle all computing needs. POS will be POS regardless; again, I'm not discussing that here.
Personally, I'm convinced that the folks pushing *aas are a combination of marketing people who are fighting to hide the fact that marketing is overhead, hated by everyone, and mostly useless, and IT folks who are in over their head and trying to hide it, mixed with otherwise useless middle management. Get rid of all these useless people, hire a decent IT staff, and do it in house. Save money. But that's just my opinion.
Agree to disagree
Fair play, I guess we disagree on this one. All I have to draw on is my own experience; years of designing and developing on-premise solutions which took ages to describe, let alone build, didn't scale easily, needed translating for each new market, took ages to re-design when the next new market turned out to do tax differently... The synchronisation issues, the backups, the integration, the failover planning...
I'm sorry if it sounds wooly and feeble, but doing it all on shared infrastructure that's someone else's headache is incredibly refreshing. Liberating, even. And it's such a joy to deliver in a couple of months a system which, in your previous life, would have taken quite literally orders of magnitude more man-hours to deliver, and which wouldn't have had a hope of lasting as well because they wouldn't scale as seamlessly, or be as flexible or universally available.
It may simply be the latest in a long line of solutions in search of a problem, and it's unlikely to become quite as ubiquitous as the laser, but I wouldn't be quite so quick to dismiss *aaS as "the old mainframe service bureau concept recycled" - there's a little more to it than that.
Since you specialise in rescuing woebegone victims from second-party-solution disasters I can see that you anticipate a beanfeast when the shine comes off the cloud's silver lining (metaphor stretched to breaking point yet?) but, seriously, I don't think I'd recommend holding my breath. SFDC's user count continues to grow apace, and their customer satisfaction rating is consistently high. People love their product! I can't tell you how refreshing it is to train end users and have them enthuse about their new system! Even the sales guys!
Again, I'd recommend signing up for a dev account and just testing the waters. What have you to lose but your preconceptions?
Jeez. I sound like a feckin' Scientologist, don't I? Xenu loves you...
Agree to disagree. Agreed :-)
"I'd recommend signing up for a dev account and just testing the waters."
Been there, done that. Even went "live" a couple times, with companies large and small. They wanted it, I needed the money and didn't argue (new water trucks and tractors are EXPENSIVE!). All of 'em had issues, and later payed me to bring it all back in-house.
But hey, if you can make a living at it in this economy, enjoy! Who am I to argue?
"Hands up anyone here who has a bank account. Anyone? ... Is that most of us? Really? So I guess we're all pretty comfortable about allowing specialists and experts in their field to look after our most valuable assets for us..."
Weren't we all the fools to let these banks look after money!!
Actually, the major thing I don't get about this whole Cloud/SaaS world is that you're relying on external internet links to provide transport to/from your mission critical services, but the tubes are getting squeezed and capacity is not as plentiful as even 2 years ago due to Web 2.0 bloat and streaming media, etc...
And yes, dedicated private lines will alleviate this problem, but then doesn't that remove one of the prime reasons of going SaaS - the ability to easily switch providers and to play the competition off between them? And yes, doing that will lead to the board going for the EDS cheap is cheerful route before realising that cheap actually means crap!