What he really said
When he talks about cloud taking off when facebook kiddies replace grizzled vets, this is code for ..' when people who don't understand how computers work and don't care about security and privacy replace people who do' .
Major adoption of public cloud computing services by large companies won't happen until the current crop of IT workers are replaced by kiddies who grew up with Facebook, Instagram, and other cloud-centric services – so says Rackspace CTO John Engates. Should we be worried? "10 to 15 years ago no one would put their credit card …
there is nothing like someone ells having all your data... no upfront costs... seems to good to be true..
i work for a rather big insurance company. a few weeks ago we had to kick an idea into the long grass for a quoting engine moving to a sharded cloud platform. it is currently running in house at no major cost on 20 blades. it need virtualizing to get more out of the tin but no major deal... we just said the words PCI and FSA and it went away... They just thought they could get it up and running for a bargin price.... idiots...
the cloud is one major hype sales pitch for hosting serives... i think after a few major blow ups and major data loss they will start selling us some new bullshit... :-D
I think it is generational. We are at the stage in IT that is similar to when farms all had their own windmills prior to rural electrification. Increasing complexity yields increasing specialization. Eventually firms won't want to deal with all the different specialties they will need. We'll pay for computing power in the same way we pay for electric power. It will be a utility.
You forget the "electricity" has no unique properties, my amperes do just the same as yours.
With IT it is the underlying problems of:
(1) Data protection. How many providers have client-side encryption by default so they (and foreign gov) can't spy on you?
(2) Bandwidth. If your task is computing-heavy, then you have a real problem with (1) (can't secure it remotely as it is on someone's VM so can be imaged while unencrypted) but not much with I/O bandwidth. However, if you work on a lot of data-heavy tasks then your ISP link will become a serious bottle neck and could run up a mighty bill.
(3) Redundancy. Oh sure, the likes of Amazon, MS, etc, have lots of hardware redundancy and similar, but also they may have a single point of failure (e.g. certificate recently in MS' case) and your ISP is likely to be the same. While your own building is also at risk, given the number of failure points for the ISP(s) between you and them it is much worse. And paying serious money for another provider is no guarantee it won't share the same fibre trunking, etc. (Actually, a local JCB mistake is more likely than large scale ISP outage).
Some may argue without an Internet like, your business is stuffed, but not always. Quite a lot of jobs (CAD, video editing, writing up reports, business planning, general HR duties, etc) can be done for a while with no external link at all.
(4) Lock-in. Once your balls are in the vice, they can turn the screws on you and what can you do? How do you migrate TB/PB of data from one cloud to another? Is that data really open, or locked in to some software they had you use? Have you actually tested such a migration to see if and how painful it can be?
Sadly the article is probably right, as a lot of "youf" seem to care little about privacy or the long-term implications of their choices.
Until all the "cloud" storage companies that over GBs of free storage go out of business for not having a sustainable business model.
I have too many GB of files to be able to store them on a cloud storage service.
I updated the photo gallery software on my website to better manage the photos I have archived there a couple of years ago and now have 16,082 photos on it and that is only the best 10%. I have taken 390,034 photos in the past 10 years, not all of them are worth putting online.
Experienced IT folks are trying to let newbies know that we've been there, and done that, and that outsourcing computing capability is a waste of time & resources, and is a major security headache.
Unfortunately, the Nintendo generation-bred iFad generation is completely clueless as to how the guts of the system actually works ...
My guess is that it'll end in tears, and probably be the source of the first proof that Humanity is a dead-end species from an evolution perspective.
Actually, I've given up. Three years to retirement and I can't be bothered to try to explain things to the newbies any more. They know that bandwidth is unlimited, that code optimisation is a waste of time, that frameworks are Where It's At, and that we old farts don't know anything at all. On the other hand, the solution to their problem
posted on $random_website can be dropped straight in even though it has no protection against anything from SQL injection to simple webpage corruption.
I've just tried to explain to someone that a clustered server on one physical location is nowhere near as good a guarantee of high availability as having a fallback server in a different country, and had back mumbles which translate, I think, as "The manual tells me how to do clustering but what you're suggesting looks complicated".
I'm in favour of migrating some things to the cloud. I also come up with things like more efficient ways of replicating essential data back to the ground (the Microsoft replication solution is all inclusive but inefficient in storage). Use the cloud but do not turn your brain off, is my advice.
But of course that assumes you have enough experience of what can go wrong that your brain is capable of processing the risks.
Don't forget that they can write all the software in Java, and like totally run it on all kinds of stuff dude. /sarcasm
Gone are the days of writing applications in assembly to optimize it, hell they aren't even writing crap in C, its all in Java. I'm looking right now at JavaW using 800mb of RAM, that's nuts, that is 8 times the RAM needed to run the whole XP OS just to run a program that if written in a lower language and optimized would probably be less than a 100mb.
Nowadays the attitude is: "The application is running slow. Have you tried throwing RAM at it?" What do we expect?
Is this not in a way similar with those (young) folks at Lehman Brother that found out you could sell any financial product and pass the risk to someone else ? How well did it go ? Who paid at the end? Aren't we insured for everything today anyway? Who pays for that when it goes ok? Who gets paid for that when it goes OK? Who pays for that when it all collapses?
It wasnt so long ago every man and his dog was offering network connectivity, now you probably get your connectivity from just a handful of providers. The same with desktop machines, now you probably get them from Dell, Fujitsu, or IBM.
Now its the turn of server infrastructure, much as you like to think your server you personally hand built will never die that isnt the case. You therefore would not be crazy to outsource this commodity to someone else, who has all the expertise inhouse (not in the head of some smartarse) and 24/7 support, every piece of spare hardware to hand. You couldnt hope to replicate this yourself unless you want to give up your life and sleep next to the server.
Time to move on guys, its not the 90s anymore
I don't see the interest of running mundane tasks like Bugzilla, TinyPM or Salesforce "in the cloud" once stupid problems of maintenance, upgrade and upkeep have been ironed out. And if we expect BYOD to be made viable, these WILL be ironed out. You will just need a fiber optic cable and a 1 dm³ ComputerCube made by some white-label asian company, probably sitting atop the microwave oven in the company kitchen, running the stuff of 20 employees. However big computation will remain TEH BIG....
It was only 8:15. The whole day loomed ahead, promising nothing but bills. With no contract work coming in for the past two months, Maria had written half a dozen pieces of consumer software -- mostly home-security upgrades, supposedly in high demand. So far, she'd sold none of them; a few thousand people had read the catalogue entries, but nobody had been persuaded to download. The prospect of embarking on another such project wasn't exactly electrifying -- but she had no real alternative. And once the recession was over and people started buying again, it would have been time well spent....
Maria put an end to her indecision in the usual way. She logged on to her Joint Supercomputer Network account -- paying a fifty-dollar fee for the privilege, which she now had to make worthwhile. She slipped on her force gloves and prodded an icon, a wireframe of a cube, on the terminal's flatscreen -- and the three-dimensional workspace in front of the screen came to life, borders outlined by a faint holographic grid. For a second, it felt like she'd plunged her hand into some kind of invisible vortex: magnetic fields gripped and twisted her glove, as start-up surges tugged at the coils in each joint at random -- until the electronics settled into equilibrium, and a message flashed up in the middle of the workspace: you may now put on your gloves.....
Maria reverted to the standard clock rate, and a macroscopic view of her twenty-one Petri dishes -- just as a message popped up in the foreground:
JSN regrets to advise you that your resources have been diverted to a higher bidder. A snapshot of your task has been preserved in mass storage, and will be available to you when you next log on. Thank you for using our services.
Maria sat and swore angrily for half a minute -- then stopped abruptly, and buried her face in her hands. She shouldn't have been logged on in the first place. It was insane, squandering her savings playing around with mutant A. lamberti -- but she kept on doing it. The Autoverse was so seductive, so hypnotic . . . so addictive.
Whoever had elbowed her off the network had done her a favor -- and she'd even have her fifty-dollar log-on fee refunded, since she'd been thrown right out, not merely slowed down to a snail's pace.
Curious to discover the identity of her unintentional benefactor, she logged on directly to the QIPS Exchange -- the marketplace where processing power was bought and sold. The connection to JSN had passed through the Exchange, transparently; her terminal was programmed to bid at the market rate automatically, up to a certain ceiling. Right now, though, some outfit calling itself Operation Butterfly was buying QIPS -- quadrillions of instructions per second -- at six hundred times that ceiling, and had managed to acquire one hundred percent of the planet's traded computing power.
Maria was stunned; she'd never seen anything like it. The pie chart of successful bidders -- normally a flickering kaleidoscope of thousands of needle-thin slices -- was a solid, static disk of blue. Aircraft would not be dropping out of the sky, world commerce would not have ground to a halt . . . but tens of thousands of academic and industrial researchers relied on the Exchange every day for tasks it wasn't worth owning the power to perform in-house. Not to mention a few thousand Copies. For one user to muscle in and outbid everyone else was unprecedented. Who needed that much computing power? Big business, big science, the military? All had their own private hardware -- usually in excess of their requirements. If they traded at all, it was to sell their surplus capacity....
However, what actually happens is that when a problem arises, the outsourced company addresses every single incident individually and leaves the problem untouched, since from their point of view it's the golden goose laying eggs for them to collect.
I have seen exactly this happen, and to be fair they were paying per incident rather than for a specific level of service to be maintained which probably made that ending inevitable, however an SLA only seems to move the problem elseware. It doesn't address the problem that a company hoping to save money outsources IT to a company who's business who is intending to make a profit by charging as much as they can for as little as they can get away with.
My company buys crap bulk. I still build my own and so can anyone else. But nobody wants to pay for in-house wrench turners.
ISP? Not given a choice. Which is why we pay too much for the worst service.
Cloud? Putting your critical data on it is teh stupid. Just ask the now hundreds of companies who've had their data hacked to the tune of millions of users.
Go on and see for yourself. You can use the google. It's a cloud thingy, too. (whose search engine is becoming less reliable year after year)
"Major adoption of public cloud computing services by large companies won't happen until the current crop of IT workers are replaced by kiddies who are too stupid or simply don't care enough that cloud services have major outtages every other day".
If you want people who grew up caring about what happens to other people - i.e. the people we work with and support - to adopt cloud services you'd live up to those 99.99% uptime claims that generally go *poof* as soon as it gets a bit windy near your datacenters or some noob you hired on the back of his ability to fuck around with photos in Instagram decides to press the wrong button.
I would be in favour of much increased use of Cloud infrastructures.
Not that long ago a mansion that wanted electricity had to get it's own steam-powered generator. It had to be switched off at night because it could not be safely and cost-effectively run, etc. etc. all the technical hero stuff. It has now usually been replaced by centralized networks that are much more reliable and still hire technicians and other staff to keep things running. They do not change Voltage on a whim of management but they keep everything running.
The solar-panels, generators for special purposes etc, also exist (usually with very little maintenance). I do not see why people should be so hung up on co-locating services, that require high up-time and standardization for many companies, with the business that uses it.
For most standard usage standardized affordable services can be the norm. why should one be against that?
Salesforce, Amazon AWS etc. are already being widely used and I see no reason to expect that to go down.
Different service-levels for different tasks and kinds of work are possible over networks, why try to block that?
Bring it on, responsibly I say.
I remember living in a house in Africa where we had a steam powered generator because the power from the centralised network kept failing all the time. I hope they don't move the cloud to Africa... Or for that matter rely on BTs network to connect us to the cloud.
The good thing about 2000v cable is, people are less likely to mess with it.
Best way to do it.. if you are going to make some "cloud" app, make sure it runs on OpenStack (which is Amazon EC2 and S3 compatible). Then you can switch between Amazon, a couple other OpenStack providers, or getting some Ubuntu and/or Redhat servers and running it on your own "private cloud".
This is pretty much how things used to be. Computer shops with a mini-computer and a bunch of tethered users with terminals.
The cloud will take off when the kinks are worked out. For many applications the kinks were worked out decades ago already - e.g. web shopping and shipping. The rest will follow in due course when it makes sense to do so.
"The next 18-year-old guy that we hire to be a developer here will have never experienced the client/server era."
Not quite sure of the point being made here, but my take in it is that the next 18-year-old will have no concept of enterprise application design and programming, since Client/Server (2-tier, 3-tier & n-tier) is a design paradigm, which is a slightly different issue to physical v. virtual hosts and still applies to cloud deployments...
"Who employee's 18-year-old's and expects them to actually be able to write business fit for purpose programmes without supervision from their elders..."
The word you were reaching for is "employs". HTH, HAND.
But to answer the question, off the top of my head? amazon, facebook, twitter, yahoo, google, myspace, youtube ... Starting to see the problem yet?
Yes sometimes the brain-hand-eye system malfunctions, but at least you understood the point.
"Starting to see the problem yet?"
The problem is much worse! as the majority of 18-year-old's will have no real idea about computing full stop (if you are 18 and reading this, I suggest you will understand what I'm saying when you're gained a few more years experience). Remember even graduates aren't considered 'suitability skilled' until ~2 years postgrad...
10/10 for bias reporting.
Cloud isn't a technology, it's an ABSTRACTION layer of the technology underneath.
In all honesty, "kids" coming through wanting to just "spin up" VM's and talk about infrastructure at a high level doesn't remove the existing technical expertise underneath required to underpin it.
It's all very well the likes of Rackspace preaching about Cloud, but from experience they have NO IDEA what so ever about segregation of data, cryptographic controls, secure DMZ environments and they never will, because the clever guys cost money, and hosting companies don't pay top money!
I've seen this and similar opinions on several articles from various sources and all I can say is this: It needs to die. Not only is it counterproductive, but it lends credence to the folks who read a sales pitch and believe that a new service can replace an experienced IT team.
Instead of wasting time wishing IT would go away so we can start living in the clouds with perfect outsourced environments, perhaps we should focus on growing the IT role in the new deployment models. Giving businesses more control over their IT infrastructure and strategy is great; it is the cornerstone behind the success of "as a service" architectures. But responsible expectations should come with it.
From the IT point of view, we need to wise up too. As a long time DBA I am well acquainted with many people who cringe when NoSQL is mentioned, walk away when Hadoop is discussed, and flat out go into a babbling stupor when cloud is brought up. By growing ourselves and understanding the new role of IT in business, we can easily become a part of the process and not the bottleneck we're perceived to be. From the DBA point of view, I wrote about it here in an article titled "DBA, Grow Thyself": http://www.oraclealchemist.com/news/dba-grow-thyself-moving-and-shaking-in-the-era-of-data-dominance/
Do IT professionals need to learn new things? Of course. Should businesses be able to control their IT destiny a little better? Absolutely. Does everything need to get turned upside down just to make it so? Not a chance.
The Cloud is just a remote server you have no real control over. No privacy. No proper management.
So why would anyone with critical data have it all in the hands of some third party outfit, thousands of miles away, with no control of how their data is managed and protected?
But hey, I'm still waiting for the Flying cars too. Surface roads were to be obsolete by now, remember?
As an IT Grumpy Old Man I have deep reservations about using Cloud Hosting services.
If you use a company that even has a presence in the US your data is there for the Feds to snoop. You won't know about the court order graned by some drunken judge in Hicksvile Nebraska but they can get access to everything you put in the cloud.
Even German Privacy Laws can't help you here.
All it takes is one Fed who has a friend/brother who works for a rival and your company is probably down the tubes and you would never know why.
The Feds openely declared aim to snoop on computers everywhere should be taken as a direct threat.
Don't put your stuff in a public cloud.
Don't let 'others' have access to your data.
Sigh, the modern generation don't give a damm about this. They just think it is cool to have everything hived off into the cloud.
Black Helicopter naturally.
The average company connecting to cloud services is doing it over their standard Internet connection without dedicated leased lines etc direct to their cloud providers. If there is a big outage on the Internet, however localised, a lot of companies are going to be a lot of thinking about resiliant access to the Internet and hence accessing the cloud services whereas most have been worried about uptime of the cloud services themselves.
Not to mention that your Internet connection is a prime target for DDoS attacks when you rely on it for the internal functioning of your business as well as its external communications.
You can't/shouldn't simply suck the IT out of a physical business location and dump it somewhere else - it really only makes sense to "clolud" your IT if you do the same for the people and processes in your business.
Which of course would suit Amazon just fine: you don't need any warehouses or store, they'll provide an entire retail infrastructure for you. And that's the problem - any business model that can be put into the cloud will end up being a business model in the control of the cloud provider.
Not exactly new though is it? Back to the DP centres of the seventies methinks. In IT we do a bloody good job of reinventing the wheel over and over and over and making it look like innovation. I mean, look at web apps for instance. Basically they equate to a pretty dumb terminal app of yesteryear. Browser = dumb terminal and app runs on server. Not exactly innovation. Of course I am oversimplifying but there are only so many ways you can skin a cat. The industry need you to believe this is new so it can sell you more. It's like the death of the desktop in favour of the mobile world. We are sold on it. We haven't stopped to see if it really is a good idea. I mean using an app on a phone. Its a pain in the arse if you ask me. Typing a letter on a tablet. Do me a favour. The keyboard has never changed because it works. Beautifully. Voice recognition ??? Bloody noisy world if that was the chosen input mechanism of the masses. I saw 2 people using their tablets to video a darts match the other night. Bloody hilarious. Quite simply the most inappropriate tool for a job but the marketers would have you believe otherwise. Final rant :-) 3D TV. Now that's a laugh. Its not 3D for a start because you still cannot decide what you want to focus on. That is predetermined by the film maker. It has to be. So, it screws with your head and gives you headaches. The marketers wont tell you that cus they need something new (which it isn't) to sell more TV's that we don't need. Most of don't have room for TVs much above 42" in a living rooms so why do we need 4K (or indeed 1080p). Our eyes cannot resolve that detail at the distance we are likely to be away from the TV.
I am think things may be a little more cloudy in the future, but me thinks it be more private...
and if hes relying on Facebook for future sysads...shutter*
I don't know why they keep trying to make cloud out to be some abstract idea. Its as if these marketers pushing cloud don't think we have the foggiest idea of how it operates.
"You can have my servers when you pry them from my cold dead hands!"
Like I am going to trust a 3rd party with my proprietary data/code/stats.
Oh the cloud is our raining data everywhere! waily! waily! waily!
I have no access to our production data!
we have been HACKED!
Our data is in WHAT COUNTRY!?
unless these companies can offer the same level of accountability that our internal IT department has for security/data access/SLA uptime/etc and offer the quick turn around and fix that an internal group who is dedicated to the functioning of the company (and not simply say Oh well not my fault it is our cloud supplier, which seems these days to be the classing newgen lack of personal responsibility thing they show an awful lot of these days. There is no way in H E Double-hockey sticks I will let anyone else fondle my data.
We recently had a problem that one of our competitors bought out the company that supplies us with a certain application we use for business and offered cloudy based services for the same app... erm no! we are not letting you anywhere near our data be it from our clients or our own.
and the next time someone is doing something illegal on their part of the cloud and the feds come in and shut the entire thing down so potentially thousands of clients will be locked out of their data..
the cloud is good for an (encrypted) backup solution to compliment a rigid and accountable internal infrastructure, it should never be used as a primary data source unless you are a small startup, I can see this is good for small business, but for the price of a typical cloud solution you could probably have your own small infrastructure and know exactly where your data is and who is looking at it.
just my opinion
I am going back to field stripping my servers while muttering something about hari Krishna now...
brains enough to know that so long as organizations like Sony, banks, US gov, Can. Gov, etc. get hacked and things Iike DDOS, botnets, etc. exist it makes absolutely no sense to put your business in a cloud on the Internet.
I got rid of previous hosting company because email went down and was down for days with no response from their support, I could just imagine having the entire company in a cloud.
The thing that ensures that the cloud will never be fully adopted is the fact that any company that hosts cloud services cannot guarantee that they will not go bust at some point. Not possible. Bearing that in mind, what right minded CIO/CTO/CFO would move his company's sensitive data into the cloud (or any data for that matter). The cloud (or data centre, or DP centre or whatever name you care to choose) merely exists to gain control. Think about it. Host all your data and apps with us. We'll save you a bomb. OK dude, why not. Hey now you host all your crap with us how about you pay us a shedload more. I don't want to. Well tough crap dude you aint got a choice anymore.
About how large the area for attack could be for cloudy solutions, currently you host in house you have a nice server room with all the cool stuff, only access by a few special people who know what they are doing, or at least should know what they are doing... You sql talks to your web machines and other machines, over a nice secure internal wired network, you have total control over the gubbins of the stuff, and are not limited by latency or other fun stuff, as far as in practical anyway. So now we host all our stuff in the cloud, which by definition is basically a colo or lots of colos now do you attach to your app via the same interface as joe public do? what about any backend work you need to RDC or equivalent to the box in the colo(s) which you have no physical access too? so now you need RDC or other ports open to access the machine that you, before could have walked through nice secure door to, and opened up the kvm console... now it is up to someone else to do this, probably of the same generation who want to put everything in the cloud ( hell why don't they put the cloud in the cloud then the cloud is always available) so now potentially some twat who is probably posting pics of his work on facebook with his phone can potentially get access to your hardware and software though admittedly he is probably too dumb to do anything with it.... "err it has lights...they blink.... if they stop I hit this button... oooOOOOO" either way you have opened up a potential hive of MITM exploits. the more I think of it the more I think it is horrid.
the cloud is for your drunken spring break pictures not important stuff that anyone can grab hold of.
10-15 years time and businesses will all be using the cloud? The 'Kiddies know best?
Guess I must be a dinosaur as I'm over 50 and one of the people holding the cloud back unnecessarily. Couldn't be a bit of self interest in making that statement could there John? After all you are Rackspace CTO. But as CTO how are you at running a business?
Would you trust AWS or MS to host your whole IT operation? Would you recommend it to your board? Yes I know its a silly rhetorical question, but there are many reasons why it's not being adopted as fast as you may like.
1. It's still new and having teething problems. Yes it will settle down, we'll learn from major outages and we'll improve it. Personally I think adoption in 10-15 years is way to far off. If the suppliers get it right, 3-5 years may be more appropriate.
2. It's not a strategic option for ALL business data. Strategically I support cloud computing for customer facing activity. Web site are a great leader in this field. Can anyone accurately predict how many people are going to hit a web site after a TV ad? With the cloud, its not a problem. In fact a perfect solution. With business critical data.... Much more caution is needed. Even the humble finance system's data if it got into the wrong hands by mistake could give a competitor a massive advantage....
3. Can we really trust the suppliers? Nothing personal, but in my time I've seen suppliers be really nice and friendly to get your business, but if you want to move your business or they go out of business. Who own's YOUR data? In todays world, its simple, if a supplier get shirty when you want to change, or goes bust you have protection because it's all running on your hardware and will continue to run on your hardware (yes you will have problems) so your business will still function, it will also only be one supplier not every single bit of IT you use that goes or has problems. Even your 'kiddies' understand this argument, so you can imagine how easy this argument is to sell to a board.
May be I'm a dinosaur - but I make a lot of money out of being a dinosaur and keep being re-employed simply because I keep my thoughts grounded in the reality of the world not the spin. :-)
This whole cloud thing is a dilemma. I want to promote co-lo, I want good hardware all under my control (And by good, I mean good, not silly expensive for kicks). However, the execs want things now and they don't keep you upto date with any of the planning or future business strategies, hence you are given a tiny window to provide a solution.
Perhaps the best reason to employ a cloud service is the willingness of the execs not to spend money on staff/kit, but to pay a monthly cost.
Biting the hand that feeds IT © 1998–2021