back to article Public enemies: Azure, Amazon, Google, Oracle, OpenStack, SoftLayer will murder private IT

On-premises IT is facing decimation by six public cloud enemies: Amazon, Azure, Google, OpenStack, Oracle and SoftLayer, who are on course to have the majority of customers' IT spend by 2018. An Internet retailer, started up as an online book store, and a website search facility have kickstarted an IT phenomenon which could …

  1. Ken Moorhouse Silver badge

    Centralisation/decentralisation Insourcing/outsourcing

    This is another one to add to the list of cyclical fads.

    Yes Cloud will get bigger, but then I reckon that there will be a reversion to on-premises. A few more security scandals, a few more changes of ownership of players in the market and people will be running for the exits.

    The "edge" that corporates strive for is how to differentiate oneselves from the competition. "Everything we do is in-house, we care about security" is a strap-line that will be wheeled out as a differentiator when the cloud cycle takes a downward turn.

    1. SirWired 1

      Re: Centralisation/decentralisation Insourcing/outsourcing

      I strongly suspect that what does move out of the cloud back into the DC will move into "private clouds", or "cloud-in-a-can" solutions that remove most of the complexity of traditional IT solutions.

      A cloud-in-a-can requires sophisticated software. Hardware? Not so much.

      1. DougS Silver badge

        Not so fast

        I agree that privacy concerns may cause those future projections to be a little iffy. All these big cloud players are US based. Even if they maintain data centers in Europe, are European companies comfortable with housing their really important data on Amazon or Google servers? Are they truly compatible with EU data protection laws? Even if they are today will they be tomorrow if a new administration and new terrorist attack makes people forget Snowden and rush headlong into more corporate cooperation with the US government?

        1. a_yank_lurker Silver badge

          Re: Not so fast

          A couple of reasons for some wariness about the cloud. It is being hyped a panacea for all that ails the DC. It is not, it solves problems but creates its own. Secondly is control of the hardware. If you have the DC and staff in house you have complete control. If it is outsourced, the Cloud, you only have partial control at best. Are these deal killers only specific situations.

          My suspicion is the cloud will continue to grow until an equilibrium is reached with in-house and cloud solutions. Each company and industry will have a different ratio between the two.

          1. Ken Moorhouse Silver badge

            Re: equilibrium

            I used to work for an organisation where mention of the word "software" in a proposal would get waved through much more easily than if big amounts of money were to be spent on "hardware". As a result, the programming team had to deduce, using software, events that would have been obvious to detect if they were presented to us using hardware. This was at the time when software was "fashionable", whilst hardware was "so yesterday". The software costs ballooned as a result, compared to if a more hybrid approach had been chosen.

            For a contemporary comparison, from what I have seen, the ongoing costs of providing say 20 mailboxes in the cloud is ludicrous compared to the cost of hosting your own mail server, yet the finance chiefs are authorising it, purely because it is a fashionable (Cloud) solution. The hidden software costs of that historical project are analogous to these ongoing rental costs.

            1. Vinyl-Junkie

              Re: equilibrium, @Ken Moorhouse

              "the ongoing costs of providing say 20 mailboxes in the cloud is ludicrous compared to the cost of hosting your own mail server"

              However, if you're hosting 16,000 mailboxes the costs are about 40% over 10 years of an on-prem system...

        2. Tom Samplonius

          Re: Not so fast

          "...are European companies comfortable with housing their really important data on Amazon or Google servers? Are they truly compatible with EU data protection laws?"

          Apparently they are comfortable with lots of hypocrisy. The UK intends to require ISPs keep records on every accessed internet site for 12 months. Oh, its protected by a court order and so it is fully compliant with EU law. Try doing that in the US.

          And what about the German "outrage" two years when NSA was revealed to have spied on Merkel? German investigators haven't been able to prove that actually happened. But they have dug up evidence that the German BND spied on pretty much every country they could:

          http://www.worldpoliticsreview.com/articles/17182/spying-on-friends-germany-s-bnd-scandal-puts-snowden-leaks-in-context

          Why is there a widely held belief that US data protection is worse than EU, when evidence indicates it is the same or better? Euro-jealously perhaps?

          1. Anonymous Coward
            Anonymous Coward

            Re: Not so fast

            Why is there a widely held belief that US data protection is worse than EU, when evidence indicates it is the same or better? Euro-jealously perhaps?

            It is not a belief, it is a fact. Don't look at laws in isolation, look at how they fit in with other laws, and look at how they are enforced. From what I have seen over the last few decades, US companies are apparently allowed to treat laws as optional guidelines that can be ignored if they get in the way of profit, all you need to do is grow fast enough to buy your way out.

            Simply due to structure, this is at present much harder in the EU. Give it a decade or so, though, and it will be similar as EU parliament cooks up stupidity that eventually even the European Court of Justice cannot undo - I reckon Facebook v Max Schrems was the last time the ECJ is allowed to annul a comfy arrangement by EU Parliament to allow US companies to ignore EU laws too. Some lobbyists must be *seriously* pissed off that their investment in buying MP favours didn't deliver the required results.

            /cynic

          2. LucreLout Silver badge

            Re: Not so fast

            @Tom

            Why is there a widely held belief that US data protection is worse than EU, when evidence indicates it is the same or better? Euro-jealously perhaps?

            Please explain how American data protection provides equal of greater protection to EU citizens within the EU? It may or may not provide better protection for US citizens in the US, but surely you can see why those of us not in that boat are asking questions?

          3. KeithR

            Re: Not so fast

            "Why is there a widely held belief that US data protection is worse than EU"

            Because it bloody well is - you buggers haven't got a CLUE what privacy means: it's just a barrier to profit, according to you lot.

            "when evidence indicates it is the same or better?"

            "Evidence?" Let's see it - not one cherry-picked example that confirms nothing.

            "Euro-jealously perhaps?"

            I can't think of a single damn' thing to envy the US about. You've got TRUMP, FFS - that right there is enough of a reason to mock and pity (I guess that not EVERYONE in the US buys into his shite) the US for all time.

            1. John 104 Silver badge

              Re: Not so fast

              @KeithR

              "I can't think of a single damn' thing to envy the US about."

              I can think of a few:

              The lack of cameras on every street corner watching your every move.

              Ridiculously high tax rates. (seriously, a tax to use the receiver in your TV? Who comes up with this stuff?)

              A DNA swab upon every arrest, stuffed into a database for ever, even if you are acquitted or no charges are filed. (not that I've ever been arrested)

              Sounds like the same old nanny state Britain that we left behind a few hundred years ago...

      2. Anonymous Coward
        Anonymous Coward

        Re: Centralisation/decentralisation Insourcing/outsourcing

        "A cloud-in-a-can requires sophisticated software. Hardware? Not so much." Which is the strategic differentiator the Microsoft/Oracle/IBM want to draw with AWS/Google. That's going to be fun, as in popcorn, to watch.

        And there's always OpenStack for Cloud-in-a-Can but for some reasons/feelings I have, it may not break out of being only a niche play.

      3. LucreLout Silver badge

        Re: Centralisation/decentralisation Insourcing/outsourcing

        I strongly suspect that what does move out of the cloud back into the DC will move into "private clouds"

        My bank is moving wholly to virtualised PCs in the near future, which is what we'll be using our old datacentre to host in a private cloud. The stuff that was in the DC is already being gradually moved to the cloud, with a lot of the major software components being moved to a SaaS subscription.

        I expect there will be some form of market for that, while PC manufacturers swallow both barrels, having been relegated to the retail only channel.

    2. Destroy All Monsters Silver badge

      Re: Centralisation/decentralisation Insourcing/outsourcing

      It's as cyclical as the adoption of the Internet. Cloud infrastructure will become better, kinks will be ironed out. The problem of the "big 3" providers will be constant though, so small-scale players may stand a chance. Alas, it will be niche same as companies doing "Everything we do is in-house, we care about security" will be niche.

    3. allthecoolshortnamesweretaken

      Re: Centralisation/decentralisation Insourcing/outsourcing

      Pretty much this.

      Also, sooner or later a bit farm is bound to crash, big time, the equivalent of a Tchernobyl meltdown. Or at least a "The hell! Goddamn machines, anyway! That's not supposed to happen!" event. Every system can fail, and some of them will fail, no matter what.

      BTW, there is a word in German for cyclical fads: Schweinezyklus. Says it all, really.

      1. allthecoolshortnamesweretaken

        Re: Centralisation/decentralisation Insourcing/outsourcing

        More a nuisance (if you're lucky) than a meltdown - still, this is the sort of thing that will happen from time to time:

        http://www.theregister.co.uk/2016/02/18/azure_lost_some_virtual_machine_backups_for_eleven_hours/

        Cue PHB: "We must prevent that sort of thing! I know - let's back up all our cloud stuff on premise!"

  2. SirWired 1

    Yeah, I've noticed that too

    I've been an IT Infrastructure guy my entire nearly two-decade career. (Started out with tech support for LAN gear, shifted to 13 years in tech support for SAN gear, and now I'm a DR solutions architect.)

    On a lark, I took a cheap AWS Architect course, and it totally dawned on me: My entire profession, IT infrastructure, is toast. Kaput. Kablooie. There are few reasons for anybody setting up a new IT shop to use the sort of environment I've spent nearly a couple of decades becoming a fairly well-paid expert on. The flexibility and ease-of-use the cloud provides means nobody has to pay somebody like me to make it all work. I'm really good at SAN design, remote mirroring, nitty-gritty protocol gnarliness, etc., and now there's no need for any of it.

    And yes, it terrifies me. At this point, I'm learning as much as I can about cloud architecture as fast as I can, but I fear it may not be enough. (As in, I'll be able to find a new line of work, but it certainly won't pay as well as what I'm doing now, because I won't have the experience levels I have now on the new stuff.)

    Now I know what the proverbial buggy-whip makers felt like.

    1. Adam 52 Silver badge

      Re: Yeah, I've noticed that too

      We still use infrastructure engineers. They do slightly different jobs - provisioning via software rather than hardware - but many of the disciplines (monitoring, backup, DR) remain in a different form.

    2. Esme

      Re: Yeah, I've noticed that too

      @SirWired1 - buggy makers? No, you don't need to go back that far. I used to be a mainframe operator on IBM 4381's and 3090's. Then tools like Omegamon started appearing, and I realised the end was in sight for most of what I did in my job, aside from changing the tapes. As it happened,my career took a sharp left turn before the demand for mainframe operators dried up, but I wasn't at all surprised, some years later, to find that the job I'd once done effectively no longer existed, as software did the same job, only better.

    3. LucreLout Silver badge

      Re: Yeah, I've noticed that too

      I'll be able to find a new line of work, but it certainly won't pay as well as what I'm doing now, because I won't have the experience levels I have now on the new stuff

      Nobody will SirWired, nobody will. What will determine your pay rate will be how much experience you have relative to your peers more than in absolute terms. Seems to me you're making the right moves and what you need to do is keep doing what you're doing.

      That being said, people like hedge funds will never use public clouds - they'll always have their own kit and networks guys. I imagine many smaller R&D focused companies will do the same, so I'd not be willing to call time on your main skill set just yet.

    4. Anonymous Coward
      Anonymous Coward

      Re: Yeah, I've noticed that too

      There's a huge shortage of skilled cloud architects and the pay's pretty good.

      Be glad that you now 'get it' very many are still comfortable in the old world, you're on totally the right track and just need to get experience in Cloud and you'll still be well ahead of the pack, they will have their moment of enlightenment just like you have but its still fairly early days.

  3. cantankerous swineherd Silver badge
    Mushroom

    triple point of failure? perhaps quadruple or quintuple?

  4. Anonymous Coward
    Anonymous Coward

    well

    It's not quite as grim as depicted. Yes, SaaS, PaaS and IaaS will continue to grow at the expense of on-premises IT. However, it is highly likely that it will never be a cloud-only computing world. The dominant deployment model for the foreseeable future will be hybrid. Not strictly hybrid cloud, but hybrid-IT as well, and they are not the same but one could argue there is overlap in the concepts. My advice is always to use public cloud where it makes sense and don't where it doesn't. And, sometimes it doesn't, for myriad reasons. Cloud is merely another tool in the IT tool chest.

    1. Preston Munchensonton
      Mushroom

      Re: well

      Cloud is merely another tool in the IT tool chest.

      Exactly. Any fool thinking that this becomes a Cloud+SDN+NFV nirvana is ignoring the various use cases for IT in general. A highly virtualized environment might suit the many cases being bantered about here, but anything that requires performance or compliance should never, ever, ever be put into a public cloud. Ever.

      This entire discussion really is not about cloud vs. standalone, but just an extension of the insource/outsource debate. There are bloody good reasons for both and, like always, you have to find the right tool for the job. Assumption of a customer's requirements is the surest way to demonstrate how quickly you can find the door.

  5. Jim O'Reilly
    Alert

    It's worse!

    Chris is pretty well on-point with his predictions about the cloud winning, but he missed a couple of other things. First, the Chinese ODMs are well positioned to attack the US incumbent traditional vendors...that's because they supply AWS etc with huge quantities of gear already and also because they make most of the gear for those US vendors! If they roll into the market with distribution and end-user sales, they'll undercut the US "vendors" by 60 percent or more on price...a classic "cut out the middleman" scenario.

    Second, the poor state of WAN connections in the US and EU makes the hybrid cloud model awkward to implement. We need fiber connections, but it takes bullying by Google to get Verizon and ATT to budge.

    If hybrid cloud isn't attractive and ODMs are cheap, the traditional vendors are going to hurt a lot.

    1. Preston Munchensonton
      Stop

      Re: It's worse!

      Second, the poor state of WAN connections in the US and EU makes the hybrid cloud model awkward to implement. We need fiber connections, but it takes bullying by Google to get Verizon and ATT to budge.

      I can't speak for the EU, since I haven't been involved in any European WAN designs since 2010. But you could not be more patently wrong concerning the US. Anyone hosting anything in the US, whether cloud or otherwise, has a bevy of carrier options available. In fact, it's really presumed now that any data center (or centre, ahem) has to have dual entrance for connectivity, with at least two Tier 1 ISP connections, to be taken seriously. And those connections are moving to 40Gbps or 100Gbps in many cases, which have to be fiber (or fibre) to reach the distances involved.

      Google itself isn't really the driver of all this. This process, at least in the US, has been going on for a long time. The aptly-named carrier hotel has been a mainstay of WAN for as long as I can remember and I'm sure that's still the case in the EU (if my recollections of Colt, BT, and Global Crossing mean anything).

  6. elDog

    And the downside is?

    A long time ago some savant said that the network would become the computer and the computers would just be access points.

    My very hefty (capabilities) but light-weight and portable tablet is the same as the old 3270 smart terminals of yesteryear. We're still in the client-server mode but with the improvements in accessibility and reliability the server is somewhere out there, and we don't care where.

  7. Anonymous Coward
    Anonymous Coward

    The potential advantages of moving to the cloud are as clear as ever, even Orcacle dropped "non-shared" as the main concept in favour of cloud.

    To what extent the world will change to one where hardware/hardcore tech are separated more from content and its owners and users will be interesting to see.

    - Will hardware and its management be largely be the domain of specialists with at least 1000s of users?

    - Will many devices be connected to data-centres to securely process the info generated (instead of users with papers reading and forwarding paper, particularly the medical word is full of it).

    - Will standardization lead to customers picking offerings from competitors?

    - Will most devices become "rolling" for their system-software with a layer of embedded on top of it?

    It will be interesting to see where we end up.

  8. LB45
    Facepalm

    Why not? What could possibly go wrong?

    The scandalous security breeches to date and data center "outages" haven't been big enough yet to inflict enough pain to deter "cloudy" business plans.

    Eventually there will be a big enough slip up that will end up in severe business losses and the usual lawsuits to recover damages that examples will be made.

    Even then it'll never go away but completely trusting others with your storage and data will not be taken as lightly as it is today.

    Eventually IT will cycle back to the next big thing and it'll be interesting to see what place in the grand scheme of things "private IT" winds up.

    1. tom dial Silver badge

      Re: Why not? What could possibly go wrong?

      The most notable security breaches seem to involve organizations that own and operate their own infrastructure - US Office of Personnel Managament, Sony, Target, and Anthem come to mind readily. It seems quite possible, or even probable, that the very large cloud services can apply the expertise and resources to make their services more secure than the average large business, let alone small or medium size organization, or individual consumer, can even remotely hope to do.

      1. Adam 52 Silver badge

        Re: Why not? What could possibly go wrong?

        I'll second what Tom said. AWS can devote much more effort to security than any SME.

        But I'd also caution that the new models make it very easy to shoot yourself in the foot. There have been a couple of news stories about AWS keys checked into github, for example, and the temptation is for developers to bypass good security (for example by using DynamoDB and S3 for personal information - DynamoDB is Internet facing only, S3 needs policies to make it work with a vpc). This is even worse with the PaaS vendors like Heroku where everything is inherently insecure.

        To the other comment, this is why you still need Int Engineers to keep the developers in line.

    2. Warm Braw Silver badge

      Re: scandalous security breeches

      I must get myself a pair for my next assignment...

      1. Vinyl-Junkie
        Coffee/keyboard

        @Warm Braw

        Thank you for a Wednesday morning coffee redistribution moment! If I could put two icons on you'd have a beer one as well; as it is, have an upvote!

  9. Steve Chalmers

    Watch the storage, not the compute

    When it comes to the cloud, pay attention to the data. Moving data into the cloud is hard. Getting it back is harder, particularly if your cloud service provider goes out of business on short notice.

    Also pay attention to concentrations of risk. It's no big deal if some archive you rarely need access to is offline for a week. On the other hand, a big bank or trading firm would be put out of business by an outage that long, and prepares accordingly. Remember Amazon's outage a few years back, when a network partition within the piece of a data center which held a lot of customer data suddenly caused essentially all of the computers holding that data to think they were the last surviving copy, each then attempting at very high priority to copy itself?

    I've often said it takes a decade to establish a new server to storage connection, and at least two decades for a server to storage connection to fade away once a replacement technology reaches critical mass. In the cloud, that connection is an API ecosystem like S3, not a nominally hardware centered ecosystem like Fibre Channel. My point is not these details, it's that it would take decades to migrate all of enterprise IT to the cloud, even if the cloud were provably a better business choice today.

    That is not to say the the cloud hasn't captured 100% of a new generation of applications, and some existing corporate IT as well. Shifts like this in customer choices were the nature of the technology world when IBM brought the System/360 mainframe to market 50 years ago, when Digital introduced the VAX series a little under 40 years ago, when Compaq created the plug-compatible, software compatible PC 30 years ago, when Fibre Channel was introduced 20 years ago, and so on. Most of the time customers figure out where the new technology makes the most sense, where the older technology makes the most sense, and the market evolves accordingly over a period of years (sometimes even decades).

    Back to the point: watch the data. If corporate customers (small, medium, and large) start moving all their data to the cloud, the compute can follow (the compute is the easy part).

    1. P. Lee

      Re: Watch the storage, not the compute

      I was just thinking the same.

      I suspect correlation rather than causation when on-prem storage from EMC takes a nosedive and cloud is on the rise. Are people really moving their EMC-based data to the cloud? I don't think so.

      Spending on EMC may have dried up and spending on cloud might be increasing, but that just means EMC has saturated its market, not that people are no longer using EMC. It is a mature market, not a dead technology.

  10. Anonymous Coward
    Anonymous Coward

    That picture of Jeff reminds me of Vizzini from "The Princess Bride". You know.. right before he falls over dead...

  11. abedarts

    Software as a service, like electricity and water...

    Doing it yourself on site will become as daft as generating your own electricity and arranging your own water supply. Of course the mains can fail or someone can poison the reservoir, but on the whole we trust the suppliers to keep things running. The tiny percentage of downtime does not justify DIY.

    1. Ken Moorhouse Silver badge

      Re: Software as a service, like electricity and water...

      I grant you that the concept of providing certain aspects of software as a service are within the realms of treating them like the supply of electricity and water. Examples of this are computational, such as SETI, and to enhance deliverability, such as BitTorrent. These are used to improve some aspect of the ability to get results more quickly than if you were doing things purely by yourself. Cloud has a role to play in users "handing-off" to some third-party "here you are, deal with this and gimme the answer, quick, please." The task being delegated is icing on the cake, it is not completely replacing something that is customarily being done on-premises. You are certainly not moving the only copy you have of something into the cloud, and delegating the complete running of your business to the cloud, which is what some people are aiming to achieve.

  12. smartypants

    Cherry-picking season

    "Not secure!"

    "Wait till that big cloud datacentre meltdown!"

    "Can we really trust them...?"

    etc...

    All problems with on-premise IT, but funnily enough only cloud gets this sort of criticism. In my last job, the on-premise IT was so bad that a £3 million project was lost. Only you never heard about it, or the thousands of other daily disasters because they don't make the news, and why would they? They're like car crashes... too numerous to be newsworthy.

    IT means a lot more than just buying or selling boxes and wires and plugging them together. Cloud takes away much of the need to bother with the physical boxes and wires but really doesn't change IT where it matters - the design, engineering and maintenance of systems.... And it doesn't remove the business need to manage cost of provisioning - just changes the rules you play by.

    If you're a box-seller, yes you're probably going to wave goodbye to much of your business today. If you're a box-plugger, look out for the latest mega-datacentre opening up round the corner from you next year!

  13. Locky

    Is it me...

    ... or does that picture of Jeff Bezos look worrying similar to Dr Evil?

    1. KeithR

      Re: Is it me...

      "Is it me...

      ... or does that picture of Jeff Bezos look worrying similar to Dr Evil?"

      Dunno - I've never seen a picture of you...

  14. elan

    its the economy.stupid.

    - analysts just extrapolate the IT production function.

    - advantages of the cloud compute model are already worked-in in new products

    - the new generation of start-ups got the IP and get the same resources for their hw.

    - remains a certain cost advantage for the bigX based on size, but not that big (10% ? )

    - well, balancing that with the risk (for your enterprise IT) completely loosing control of your assets (data, operations) ....Hmmm....I would not prefer to have a seat in the board ...

    - the cloud is an option, not more, not less

    -

  15. GrumpenKraut Silver badge
    Thumb Down

    Example chart modelling on-premises vs public cloud storage growth

    (the image at page 2).

    That is single handedly the most bizarre prediction I have seen in a decade.

  16. Anonymous Coward
    Boffin

    W00t!

    I am a cloud Ninja, that means I can invoice at £1,500/day

  17. Anonymous Coward
    Anonymous Coward

    You mean two

    I think it is more like Azure, AWS....... and the rest.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2020