back to article Look out, Amazon Cloud! HP's on the warpath

You might think the hard work for Hewlett-Packard is done, after it came from behind to build its own Amazon-style cloud so quickly. But the difficult part – taking on Amazon and winning with open source – lies ahead. HP The world's biggest PC maker lifted the lid yesterday on the biggest change to its business in recent …

COMMENTS

This topic is closed for new posts.
  1. Anonymous Coward
    Anonymous Coward

    As a long time observer of

    As a long time observer of (and former occasional purchaser of kit from) HP and more frequently its predecessors (CPQ, DEC), this looks like an interesting move. In some ways it should have been done five or ten years ago, but would the market have been ready?

    And customers of this service will likelystill be horrifyingly dependent on the relevant telecom infrastructure. This morning in the UK, the lovely people at BT Wholesale have an outage at one of a handful of broadband datacentres (Manchester, fwiw) which has blacked out broadband customers of smaller ISPs (including "boutique" ones which are allegedly quality focused for the smaller business) across much of the UK.

    Obviously bigger businesses will have invested in resilient redundant connectivity where it matters so incidents like this don't affect them and their usage of "the cloud". Won't they?

    What price the cloud when "stuff happens" like that?

    And no one really wants a Microsoft cloud, not even MS Premier/Gold/etc partners. Who'da thunk it?

  2. Anonymous Coward
    Anonymous Coward

    The Cloud - When your Data is not really your data

    If you are so silly to move your whole business computing into the cloud then frankly you deserve all you get.

    Telecoms outages are a fact of life.

    Try telling HMRC that you couldn't produce your accounts because the line to "somewhere over the rainbow" was down.

    Then what happens to all that lovely data of yours when your cloud supplier goes tits-up?

    What about security? How really separate is one bit of the cloud from another? Just wait until the script-kiddies start going after 'The Cloud'.

    Far too many risks for me.

    Another bubble to burst IMHO.

    1. Destroy All Monsters Silver badge
      Meh

      Re: The Cloud - When your Data is not really your data

      This is becoming tiresome...

      Shift the barrel "cloud" <--- "electronic records" <--- "paper records" and you have the bitching and moaning of the 70s.

      Turns out that some problems were solved, some new problems were discovered, some problems turned out to be unsolvable, but in the end, a new balance between paper records and electronic records was struck.

      It will happen again.

      Also: "What about security?" ... it's an USB stick ago.

      1. P. Lee
        FAIL

        Re: The Cloud - When your Data is not really your data

        I would agree that the problems are all solvable, the question is, has your cloud provider actually solved them.

        Security is a significant issue issue. It isn't just about whether you got hacked or not. Security is about knowing the risks and accounting for them. Try setting up a PCI-DSS compliant solution using a public cloud and let me know how easy it is. It isn't just "we trust the provider" its "we have audit capability, we know who the admins are and we can show exactly who accessed what and when." Without that detail, you probably haven't gone through the rigour of designing a secure solution.

        It is also a legal thing. You can't usually export banking data or personal data around the world. You will get fined. That is awkward and commercial suicide from bad design.

    2. Anonymous Coward
      Anonymous Coward

      Re: The Cloud - When your Data is not really your data

      @ Former Job precisely all of those arguments came up - However, It was very easy to document that Former Job really don't give a shit:

      We had Students working Operations, Sacked BOFH holds the passwords to main routers, the configuration of which are crazily complex due jumping through burning tires for random customers, the same four/five passwords used everywhere - those also shipped with on-premise systems, LDAP configured, but in such a way that it has admin rights on the LDAP server (password in LDAP config EVERYWHERE) e.t.c. e.t.c.

      Then you add the general and wide-spread incompetence in setting up databases, firewalls and CISCO-shit alle over industry, where most people sell a service, at a price and production costs must be low so nobody bothers to learn the tools because they are not going to be there for long.

      All things considered, The Cloud make a lot of sense. I seriously doubt that AWS is run the same way as I have seen IT being run. At the worst it is the same level of negligence and incompetence - but with huge savings in Operations staff and hardware for the users of the service.

      Of Course people took a dim view of this being put into a report so when they sacked half the staff I was sacked too - which actually turned out to be like winning the lottery!

      PS: Former Job is still shovelling shitty software around between shitty boxen mosly by hand - not because this stuff is hard, a largeish boto-script can run up everything on AWS in minutes (been there, done that) but because the Chief Opreations Officer then would be without minions and domain, and one can never have too much management!

    3. the-it-slayer
      WTF?

      Re: The Cloud - When your Data is not really your data

      Why's cloud computing any different to hosting applications on dedicated managed services or virtual services hosted on shared hardware? None! It's just the technical definition of how hardware is setup to make the use of mass resources and being flexible in what you use.

      Script-kiddies? You lack any knowledge on this completely. Yes, your data might be spread across different physical kit, but seen as one logical instance; whether that be a computing or storage. Of course, the security will be much more complex with different cloud setups needing different firewall requirements, but you'd generally think the better providers have top notch kit that's setup in the best way.

      It's no different than a tailored setup run by the company itself in a data centre. If it's poor, then it's poor. Same as cloud provider. If it's setup poor, it's poor.

      1. Anonymous Coward
        Anonymous Coward

        Host vs Cloud for 100 please Alex

        @the-it-slayer sez "Why's cloud computing any different to hosting applications on dedicated managed services or virtual services hosted on shared hardware? None!"

        Ah, did you mean servers instead of services here ? Let's assume so for the sake of argument.

        If you're a US company then you're correct. There's not a lot of difference.

        For non-US companies however, there's a world of difference. Some of the differences are legislative. Just ask either Bodog or the poor buggers who had picked MegaUpload to host their data. Other differences are performance and uptime. If you host, you know where you data lives and can properly assess the risks of network congestion/outage. You can't do that in a pure cloud model because you may not know where the data is at any given moment.

  3. Anonymous Coward
    Anonymous Coward

    Potential blowback

    If HP goes in guns blazing with public "cloud" offerings, their IT infrastructure manager customers are not going to take very kindly to it. Amazon doesn't have to sell the guys they are trying to replace servers and other kit. Not to mention, the channel sells HP gear, not HP. How are they going to be cut into this picture? If the channel is out, there could be some defection.

    1. tillm
      Stop

      Funnily

      ..former HP Labs director Joel Birnbaum was theorizing about buying compute capacity on demand about ten years ago. It was just Amazon which made it reality.

      This is exactly the right strategy, even if they lose some Amazon business, as they would otherwise slowly disappear as a brand in the server domain.

      1. Anonymous Coward
        Anonymous Coward

        Re: Funnily

        I am not talking about Amazon themselves. I mean their rank and file customers that buy HP servers and other gear in their own data centers. Now HP is going to be trying to replace their own customers' jobs. This might be the wave of the future, but they are not going to appreciate it. They have to simultaneously push "cloud" while also pushing servers, storage, PCs, printers, etc.

        The channel could also be an issue as HP sales = channel sales. This probably won't work out as well for the channel.

  4. tillm
    Megaphone

    @HP: Provide Quality, Be Loyal To Own Stuff

    As the article has stated, it will be difficult to compete with Amazon ON PRICE. But haven't we read lots of stories about AWS being unreliable ?

    So HP, try to compete ON QUALITY: Assure competent people run these infrastructure operations, aim for high availability, resilience, proper technological architectures.

    Most of your customers will not be web 2.0 startups - they are already comfortable with Amazon. So it will be more "normal" companies who do not want to buy expensive servers just to see it rot at 5% CPU load in some forgotten corner of their datacenter.

    Finally, be Loyal To Your Own Creations. If success is initially not phenomenal, don't do the regular HP fuckup of closing down the own product/service and then support some competitor. You killed lots of High Value product lines this way (Allbase SQL, VMS, PA RISC, HPUX (almost)). Take the Larry E approach and drum up your stuff even if it currently does not fly. Loudly announce minor improvements to anybody who listens. A bit of hyperbole never damaged Oracle and Microsoft - but their mute competitors.

    1. Anonymous Coward
      Anonymous Coward

      Re: @HP: Provide Quality, Be Loyal To Own Stuff

      Agree, there is already no margin in this PaaS game. Amazon doesn't make a ton of margin pushing their cloud services. The cloud services companies that are making margin do so by attaching their own software. Cloud is a low cost and easy to manage way to get to more people using their software. Microsoft, Oracle and Salesforce are on this path. IBM is on that path as well but a little different as they sell Microsoft, Oracle, open stack as well as their own software in the SmartCloud.

      HP doesn't really have any software stack or applications to sell. If they are not careful, this could just be a way to make even lower margins on infrastructure while helping Microsoft and open stack sell software. If it is all about cost, HP is going to get hammered by Amazon and other yet to be name providers (probably the ISPs, ATT, Verizon, etc) that are going to use the Google data center model with white box x86/Hadoop. If they are going for software flexibility, security, data governance, quality of service, etc, they are steering directly into IBM's court.

  5. Mikel

    $1 per day

    That is about 4 cents an hour. I don't see how that's going to be a big profit center.

    1. Anonymous Coward
      Anonymous Coward

      Re: $1 per day

      It won't be, they would be better off selling servers. As someone mentioned above, a lot of those x86 servers are running at 5% CPU utilization (probably more like 25% in a virtualized environment). That really isn't a bad thing for HP though as they just get to sell a whole bunch of additional x86 servers. HP and Dell stand to lose the most from the cost efficiencies of the x86 "cloud" if every cloud service provider's data center looks like Google, Yahoo, fb, etc. It will primarily be a bunch of small shops with a couple of inefficient HP or Dell servers moving to a highly efficient cloud. This is more of a defensive play.

    2. tillm
      Stop

      Why ?

      It all depends on the ratio of HP's capex and the revenue generated. So if they spend 200 dollars on that service capex and 50$ on power, management, interconnect; make 365$ per year, that would already be an excellent investment.

      You can indeed built an x86 system with 1Gbyte of RAM and 1.6 GHz for 200$. The other costs are pulled out of thin air by myself and much depends on what kind of deals they can get from the telecoms which were mentioned (those who will host+connect). But I am sure some high-level deals between the CEOs of HP and Vodafone (for example) can produce much better prices than you and I could ever dream of.

      Also keep in mind they can ship the hardware in quantities of Full Containers; that immediately cuts out most of the distribution costs which are associated with the traditional business.

      1. Anonymous Coward
        Anonymous Coward

        Re: Why ?

        I am not saying that it will be impossible to make cash, just not as much as they made selling the low utilization x86 servers. HP is going to have to compete against some big competitors which are going to be running Google style data centers with 50-100 white box x86 blades in a rack. If they want to compete, they are going to have to use those servers very efficiently and not charge as much as they would have charged if some business had purchased a bunch of ProLiants and paid on a monthly basis. They may be able to leverage their buying power, but they are already doing that with x86 components. Also, instead of having two competitors, IBM and Dell, and really just Dell in most of the small businesses, they now will have dozens of competitors coming after this business which they use to own with servers (small business compute). I doubt the cloud margins are greater than their current x86 server margins. I am not saying that it isn't necessary in the long term, but it isn't a welcome change either.

        I doubt that Vodafone or any of the other ISPs will give anyone any preferable terms. This is a bad situation for the ISPs. Instead of charging a bunch of small businesses high rates because they have no purchasing power, now they are going to have to deal with huge companies with tons of purchasing power. I think, and it has already started with AT&T and CenturyLink, the ISPs want to be the cloud provider selling computer services as well as ISP services. They will probably eventually become a competitor.

        1. Matt Bryant Silver badge
          Trollface

          Re: Why ?

          Lol, always funny to see the IBM troll reflex bashing of anything hp! I bet Wunderburp would be singing the idea's praises long and loud if it was IBM doing it rather than hp.

          1. Anonymous Coward
            Anonymous Coward

            Re: Why ?

            IBM does have the same cloud service, SmartCloud. They have had it for awhile. It isn't the greatest idea, but at least IBM is getting business they probably otherwise would not have competed for with hardware (small shops, blog web servers, etc). IBM isn't going after any of their customers, or at least not many, because they are generally playing in the big shops that wouldn't touch public cloud. HP has gained a lot of good will with the techies as a result of not being thought of as an outsourcer (even though they are, they just aren't any good at it). It could cause more harm than good as HP is taking out HP servers and their customers with cloud. HP is also VAR centric. The VARs are going to lose in this big cloud push. Their SI services are going away with the cloudyness. Who knows how HP will handle cloud renewals and whatnot.

  6. tillm
    Stop

    @Wunderbar1, Cannibalization

    As they say "if you can't beat them, join them". HP and Dell know exactly that the IT industry is moving quite fast (as compared to most other industries).

    IBM tried to "protect" their mainframe and midrange business, only to see competitors such as HP, Compaq, Dell, Microsoft and Intel eat their business alive. IBM certainly had all the skills required to build a 100% PC and to eat their own mainframe business, but the beancounters in Armonk chose their short-term benefits instead of their long-term viability. Now IBM is a kind of "American Tata" with a dying hardware appendix.

    All the "PC" companies (from Intel, Microsoft, Seagate to Google and Facebook) are now pushing Computing Forward. IBM had all of these technologies in-house from the RAM chip and the harddisk up to the SQL database server. But they were incapable to make that stuff available to the little guy with few pennies. Oracle, Microsoft, Seagate, Intel and Google were happy to do so and now they own hugely profitable empires while the IBM hardware/systems business is in a process of steady decline. They are even too stupid to properly sell DB/2.

    1. tillm
      Megaphone

      IBM Systems And Technology Group - What Happened To Them

      If you think it is a great idea to "protect" some kind of legacy business, see here:

      http://www.familie-wischnewski.de/joerg/nxp-boeblingen.htm

      The pictures you see are from a once top-notch IBM semiconductor fab, which once cost 1 billion dollars. Chancellor Helmut Kohl opened the facility personally, more than 20 years ago.

      I didn't work for them but it hurts to see excavators tear it down.

    2. This post has been deleted by its author

    3. Anonymous Coward
      Anonymous Coward

      Re: @Wunderbar1, Cannibalization

      As IBM STG is the largest server provider in the world and the second largest storage provider in the world (by revenue), it doesn't seem as though they have suffered as you write. IBM is far and away the leader in Unix. They are second in x86 servers. Mainframe MIPS have actually grown in recent years.

      IBM is also the second largest software company in the world, or third if you ask Oracle. Regardless, they are much more than an "American Tata." They lead the world in patents every year, still win most of the major tech prizes awarded to private business every year, and are still at the cutting edge of IT. See Watson or Carbon Nano. IBM is by far the most inventive company in IT today if we are talking about actual computer science and electrical engineering innovations and not advertising media.

      Google may have a case for pushing computing forward with Map Reduce. I would like to hear your case for MS pushing computing forward. I cannot think of any record of innovation or invention at all commensurate with their size, cash and influence.

      Has IBM made mistake with relation to PCs? Yes, I don't think that can be argued, but it doesn't negate all of the great things they have done over the years and are still doing today.

      1. Matt Bryant Silver badge
        FAIL

        Re: Re: @Wunderbar1, Cannibalization

        <Yawn> IBM troll's mask has long since slipped.....

        ".....IBM STG is the largest server provider in the world...." Been over this before - hp is the leader in x64, not IBM, and that's where the growth is. Especially blades (hp the leader by a massive margin) and the trusty two-socket rack server (surprise, surprise, hp yet again!). And then we have hp's massive lead in Linux servers (you know, the stuff that's eating the UNIX market).

        I got very bored of the rest of your "IBM, rah-rah-rah" routine, but then I just need one link to expose the shallowness of your bluster - a list of companies by revenue. In at number 29, the highest placed IT company in the list, is good ol' hp! Waaaaay down at number 56 (below Apple and Hitachi!), ickle IBM. http://en.wikipedia.org/wiki/List_of_companies_by_revenue

        /SP&L

        1. Anonymous Coward
          Anonymous Coward

          Re: @Wunderbar1, Cannibalization

          "Been over this before - hp is the leader in x64, not IBM, and that's where the growth is"

          In other words, I am right. IBM is the world's largest server producer by revenue. As IBM sells way more z and Unix than HP, IBM is far and away the leader in profitable systems. HP is the leader in x64 servers. They are also the leader in PCs. They don't make two cents pushing that commodity gear.

          "I got very bored of the rest of your "IBM, rah-rah-rah" routine, but then I just need one link to expose the shallowness of your bluster - a list of companies by revenue. In at number 29, the highest placed IT company in the list, is good ol' hp! Waaaaay down at number 56 (below Apple and Hitachi!), ickle IBM. http://en.wikipedia.org/wiki/List_of_companies_by_revenue"

          Yes, unfortunately companies are in business to make profit, not revenue. That is why IBM is worth (market cap) five times as much as HP. Apple could buy HP with spare change. Oracle doesn't make even a third of HP's revenue, but they could buy HP tomorrow.

          1. Matt Bryant Silver badge
            Facepalm

            Re: Re: @Wunderbar1, Cannibalization

            ".....In other words, I am right....." LOL, ask the voices in your head, they might agree with you! IBM tried hard and failed in the PC market, a market they could reasonably claim to have profitised if not invented, and is failling as we speak in the x64 serevr space (every quarter seems to bring mroe lines being hived off to Lenovo). Simply yapping on about the shrinking Z and AIX market is simply being an ostrich, both are shrinking and have been shrinking for years, and IBM does not have an escape strategy now that it has sold off most of the x64 lines.

            IBM's clound offering is the technical definition of unadventurous and unsuccessful. IBM was once the largest and most important IT company in the World, but those are long since gone and very unlikely to ever return. Just get over it.

This topic is closed for new posts.

Other stories you might like