10 things you still have to worry about until cloud computing becomes reality

10 things you still have to worry about until cloud computing becomes reality

Summary: Helpstream's Bob Warfield unveiled 10 things that don't matter anymore given the adoption of cloud computing. In theory, Warfield is right.

SHARE:

Helpstream's Bob Warfield unveiled 10 things that don't matter anymore given the adoption of cloud computing. In theory, Warfield is right. The key word: Theory. Given that only 3 percent of enterprises have adopted cloud computing the rest of you need to worry. 

Warfield riffs that you just don't need to worry about the following things in the cloud:

  • De-duping and backup;
  • Server power consumption;
  • Little iron vs. big iron;
  • MIPs;
  • Bandwidth costs;
  • Load balancing;
  • Hardware monitoring;
  • Creating redundant data center;
  • Configuring complex software;
  • Engineering time spent on keeping the lights on. 

For Warfield, these worries don't exist anymore since his company, Helpstream, runs on Amazon Web Services (here's the blueprint). 

Here's the reality for the rest of you:

Only 5 percent of enterprises have implemented cloud computing and 3 percent plan to use it in the next 12 months. Assuming that 8 percent of those enterprises move their entire infrastructure the cloud they can live a worry free life like Warfield. 

The point: Let's not get too carried away with cloud computing hype. There are the first movers and then the reality the rest of you face. Simply put, a larger percentage of enterprises won't have to worry about mundane tasks in the future, but it's not likely to approach 50 percent in the next 10 years. Why? It's hard to move the petabytes over to the cloud. Amazon is still using a fancy means of Sneakernet to move enterprises to the cloud. Now that's reality.

Topics: Cloud, Hardware, Virtualization

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

12 comments
Log in or register to join the discussion
  • Someone is going to have to worry about those things

    Maybe I don't get cloud computing, but ultimately you applications are running in a process somewhere, and communicating with a process somewhere else. If that's the case, all of these things still apply, it's just a matter of who is responsible. Just because someone else is hosting your app doesn't mean you're not responsible for it's availability...your customers are going to come to you.
    bmonster
  • Sure, but, when the bank handles your money for you, you do not need

    vaults, armed guards, armored trucks, etc. That is all off-loaded to the bank. Yes, your customer WILL call you if a bank screw-up results in problems for your customer, but, that does not mean you need to be involved in how the bank is run, or that YOU could do it better or more reliable or cheaper having your own vaults and armed guards.
    DonnieBoy
  • RE: 10 things you still have to worry about until cloud computing becomes reality

    Wow this article is misleading.

    1. No Backups? If your vendor goes out of business or spins off it's data centers to another company that uses a different format, you better hope you have data back ups of your own. Most vendors have been clear that if you have data you need to keep secure, you need to maintain yourself, which includes backing said data up.

    2. Server Power Consumption? Hmmm. Somebody pays for that. Let's see the standard free market model is to pass the costs on to the consumer plus a hefty profit. I think consumers will still be paying for the power a server uses. It may not be itemised in their bill but they will still pay for it.

    3. Little Iron vs. Big Iron??? As far as I can tell it is the difference between buying a small home server vs. a corporate server. Not sure if virtualization would make this a moot point or not.

    4. MIPs??? Millions of Instructions per second? If you off load your own data center you are still going to be accessing it. Your ISP had better be ready to handle the increased load or your business may be shut down.

    5. Bandwidth costs - They might go up. If every single terminal in your business now need direct access to the internet and they still work in an office, Your internet connection had better be able to handle the daily load. Your ISP is going to have a field day with this price increase.

    6. Load Balancing. My guess is you will still pay for this service. Nothing is done for free. Again it might not be itemised in your bill but you will still pay for this service.

    7. Hardware Monitering. First your vendor will monitor the hardware, you will just pay them to do it. Also you will need to hire someone to manage the vendor relationship. If your vendor goes offline, then your business had better have a backup plan and a single point of contact to get tech support.So instead of paying a relatively cheap IT guy to watch the boxes, you hire an expensive MBA to watch the vendor.

    8. Creating a redundant data center. Well, you will need redundant cloud vendors just in case yours goes out of business, gets taken over and you don't like the new owners, or just has long down times. Otherwise you will need to maintain your own data center as a backup plan for if/when your vendor goes out of business.

    9. Configuring Complex Software. Well most businesses do this already. They hire a company to configure their system. The difference is they own the system and have a real business contract with the company they hire.

    10. Engineering time spent keeping the lights on. You will still pay your vendor for data center usage. No vendor is going to give it's customers a free ride.

    If the could ever wakes up and realizes that it will need to charge for services and utilization, businesses are going to realize they aren't going to spend less money. They will spend the same money, only on different things or hidden things.

    Outsourcesing manufacturing saves money because off shore countries have barbaric labor laws that allow economic slavery and have societies with large strong family structures that support small incomes from many working members of the family unit.
    Outsourcing data centers will only save you the cost of the IT worker. You will still pay for the electricity used. You will still pay for the amount of data resources your use.
    mr1972
    • The point is, just like with the banks, all of those details are off-loaded

      to a reputable provider. Just like you pick a reputable bank, you pick a reputable provider for cloud services. Sure, just like you would switch banks if they screw things up, you will change cloud providers.

      And, sure, you SHOULD demand the ability to backup your data from the cloud in a format that could be transferred to other providers, and you might even pay another trusted third party to do that.

      Do you get involved in how your bank constructs vaults, the hiring of armed guards, the model they use for armored trucks?
      DonnieBoy
      • I see the point but...

        Banks are heavily regulated (more so coming up) and ultimately backed by governments.
        Cloud Vendors, not so much. Basically I don't like bailing out banks when they miss-manage their businesses. I won't like bailing out cloud vendor companies because they are "to big to fail."

        Realistically cloud computing depends heavily on concepts that lead to socialism. At the same time, socialism will not provide the kind of innovation necessary for technology to continue to grow.

        Banking is socialized and will probably become increasingly socialized. However, the business of banking doesn't really require it to change models every two years because the "next biggest thing" in banking only occurs every century or so.

        Businesses rely on a stable infrastructure in order to run. Cloud computing can't provide that stability with out some type of bail out plan.

        All cloud computing is doing is shifting around where the money is spent. You will end up firing a bunch of relatively cheap IT people and hire a bunch of relatively expensive MBAs to make sure the business plan doesn't fail due to faulty vendors.

        mr1972
        • No, there will be heavy competition on cloud computing, no central control.

          And, companies already have to look for trusted suppliers for many other parts of their business, cloud computing will be no different.

          Finally IT people are NOT cheap. At least good ones that can keep it secure and running 24x7. You would not need any more oversight for this than for any other supplier that you use.
          DonnieBoy
          • Not quite there yet...

            Trusted Companies basically translates into "Companies to big to fail." Sure the best will shake out of the competition but if most companies end up with Google for example. Google failing would bring a sizable chunk of the worlds economy to it's knees.
            Sure companies use power from the local grid rather than support a power plant of their own but those power companies are heavily regulated by the government and they tend to represent local monopolies. Only in a few areas do companies have a choice in different power companies.

            Also I would put a MBAs salary up against an IT persons salary any day of the week. Even the best IT people typically make far less than an MBA.
            Plus once you off shore the data centers to India or China, you will have whatever they decide is secure or competent personal.
            mr1972
  • Network reliability

    Yeah, they still have to worry about that stuff . . .

    "De-duping and backup" - yeah, you still have to worry about backup. What happens if your provider goes down, loses data, or worse goes out of business?

    A backup plan that is "our cloud provider is our only backup" is no backup plan at all.

    "Server power consumption"

    Okay, you don't have to worry about it directly. But how efficient your provider is may still be a factor when deciding who to go with. After all, if they waste power, it gets put on your bill.

    "MIPs" and "Bandwidth costs"

    LOL - this is a joke, right? You're still paying for it! You're just paying somebody else for it rather than paying for it yourself. There are still limits to how much you and your customers can use.

    "Little iron vs. big iron" (and pretty much everything else)

    Okaay - if you require a lot of servers just to run your business, then yeah maybe a "cloud provider" may be right for you. But - if you're like most businesses and don't have any particular need for lots of servers, then I wouldn't go cloud.

    And oh, yeah - one thing extra that they need to worry about is network reliability. The Internet hasn't proven to be reliable. Outages are far less predictable and controllable.
    CobraA1
    • All the risks are greater with your own internal staff, and consumer

      facing applications that bring you down when you lose your network connection.

      And, Mips and Bandwidth costs are much cheaper with a provider doing it all, AND, you can scale your usage up and down at will. With on premises equipment, scaling up for short term demand is not even possible.
      DonnieBoy
  • Not so fast

    * De-duping and backup;
    The cloud supplier (CS) still has to pay for backup software, disk and tape resources and personnel - so those costs will be passed on to the consumer. Tapes fail. Backup schedules have holes. When the data is lost, do you feel better that it's the CS instead of inhouse?

    * Server power consumption;
    You will pay for this, but there is an incentive for the CS to lower their costs (without necessarily lowering yours).

    * Little iron vs. big iron;
    This is an issue! Most clouds are x86 servers in a grid - which are great at horizontal scaling. In order to do vertical scaling, you need very large (UNIX) servers - which will have to cost more. So you still need to architect whether to go little or big iron for each project.

    * MIPs;
    The CS can use Moore's Law to keep lowering their MIPS costs. Will they pass it on?

    * Bandwidth costs;
    Transferring large datasets across the internet will still give companies headaches.

    * Load balancing;
    You will pay one way or another.

    * Hardware monitoring;
    This will be replaced with software monitoring - as you will want up-to-date info on your running jobs.

    * Creating redundant data center;
    What level of redundancy are we talking here? SLA1 requires transaction-level redundancy - which will cost.

    * Configuring complex software;
    Yeah right. After all my years of experience watching Db programmers make tiny mistakes that lead to very slow performance, I don't think anyone that isn't intimately involved in your application can tune it properly.

    * Engineering time spent on keeping the lights on.
    True. But I "survived" the 2003 blackout, and if your datacenter was in Detroit and your business was in LA - you would be sht out of luck.

    Clouds make sense for SMBs since they don't have a lot of data, and they can buy just the capacity they need. As you scale up, more intractable issues start arising to the point where you just say "forget it".
    Roger Ramjet
  • A cloud by any other name

    Is an IT department. Just an unhelpful, usually contracted or outsourced and desperate to maintain the illusion of user/customer service, while meeting as few real needs as possible. They just went private and took it down the road. Probably a bit cynical, as I know there are a few helpful IT sections out there, but it reflects a lot of people's experience ;-)

    Now instead of having some vestige of control through salaries, contracts, KPIs etc etc, you can now compete with thousands of other companies for their attention.

    When your services go down, you can resign yourself to passively accepting your business will be in stasis for an indeterminate amount of time.

    That's not to say you shouldn't use web services, just be aware of the risks, see what compensation they offer for not meeting KPIs (or whatever metric you want to use to monitor their performance), look at business insurance, backups, local storage, data import/export and what the real cost benefits are.

    You also need to realise that you are trying to pick winners here and tieing your business with theirs. If they fail, then your current business systems (which may range from simple email/WP to accounting/payroll/crm) will also fail with perhaps catastrophic consequences.

    When there are a plethora of proprietary and open source methods to provide internal IT services (that you have direct control over), you should really consider whether relying on an Internet connection for all business IT services is truly cost effective.

    tonymcs1
  • RE: 10 things you still have to worry about until cloud computing becomes reality

    Did you ever lay on your back and watch the clouds go by, picking out shapes?
    So do we all see "Cloud" computing, it's Apps, Services, and potential differently.

    Cloud Apps may works for different size companies,for different reasons; for different tiers of service in the same company for different reasons, even different Apps within the same company for different reasons, and for some, not at all.

    One thing for sure, Cloud offers solutions to many problems affecting cost and availablility in Education, Health, and Commuting, with more coming each month.
    Welcome The Cloud.
    Hail Cloud!
    BaltimoreBarry