Why cloud computing may cost more than on-premise systems

Why cloud computing may cost more than on-premise systems

Summary: Panelists discuss how clouds are extensions of enterprises, and require the same levels of planning, architecture, integration, security, accountability, and governance that go into managing on-premise applications and systems.


The cloud and SOA are creating a lot of shifts in the market. As we've said here, the lines between service providers and consumers are blurring, and this has implications for pure software vendors.

One industry shift that is becoming readily apparent -- and this was just pointed out by Sandy Rogers at a recent BriefingsDirect podcast hosted by ZDNet colleague Dana Gardner -- is that vendors are assuming the roles of service providers. "Many technology providers have to grow into the role of a service provider," she says. This is requiring a much tighter relationship with customers, versus simply shipping CDs and cashing the check.

As Sandy points out, "cloud providers [and this means most vendors now], because they are in the business of providing a service, are starting to become much more transparent regarding the usage in order to help their customers make decisions and plan for the future."

I have spoken with a number of CIOs and IT executives who have sourced out some of their applications or systems to cloud providers, and a common thread among many is to immerse themselves within the cloud providers' operations, from participation on customer boards to frequent site visits. Enterprise managers -- at least on the IT side -- are not going to simply hand over a credit card and hope for the best.

Sandy elaborated on this theme, stating that cloud service providers are under the gun to show a great deal of accountability, something which has not been enforced as strictly with on-premise enterprise IT shops. Thanks to the potential competition of cloud, this may change.  "The ability to correlate the computing that's being used with what the value is to the business may actually move forward with cloud," Sandy says. "And put a lot of pressure on those that are providing computing resources on-premise to provide the same kind of metrics."

The bottom line is that clouds are becoming extensions of enterprises, and therefore require high levels of planning, architecture, integration, security, accountability, and governance -- the same, or even more than go into managing on-premise applications and systems.

Enterprises shouldn't be fooled for one minute thinking that outside cloud services offer dirt-cheap alternatives to on-premise IT. Dave Linthicum, who also participated in Dana's latest analyst roundtable, points out that there's a lot more to enterprise IT than simply accessing and running applications. "Cloud computing typically is going to be a better, more strategic, more agile architecture, but it's also typically going to be more expensive, at least on the outcome," he remarks.

Cloud computing more expensive than on-premise IT?  How can that be?

Dave says the expenses arise because cloud requires lots of changes. "You're going to have to redo your infrastructure, as I write in my book, to leverage newer architectural patterns, such as SOA, and that's typically very expensive to get out and access the services that are available to you on demand, out of the cloud." Add to this the costs involved in retraining staff and re-aligning the business.

"Ultimately, you get to a much better, higher value strategic architecture which is going to add more value to the business," Dave says. "But it's going to cost you some additional dollars to get there."

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.


Log in or register to join the discussion
  • Imagine that....

    Yip more money and more hacking...
    • nt

      The one and only, Cylon Centurion
  • RE: Why cloud computing may cost more than on-premise systems

    While it's nice to think of enterprises involving themselves with cloud companies, the reality is that your current in-house IT department just supports your company, while a cloud service may support thousands - good luck at getting support.

    There's only one reason for cloud services - the cloud vendors making money ;-)
    • echo that <nt>

  • Oh Donnieboy

    I told you so...
    The one and only, Cylon Centurion
    • LOL - nt

  • It says it will be more expensive...

    but it also says that it is a one time ocurrence: converting to SOA, training. Frankly, if the cloud were to be more expensive than on premises, its value proposition would be not that attractive.
    Roque Mocan
  • RE: Why cloud computing may cost more than on-premise systems

    Most outsourced ventures have resulted in higher pricing. You see a short term dip in pricing when the outsource companies compete for your business. Once the culture of cloud computing is created, the vendors have a lock in the market and can gradually raise their prices to provide more value for "their shareholders". Your shareholders will have to fend for themselves.

    Ultimately the price of cloud computing is the same as most other pricing. The answer is "What ever the market will bear." This will not directly be related to anything other than maximizing profits for cloud company shareholders. Will some cloud vendors offer "better pricing"? Sure, if there is enough competition. But if the cloud market starts to settle down to 3-15 major players, you will see a gradual price creep up.

    My home town is a great example of outsourcing gone wrong for consumers. We had a shoe factory. It employed U.S. citizens and they had a reasonably good union. The factory had 50s and 60s level technology in the 70s. The local shoe stores sold their shoes at 20 USD which is about 50 USD in today's money. The same shoes are 95-120 USD made over seas today. The local factory was shut down, 300 jobs were lost from the factory itself plus the support jobs that went away. How come a factory in the U.S. with union labor (expensive labor) with outdated technology could make shoes for 50 USDs (today's money)? But an over seas outsourced company with cheap labor makes the same shoes for 100 USD or twice the cost?

    The answer is, consumers don't pay fair prices because they refuse to go with out certain kinds of shoes. They actually pay inflated prices because that is the price "the market will bear".

    It will be the same with cloud computing. You will get a deal at first but it won't last.
  • RE: Why cloud computing may cost more than on-premise systems

    You're right, but that extra expense should be for initiating the services and using them in the most fully exploitative, optimized way (e.g., changing to SOA architectures). Once this is done, there is no reason why the cloud, offering better economies of scale and higher percentages of usage should not be cheaper.

    In both cases (start-up and more mature usage), there are still measurable benefits in Time to Usage/Market, agility, and lowered cost of Capital Expenditures and operational costs.
  • RE: Why cloud computing may cost more than on-premise systems

    Joe, you raise some good issues about getting into the cloud. We believe that you shouldn't have to change anything about your app or how you manage it when you move to the cloud -- you shouldn't need to re-do your architecture at all. Our customers are taking existing VMware-based Windows and Linux apps and moving them to EC2 right now with 1 click.
    Regards, Ellen Rubin, Founder, www.cloudswitch.com
  • Cloud services AREN'T good for small businesses

    Most small businesses that don't have a downtown highrise office (and let's be honest - that's most of them) won't have a super-fast internet connection with a service level agreement because of one of two reasons: a) the ISP can't offer high reliability in the area (cough, DSL, cough), or b) it's way too expensive to pay for an SLA. If they have a need for business-critical data storage and don't have a bottomless wallet, they're better suited to an on-site IT infrastructure.
  • RE: Why cloud computing may cost more than on-premise systems

    Finally someone actually understands the financial implications and has the #$%% to stand up and say it.
  • Cloud services ARE good for small businesses

    To extend the comments made by Amy Wohl, the current servers / applications do not necessarily need to be re-architechted to run in EC2 or AppLogic.

    We have several clients that we've basically "forklifeted" their infrastructure into the cloud and the end result was pretty seamless. They are happy with the end results and we were happy to provide the migration / implementation service.

    And outsourcing to the cloud doesn't mean outsourcing to India or China. We run all of our clients in the USA. In both Applogic and EC2 environments. Granted, we provide Infrastructure as a Service and not SaaS but, the distinction needs to be made here.

    For small business, startups, etc, the cloud is a perfect model, particularly if you're using Linux / Unix based server infrastructure. Cloud basically levels the playing field for these companies. They can now afford to play with the big boys without having to fork out precious capital resources. If they are just starting, cloud is even a better idea because the SOA architecture can be designed into the infrastructure from the very start.

    Cloud implementation is the wave of the future, particularly in the IaaS scheme of things. People can either adapt and benefit from the change or stay rigid and inflexible and be left behind.

    Mike - CEO
    • cloud versus web hosting

      I see how cloud is most probably cheaper for SMB when compared to self hosting but how can cloud be cheaper when compared to using a WHSP? For example, Networks Solutions offers a 300GB disk space and 3000GB bandwidth package for $10 per month. Plug that into the EC2 billing calculator and you get $15 per month.
      • in this instance you are correct.

        If it's just a web site that is being hosted, then you are correct, there are even cheaper alternatives the EC2 or Applogic, no doubt there.

        But, when you're talking about a small business that has multiple server based applications, databases, etc. that need to run, the Cloud based solutions start to look more attractive and feasible. Plus, it allows the in house staff to focus on the core-competencies of the business.
      • Chalk and cheese

        No, a site capable of being hosted for $10 a month (definitely at that price shared by a LOT of other users) would not be a likely candidate for the cloud, or at least with any real benefit.

        Although not exclusively, the two scenarios where the cloud may be an ideal choice (both to do with scalability) are...

        1. Sites with periods of low useage but spikes of extremely high usage. For example an event ticketing site.

        2. Start-up enterprises where growth cannot be predicted but could be extremely rapid, and infrastructure demands may rapidly increase. For example, amongst many, a SaaS start-up enterprise. Crudely speaking, whether a success or not the costs scale to income - a much lower initial risk.

        Serious business aplications requiring a high level of dedicated infrastructure, but where the actual demands are nowhere near 24/7 or are not easy to predict tend to suggest a cloud alternative.
        David S M
  • To cloud or not to cloud . . . .

    I agree with most of the assumptions in the article and many of the comments. It does feel a bit recalcitrant though by focusing on the short-term, conversion side. Has anybody measured the lost opportunity side of the equation when in practice, most IT depts traditionally short change the architecture, coordination, alignment, documentation and training efforts due to the resources and focus necessary in managing day-to-day hardware, software and user problems? While you are not saving costs it feels like you are shifting costs to a more efficient platform? Seems to me moving to the cloud reinforces placing IT resources where their SME can do an enterprise the most good.
  • The cost benefit depends on the client

    We are a small solution provider in the West Palm area. We develop custom SOA based setups for our clients, rent them the equipment and charge for the service.

    Being in IT for the last 15 years, one of the biggest problems we see is that small business clients sit on their IT equipment for way too long, and it becomes either outdated or fails before they start replacing it. Regardless of how much we press the issue, many of these small businesses don't phase their purchasing so it's a $20k+ bite out of the bank account when they decide to replace equipment.

    Being a small service provider we can give them the support they require, and because this is our main business we have redundancy in place that they otherwise couldn't afford. We also replace all of equipment every 3 years and keep spares on hand.

    The need to uproot their existing architecture and replace it isn't exactly true either. Unix/Linux architectures are easier than Windows ones, but you can work wonders with RDP. True, Internet connections can cause issues, but hybrid SOA setups can be implemented where low cost Atom based on-premises data caching servers can be used.

    My point is that in some cases it is cost prohibitive to implement an SOA, but in many cases for small business it can be very beneficial.

    One thing to keep in mind is what happens when that on-premises server goes down and you have to migrate data and software from a Windows Server 2000 network with XP machines to a Windows Server 2008 network with Windows 7 machines in a pinch (such as a break-fix scenario). Not only do they have to replace hardware but update software, and account for the downtime. This can be VERY costly to the client.

    I think with a large cloud provider that mass produces solutions, you're asking for trouble. With small local service providers, you don't get the AT&T / Comcast customer service, you get real customer service, from someone that knows the clients needs, and cares about the clients satisfaction.

    • RE: The cost benefit depends on the client

      Brian, I agree with you. You hit the nail on the head here. We see the same thing with our SMB practice...the customers will run the iron into the ground, let the service contracts expire, you can't get parts except maybe on EBAY, and they want the operation back up and running immediately, understandably.

      From a SMB perspective, why would a biz owner want the hassles and headaches with owning a bunch of assets that are worthless from a balance sheet perspective anyway? Potential buyers of businesses are interested in the data, not the computer hardware and software the business declares on its balance sheet.

      IT isn't happy about all this because they perceive it as risk to their job security. We see that all the time.

      Our message to the CxO is, save your cash, eliminate the headaches and invest it in appreciable assets like their core competencies,business processes and intellectual property and rather than computer equip and software that is worth .10 or less on the dollar. CxO's eventually will see the light if they are P/L and market share driven.

      I also agree with you about picking a hosting partner that is local to your geography (not a national commodity type provider), that has a solid business model, perhaps targets or services a specific vertical or application suite rather trying to be all things to all people regardless of industry or size of customer.

      From my perspective (yes, we are a SaaS provider), it is good value for our SMB customers, they see it and are much happier because they can get projects turnaround much quicker, its less expensive, no huge upfront cash hits, no hotfixes, patches, or refreshing of servers/OS/appplications every three years or so, they don't have to worry about whether or not their backup jobs are getting done, nor have to worry about disaster recovery plans for their IT infrastructure. It is all generally good news.

      One other point, the Enterprise customers love it too. Our enterprise customers line of business owners have shared that their internal IT charge backs are $10k to $15k per server annually to have IT infrastructure partners maintain their application servers. When the LOB owner realizes they can rid themselves of those annual charge backs,they are generally sold.

      All in all, we see this as a long-term, sustainable growing trend in the IT world for many years to come. Embrace it, don't run from it. This biz model has legs and will continue to expand, and SMB and Enterprise markets appear to like the results so far.
  • RE: Why cloud computing may cost more than on-premise systems

    Joe, great post. I think Dave's point is a good one about cloud being a more agile, better foundation for a business than traditional IT. The headline is attention-grabbing but I think the point is more around transition cost rather than run-rate TCO. At a steady state, a cloud application should be significantly cheaper to run/maintain/enhance than its on-premise counterpart. Of course, taking full advantage of cloud apps require migration, integration and change management along the way. In a steady-state, a cloudsourced IT architecture will enable significant TCO savings, better user effectiveness and greater business agility.