Is Cloud Computing Really Cheaper?
Source: Reuven Cohen
Is Cloud Computing really cheaper?” Of all the questions asked at a recent event, this particular one was the most difficult to answer. Earlier this week I was at the New Jersey Institute of Technology as the host of CloudCamp, a “un-conference” created to inform and educate on various cloud computing topics. Since the launch of the unconventional series of events more than four years ago, it has grown to more than 300 cities around the globe. Over that time a lot of things have changed in the tech world. The discussion has shifted from a question of what is or isn’t cloud computing, to one of what can be done with cloud computing?
I set out to attempt to get an answer. Many of the people I spoke to have described cloud costing as a “Dark Art”, a surprise at the end of the month. The most common response was “it depends.” Not satisfied, I continue on with my mission. I set forth asking as many people as I could “Is Cloud Computing really cheaper?”
During my quest, one particular analogy was often repeated, the “Rental Car Analogy“. Microsoft‘s Alan Merrihew, Senior Director of Government Technology Strategy explains;
If you have guests coming into town to visit, do you buy a car for them to use while they are in town for two weeks? And then does that car sit idle until the next time they come to visit? No, of course not! Logically, you would rent a car for them to use while visiting. Cloud services are effectively like a renting a car service, only for ICT. Let’s look at the five characteristics of a cloud service again using the rental car analogy:
On-demand self-service (it is easy to rent a car, you can book a reservation by phone or online)
Broad network access (there is a broad network of rental car agencies around the world to give you access to a car rental.)
Resource pooling (The rental car companies manage a pool of cars in any given city to meet demand. You don’t have to worry about it. If one agency is out of cars they will often refer you to another to help you find a car.)
Rapid elasticity (Rental car companies move cars into a particular location when there is a large event and they know demand will be high. They scale up and down to meet the demand.)
Measured service (You pay only for the time you used the car. Once you turn it back in you are done. No maintenance, insurance, fuel, tires, etc.)
Continuing my hunt for an answer, I posed the question to Joe Weinman, a well-regarded cloud computing thought leader and the author of the forthcoming book, Cloudonomics: The Business Value of Cloud Computing. He believes that most people use simplistic models to evaluate the economics of the cloud. “The standard argument for the cloud is that large providers achieve large economies of scale, and thus will be cheaper than a ‘do-it-yourself’ approach to IT,” he says. However, he argues, this is neither a necessary nor sufficient condition for cloud computing to be of value to companies. After all, people rent cars all the time, at a unit cost per day much higher than that of owning. Similarly, Weinman argues that the true cost reduction value of cloud infrastructure has nothing to do with lower unit cost, but with a no commit, pay-per-use model. “In effect, it doesn’t matter that much what you pay when you use cloud services, the key cost reduction driver is what you pay when you don’t use them: zero.”
Jason Evans, Co-founder and CEO of New York City based Stackpop, a B2B marketplace that connects service providers and buyers of Internet infrastructure, doesn’t think Cloud is a particularly good model for savings. Evans says that although “cloud is rapidly evolving (dedicated ‘cloud’ servers, guaranteed IOPS, cheaper pricing, etc.) for the most part - today - cloud economics across any cloud provider do not make sense for larger scale applications. For companies with any type of scale whatsoever (around $15 �C $20K per month from a consistent baseline traffic load) it is significantly cheaper to move the baseline load to your own, collocated environment. Even in places like NYC where power can run as high as $500/kW the math usually works out to show a CapEx payback in 4 �C 12 months, and the savings over a 3 year period are in the hundreds of percentage points. Most colocation customers who tweak and run their own servers (with no worries of multi-tenancy and hypervisor contention) also see significant improvements in performance and consistency.” (Evans has created an Amazon Cloud vs Colocation spreadsheet to help illustrate his point.)
On the other side of the argument, Weinman points out that the cloud is just as much about or more than simply cost reduction, or even “business agility.” He suggests “the really interesting use cases for cloud are strategic, even existential.” In a recent post here in Forbes, Weinman outlined how cloud can be used to support four major strategies: operational excellence, customer intimacy, product leadership, and accelerated innovation. Cloud and mobility, e.g., can be used to optimize field support and logistics; big data aggregation and deep analytics leveraging the cloud can support a customer intimacy strategy, as Amazon or Netflix seemingly know more about your preferences than you yourself do; cloud services can differentiate products, such as when Nike+ differentiates an athletic brand via cloud-enabled data collection and social networking; and cloud-based contests, such as the Netflix prize, can support an open innovation strategy, leveraging insights well beyond the limits of internal R&D.
The true value of cloud computing may very well go beyond the specific dollars and cents that are actually spent, but instead it seems to hit at the heart of innovation. And moreover, this potentially un-quantifiable value that innovation gives a company or organization. The ability to do things that were not possible before is a primary driver. When looking at the question from an empowerment point of view, the value proposition quickly changes from how much does it cost, to what value does it create?
Not completely convinced, I pushed Weinman further, “There is a theory that ‘IT Doesn’t Matter.’ But if so, explain Google. Starting with no brand, no capital, no preferential access to resources, and no preferential access to capital, Google leveraged algorithmic superiority, embodied in IT, as a cloud service, to generate a two hundred billion dollar market cap, with $25 billion in profit on $40 billion in revenue.”
Unfortunately innovation and opportunity are very difficult to quantify. The dollars spent on the other hand are much more easily explained. Recently an emerging group of upstarts are focusing on helping companies gain greater control of the costs associated with cloud computing. These companies have sprung up in an attempt to answer the question of cost, not just from what did it cost last month or last year, but what will it cost in the future. Companies with names like Cloudability, UptimeCloud, CloudCruiser and Cloudyn are leading the charge into an uncharted area of Cloud economics. The space is quickly becoming among the hottest in the cloud marketplace.
I reached out to Portland, Oregon based Cloudability to get the company’s opinion on the matter. The start-up has just landed $8.7 million in Series A funding from an all star list of investors including The Foundry Group as well as contributions from 500 Startups, Trinity Ventures and ad firm Wieden + Kennedy. It has created a platform that enables cloud customers to create simple daily reports and budget alerts that ensure that they are never surprised by overages and spikes. Cloudability now claims to manage $140 million worth of cloud service for 3,000+ companies in 80 countries.
Mat Ellis, Founder & CEO of Cloudability, says people who think cloud is more expensive will ultimately lose their job, in part because they will be left behind. He tells me that companies must, Adapt or die.” He concedes that his customers are “Paying a premium for new stuff, there are some additional costs of being an early adopter. But he says “the opportunity outweights the drawbacks.”
Sharon Wagner, Founder and Chief Executive of an Israeli startup Cloudyn, focuses on cost management that analyzes customers cloud usage and spending, and recommends how to optimize it. He says for a smaller web company, “the cloud is not only cheaper, it’s really the only option.” As companies grow, the question of spikes and general usage becomes extremely difficult to predict. He says you basically have two options, use you’re own data center, or someone else’s. Cloudyn attempts to dynamically allocate cloud resources via what he describes as an “optimal cloud configuration”.
I asked what kind of customers are using the Cloudyn service, Wagner told me his average customer spend is now spending upwards of $150,000 a year on cloud services with their largest customer spending more than $8 million / year. Cloudyn closed its seed round of funding last fall for $1.5m and expects to announce a “significant round” in the near future.
So is cloud computing really cheaper? The answer really does come down to how closely you are able to manage, track and adjust your infrastructure. If you’re taking a blind approach to just throwing your apps in the cloud without any kind of tracking or accountability than the answer is mostly likely no. If you take the time to clearly analyze your IT objectives, do the comparisons, the cloud is certainly cheaper.
Interest in Cloud Computing Has Peaked Reuven Cohen Reuven Cohen Contributor
13 Steps For Raising a Fortune in Start-up Funding Reuven Cohen Reuven Cohen Contributor
In short, Weinman says, “the cloud can help reduce costs, but more importantly, can be a core element of strategy, and occasionally, be the cornerstone of competitive superiority and market dominance.”
| }
|