Busting five misleading myths about cloud
- Summary:
- Cloud computing is prone to false analogies and misconceptions because so much of what it does is alien to conventional thinking.
As with any innovation, the problems come when people try to fit the new thing into their existing worldview. They make false analogies between the old and the new. Cloud has been especially susceptible to these misconceptions because it bundles up several separate and important new things all at once:
- Delivery of IT as a service rather than a set of products. When the accumulated wealth of the established industry is founded on a product sales mentality, that's a huge adjustment to make, and few have even got started.
- Computing as a shared utility. Instead of everyone buying their own IT products, we all have to get our heads around the idea of on-demand access to a pooled, shared resource. It's not an easy shift to make.
- Everything connected. Instead of every organization of a certain size having its own separate network, everything is on the single, shared network of the Internet, available instantly and 24/7. We have barely begun to scratch the surface of how that will transform the world we all live in.
Each of these on their own entails a big change of mindset. When all three of them come at the same time, no wonder people can't keep up. Below are five of the biggest myths about cloud computing that have emerged as a result.
Myth #1 - Cloud is just a deployment option
Many IT people make the mistake of assuming the cloud is just another place to put their existing applications and servers. They don't realize that it provides an entirely new way to do computing, one that emphasises connectivity, elasticity and resource pooling.
A better way to look at this is to say that there are two forms of computing — one designed for the cloud and the other designed for a non-cloud, on-premise world. You can run on-premise computing in the cloud or you can run cloud computing on-premise, but neither performs optimally when they are out of their native environment.
While there's often merit in 'forklifting' existing IT assets away from standalone servers into a shared cloud infrastructure, this relocation exercise fails to make the most of the cloud ethos. We may access it in the cloud, but on the inside it behaves exactly the same as it did when it was on-premise.
Putting it in the cloud doesn't automagically make it cloud computing. A resource that hasn't been thoroughly rearchitected for the cloud isn't taking full advantage of the model. Conversely, once it has been changed to adapt it for the cloud, then it's no longer a simple matter to deploy it back on-premise.
Myth #2 - There is no security in the cloud
We all know that the Internet is a wide open space where rogues and miscreants roam at will. It's therefore easy to make the conceptual leap to assume that cloud computing, which runs in the Internet, is equally exposed. But this is a misconception. In our minds, we are confusing the wide open spaces of the Internet with the narrow, closely guarded confines of a cloud service.
While the access to a public cloud service is entirely open — anyone can walk up and request an account — the provisioning of such a service is anything but. There's an admission fee, which can range anywhere from merely handing over certain personal information to paying a hefty subscription, especially in the B2B world. Users have to be authenticated when signing in to a tightly controlled private area where their access to resources is carefully managed and their every move is monitored.
This is exactly the same as the security regime that exists in any well-managed enterprise — but without the complacency that typically accompanies the assumption that an enterprise can trust everyone within its perimeter. Cloud providers offer better security because they are more alert to all of the risks:
- Cloud providers know they live in the cloud and ensure they take effective measures to stay secure.
- They know that if anyone can sign up then they cannot inherently trust their customers or control who they nominate as users. Therefore they make sure that user access controls within their domain are effective and tamperproof.
- Because they can pool the collective buying power of all their customers, they are able to afford the very best security equipment and expertise.
- Their mission is to provide a cloud service and therefore everyone in the organization — especially those who work in IT operations — understands that their livelihood depends on maintaining and upholding rigorous security practices.
For all these reasons, public cloud providers offer far better security than most enterprises. That's why the only headlines you see about cloud security breaches are due to enterprises not keeping up with the demands of operating public-facing services, such as ecommerce, in the cloud. If these organizations had entrusted the data to cloud specialists in the first place it is far more certain the data would have remained safe.
Myth #3 - You can buy your cloud in a box
The popularity of cloud has led many vendors to attach the label to hardware products and systems, even where the link is tenuous at best. They are trying to turn back the clock because they only understand how to sell products.
No wonder people get confused about what cloud means when a vendor like Western Digital goes around offering people their own personal 'My Cloud' — a web-accessible disk unit designed to be installed in a home or a small office. The vendor shamelessly taps into myth #2 by implying in its ads that it's somehow safer to store your data here than entrusting it to public cloud services.The recent revelation that such units have been compromised by the Heartbleed security flaw demonstrates the danger of such inane marketing. Such products have none of the attributes of a robust cloud service, where necessary patches and upgrades are applied instantly by skilled staff. Instead, 'My Cloud' owners have to become IT technicians, installing a firmware update, performing a factory reset and then reinstalling user access.
A box only becomes a cloud computing resource when it's been incorporated into a shared service. It may be virtualized, it may be automated, it may be accessed over the Internet, but it's still only a box until someone takes responsibility for operating it as a pooled resource. That's what makes it a cloud.
Myth #4 - Cloud costs more in the long term
Product-centric thinking gives rise to the frequent misconception that cloud computing costs more than an on-premise alternative. The argument goes that renting or leasing is always more expensive than owning, and therefore the pay-as-you-go cloud option must be more expensive.
This argument would hold water if myth #1 were a reality and we were in fact comparing two versions of the same thing. But in fact a cloud service is not a one-off product that is delivered separately to each individual customer and then never changes. As a share of a pooled resource, it is inherently cheaper than leasing the entire resource. What's more, it continually evolves and improves, with many functional and operational enhancements delivered over the course of each year.
Another argument is that cloud providers are merely undercutting rivals to put them out of business, at which point they'll jack their prices back up. As I said to a recent commenter who advanced this argument in response to a diginomica article, having significantly lower prices than we currently pay for IT would be a good place to be.
Personally I think the risk of an oligopoly developing that ultimately dominates cloud computing is less than many people fear. But do we really prefer today's alternative, where IT is not only costly to acquire and maintain, but can only be upgraded with new functionality at significant additional expense? Don't underestimate the cost of being unable to change and adapt to new business challenges and opportunities.
Myth #5 - Cloud is a bubble that will pop
Perhaps the biggest myth of all is that cloud is a passing fad that people will have forgotten about in a few years' time. In some senses this is another variation on myth #1, fueled by a misconception of how innovative the cloud architecture and business model really is.
But there's another factor that adds fuel to this misconception, which is the often cited evidence that cloud providers are all loss-making businesses. This is another myth that comes from looking at cloud computing as if it were a product business. Cloud providers are subscription businesses, and the financial dynamics of such businesses are very different.
Tien Tzuo, CEO of subscription management vendor Zuora, does a great job of dissecting all that's wrong with the way that traditional accounting fails to capture the dynamics of subscription businesses, most recently in his analysis of Box's pre-IPO financials. To appreciate the full argument, I encourage you to read the whole piece. But here's a key passage:
In the subscription economy, you invest in sales & marketing to acquire customers, and you recognize revenue from those customers over their lifetime, which often can be 3, 5, 10 years or more. However, accounting rules today do not let you spread or depreciate sales and marketing costs over time.
In other words, these cloud businesses are choosing to post financial losses because that's how they invest in growing their revenues. (Note that this only works if they keep their churn rate relatively low — it's not a carte blanche to buy customers at any price). Or as ServiceNow's CEO Frank Slootman recently explained:
"We are growing at a rate and at a scale that has rarely been seen before in the history of enterprise tech, so saying we are not profitable is just repeating mindless drivel. What good is it doing to stick money in the bank and let it sit there and put on 0.1% interest, when we can invest it in a really great enterprise?
"... we are going to invest every dollar that we know how to invest because we have such a great opportunity. If you don't want to be a part of the journey you can go and buy HP, or IBM, see how exciting that is for you. I'm not going to prematurely force this company into profitability just to please some people on Wall Street."
Whether there's a stock market bubble in technology stocks more widely is a different question. I'm not arguing for any particular level of stock prices for cloud providers — the market will decide that. However their businesses run to different financial metrics than traditional product businesses. Therefore a healthy cloud provider may make a deliberate choice to post a loss rather than a profit in times of rapid growth.
Image credit: Edge of the world © Moenez - Fotolia.com; WD My Cloud poster by @philww