Cloud Economics 2: Definitions
Like any good mathematician, Joe Weinman in his book Cloudonomics lays out some definitions right up front. He chooses to define the concept of “cloud” in cloud computing in a way that brings out five essential attributes of cloud economics that are common to other cloud-like systems in business and life in general. To make it easy to remember he gives his definition as a mnemonic: C L O U D.
Let’s see how these five attributes of any cloud system fit in with our understanding of real-time cloud computing:
C – Common infrastructure – refers to the ability to share resources. A city park is like a cloud in that it can meet the needs of millions of apartment dwellers for some quality outdoor space—gardens, walkways, playgrounds, and sports fields. Nobody feels overly crowded because they don’t all use the park at the same time, or in the same way.
As Wienman explains in detail later on in the book, non-cloud computing resources are often underutilized, which becomes a cost. For example, some industrial applications require their software to run alone, on a separate server. As the number of this kind of application grows, the waste of resources increases. Where possible, using virtual machines is one way to share the resources of a single server to reduce this kind of waste. This approach to sharing infrastructure is often used in cloud systems, as well as private systems.
L – Location independence – means that the service is available pretty much everywhere. You might not think of a fast-food franchise as a cloud service provider, but in a sense it is similar. Just as you can get order-in or take-out service from your favorite burger outlet in many places around the country or even the world, so also can you access the cloud from practically any location.
The value of location independence for real-time systems is just beginning to be realized. For decades data from industrial systems has been tightly locked down, behind firewalls and physically isolated systems. But now, perhaps to the dismay of engineers and system integrators who rely on isolation for security reasons, upper management in many companies is waking up to the value of accessing that data from anywhere.
Of course, there is always a need to keep raw process data secure and free from interference, but advanced methods of keeping firewalls closed and permitting read-only access can help bring key real-time performance metrics to analysts and decision makers in the office, at home, or on the road.
At the same time, many embedded systems once lacked the power or connectivity to put their data online. With the advent of the Internet of Things connecting cars, appliances, remote sensors, and a host of other devices directly to the Internet, we are witnessing a huge growth and interest in accessing live data from all kinds of sources, independent of location.
O – Online accessibility – is the availability of service via a network or the Internet. Every service needs some form of access. A restaurant needs an eating area, a movie theater needs seats and a view of the screen, a radio show needs transmitters and receivers. As Wienman sums it up: “Without networks, there is no cloud.” Real-time cloud systems can function well on private networks, and in many cases access to the Internet and public clouds will provide additional value.
U – Utility pricing – like the Water Works and Electric Company in the game of Monopoly, utility pricing means you only pay for what you use—be it water, electricity or computing power. Usually this aspect of cloud computing goes hand-in-hand with on-demand resources.
D – on-Demand resources – the ability to bring in additional resources, or remove extra ones, to cope with variable demand. For example, your house has plenty of space for your family and an occasional guest, but on special occasions like a big wedding you may need to engage the services of hotels or restaurants.
The flexibility to respond to market fluctuations is a real boon for retail and consumer-oriented companies who may see significant peaks and valleys of seasonal or irregular demand. In our experience, most industrial and embedded real-time systems don’t undergo such large variations in demand for computing resources. However, for systems too small or too dispersed to justify a dedicated, in-house SCADA system, (such as mentioned in our SCADA for the Masses discussion), on-demand resources and utility pricing may help make the cloud a viable solution.
Given the above C L O U D definitions, the economic value of any cloud computing system, real-time or not, depends on a number of variables and circumstances. We need to consider these in their appropriate context to determine how real-time systems can benefit.
