Economies of Scale

The largest cloud providers leverage a variety of techniques to excel and compete, such as building reliability in at the software layer, focusing on reducing power and cooling costs via site selection and engineering, negotiating with suppliers, optimizing taxation, acquiring top engineering talent, and more. However, there is probably no phrase associated with cloud service providers more often than “economies of scale.” Although there surely are such economies, the logic regarding them may not be as simple as some would suggest.

Assertion 3: Large cloud providers enjoy massive economies of scale.

There are no doubt various elements of cloud cost structure that exhibit economies of scale, but there are also scale-invariant costs and diseconomies of scale. Some aspects of cloud architecture, such as network bandwidth and power and cooling equipment, are typically “scale-up,” utilizing larger components with lower unit costs. However, some key building blocks of clouds are “scale-out:” they grow by increasing their quantity. To clarify the difference: an infant grows into an adult by scaling up; an ant colony grows by scaling out. In a scale-out architecture, there can be limits to the applicability of the economies of scale concept. Scale economies require nonlinearities in cost structure, which may not always exist in some of the elements of cloud. Today the building block for many clouds is the container or pod, packed with equipment, and also available ...

Get Cloudonomics: The Business Value of Cloud Computing, + Website now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.