Cloudonomics: The Business Value of Cloud Computing, + Website

CHAPTER 17

Parallel Universe

To specify any event in our universe—a car accident, a flight departure, or a royal coronation or presidential inauguration—requires both a location in space—9233 Elm Street—and a time—4:00 P.M. ET on November 3, 1932. Albert Einstein and his teacher Herman Minkowski proposed that space and time are components of an integrated entity: the space-time continuum. Rather than merely combining two unrelated concepts, their point was that, at relativistic speeds, time slowed down, so that space and time are inextricably linked.

What does relativistic physics have to do with cloud computing? Plenty.

Although the term “space-time” originated in physics, computer scientists have co-opted it. The complexity of a problem can be measured by the amount of time it takes to solve it as well as the amount of space—for example, storage or memory.1 Moreover, as with real space and time, one quantity can be related or converted to the other. More processing can be used to make up for a lack of storage space, or more space can be used to save processing time.

Here we use the term “space-time” even more broadly to relate to general trade-offs between resources and time, in the spirit of the aphorism that “Many hands make light work.” Specifically, with more processors—central processing units, cores, or graphics processing units—available to perform a calculation, often processing time can be decreased.

In the easiest case, time and space can be inversely proportional to ...