Repository

The cloud can also act as a repository of information. Consider the Library of Congress, Google Books, or the index of all the accessible Web sites on the planet that each major search provider maintains.

The benefit of a single repository versus multiple copies is a function of the cost of storage relative to the cost of networking and the pattern of access. When storage is cheap and networking is expensive, it’s better to maintain a local copy of the information. Conversely, when storage is expensive and networking is cheap, it’s better to access a remote copy.

When both storage and networking are attractively priced, but nonzero, there is a breakeven point based on the pattern of access. Frequently accessed items should be stored locally, and infrequently accessed items should be accessed remotely.

If we exclude network costs, the benefit of a single repository is clear: It is less expensive to maintain one copy than n copies. If the cost to store a copy is s, the cost to access a copy is a, and the number of accesses is k, then the break-even point between strategies occurs when the cost to store n copies locally—n × s—is equal to the cost to remotely access a single copy k times, which is s + k × a.

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.