Summary

From an architectural point of view, there are fundamental benefits to consolidation, such as overhead reduction, ability to speed up tasks via parallel processing, de-duplication of some resources, and aggregate demand smoothing and thus better capacity utilization and lower delivered costs.

There are also benefits to dispersion, such as latency reduction leading to productivity and commercial benefits through an improved experience, and improved business continuity.

From an economic point of view, a do-it-yourself strategy for parallelization is costly, since for a given volume of work, acceleration implies greater resource expenditures and lower average utilization. A do-it-yourself strategy for dispersion is also costly, because quadrupling the number of nodes only halves the latency, and at some point adding billions of nodes, say, may reduce latency only by microseconds.

The cloud solves these dilemmas, offering free acceleration through pay-per-use parallel processing and, in some cases, free dispersion via partitionable resources.

Instead of treating each dispersed node as a stand-alone entity, weaving it into a global cloud fabric of processing and storage nodes integrated via an intelligent network can provide the best balance of consolidation and dispersion.

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.