Cloudonomics: The Business Value of Cloud Computing, + Website

CHAPTER 12

Hybrids

A hybrid is “anything derived from heterogeneous sources, or composed of elements of different or incongruous kinds.”1 In cloud computing, a variety of architectures meet that definition. As we’ve seen, in the presence of variable demand and a utility premium for cloud resources, hybrids can be cost optimal. However, rather than just one hybrid, there are several different permutations of users, dedicated flat-rate resources, and on-demand pay-per-use resources, and several different means of linking them together.

For example, one such permutation is a mix of owned resources in a data center with flexible resources in the cloud. Another is owned resources in colocation space with physically adjacent flexible resources sold on a pay-per-use basis. Yet another type of hybrid is the Web app or cloud-enabled app. For example, the Dragon Naturally Speaking dictation app on a tablet converts spoken words to text. The client-resident functionality is limited primarily to an interface; the spoken words are captured as an audio stream, compressed and encoded, and then transmitted to the cloud for speech-to-text processing, and then the text stream is sent back to the device. Google or Bing search running in the cloud is accessed via a client interface—a dynamic HTML page or a smartphone widget, say—but the heavy-duty processing is done in the cloud, and search results are returned to the device. Or those functions can be combined and semantic analysis included, which ...