ABSTRACT

Clusters are often specialized computing infrastructures. An alternative to this is to use the resources on a wide range of computers that have resources

to spare. For example, many computers spend most of their time with less than 5% CPU utilization, and could thus offer some of the spare CPU resource to another application (as long as it was trusted, of course). This is the main concept behind grid computing, where a distributed computer can be set up, with a range of computer architectures and operating systems, spread over a large geographical area. Normally with grid computing the required resources would be mapped to physical computers, which had resources free at a given time. The advantage with cloud infrastructures is that the actual resource does not need to exist, until it is actually required. In this way computing resources can be created, consumed, and even deleted, whenever they are required. In this way, a hardware, network, server, or service instance can be created at any given time, and then deleted when not required anymore.