What are the difference between Cloud, Cluster and Grid Computing?

Often the valid question arises if cloud computing is actually something new and, if so, what exactly distinguishes this new paradigm from former computing models like cluster and grid computing. Summarized in table 1, five essential characteristics of cloud computing are applied to cluster and grid computing. All three systems are distributed and share similar characteristics. The similarities relate to resource pooling and broad network access – two criteria that are fulfilled by all systems. Network access to cluster and grid computing systems usually takes place within a corporate network, while the services of a cloud computing system can also be accessed through public network, i.e. the Internet.

Table 1: Comparing Cluster, Grid and Cloud Computing

Table 1: Comparing Cluster, Grid and Cloud Computing

The differences between cloud computing systems on the one hand and grid and cluster computing systems on the other are attributable to the system dynamics. Resources in grid and cluster environments are generally pre-reserved, while cloud computing systems are demand driven, i.e. operation of these systems is geared to consumers’ actual needs. Another difference concerns the “rapid elasticity” criterion, which forms an integral part of cloud computing systems but is not normally supported by cluster or grid systems. Service usage only tends to be accurately measured in grid and cloud computing systems, whereas the majority of cluster environments simply provision rudimentary metering functions.

Compared to other distributed systems such as grids or clusters, cloud computing solutions give enterprises significantly more flexibility. They can dispense with IT infrastructures of their own and only have to pay for the resources and services they actually use (“pay-per-use”/ “pay as you go”). These can be dynamically adapted to changed business requirements and processes with the help of virtualization technologies and service oriented, distributed software systems.