Presently, cloud computing is being used to store large amounts of data for sharing among users. However, it imposes cloud data-out charges and data access delays on organizations. These problems can be relieved by using cloud caching to prevent repetitive data loading from clouds. This paper proposes technical and economic models and a system for operating a client-side shared cloud cache to be turned into a Cache-as-a-Service (CaaS). The models and the system are our novel contributions. Evaluation by simulation using two experimental data sets, representing large-sized and small-to-medium-sized cloud data, demonstrated that the technical model achieved cost-saving ratios up to 56.20% and delay-saving ratios up to 56.65%, and that the economic model recommended a monthly service charge of CaaS to be 24,599.70 USD on average based on a largesized cloud data set and 657.93 USD on average based on a small-to-medium-sized cloud data set. The CaaS system gained a good level of user satisfaction.
