Location-Based Cloud Resource Allocation Based on Information of the Social Web

Abstract

The development of mechanisms as virtualization, load distribution, and data sharing has benefited the IT-evolution towards to cloud computing with its seemingly infinite resource capacities. This leads to novel approaches in resource provisioning and releasing. However, managing the cloud resource pool still needs manual configuration to react on changes in system load behaviour. This paper presents simulation results of the already described idea, which procures user-data based on social Web applications and uses this information for load forecasting. The solution includes the data of geocaching- or geotagging-services to determine how many people are located at a particular place with a specific interest focus. By inferring the pure load volume at a specific location and by the presented formulas, our system is able to provide an appropriately amount of resources. Thereby, we can optimize the trade-off between low costs with less resources and high user satisfaction with use of many machines.