Green Cloud Framework For Reducing Carbon Dioxide Emissions in Cloud Infrastructure

The increasing deployment of datacenters and cloud resources around the globe escalated by higher electricity prices advanced energy cost, cooling and communication cost, and carbon dioxide consumption. To curb such ever-increasing problem complexity, we have formulated a scientific workflow-based cost-effective paradigm based on rigorous mathematical model. Multiple techniques have been considered to increase system utilization rate within acceptable performance bounds. First, we have applied Dynamic Voltage and Frequency Scaling (DVFS) approach to scale down the power consumption by cloud servers via calculating the best near-optimal frequency. Then, we have reused cloud-based VMs to execute as many scientific workflows as possible using fewer cloud servers which conserves a tremendous amount of energy cost including carbon dioxide emission consumption and electricity cost. This has been done through eliminating the overhead of sharing multiple VMs the same server's capacity. Moreover, the aforementioned objectives have been achieved without degrading the Quality of Service (QoS) specified in Service Level Agreement (SLA). However, we have simulated our heuristic using open source CloudSim and compared with algorithms such as the Rank and EARES-D. The results have showed that our paradigm model is better than other heuristics with an average energy reduction of 70%.