A Cloud Monitor to Reduce Energy Consumption With Constrained Optimization of Server Loads

As the energy consumption of cloud computing increases, resource optimization becomes more important in this dynamically changing computation environment. Cloud monitor is an effort for better job distribution models and cloud load optimization on cloud computing platforms to reduce energy consumption, decrease computation and waiting time in the queue, and improve the overall utilization of resources. We use linear programming to optimize the distribution of jobs whose arrival and duration times are Poisson distributed to the servers. Our simulations show 1) improvements in job distributions with better correlation with the energy consumption of servers, 2) possible increased utilization of resources in some settings, and 3) significant reductions in energy consumption that reach 42%. We present the mathematical formulation of our model as well as how to represent it as a linear programming problem. The constraints introduced allow the mathematical definition of the optimal distribution, optimal redistribution, and optimal elasticity for cloud computing. Cloud monitor has potential to contribute to more sustainable and efficient cloud computing systems.