Predictive Elastic Load Management for Cloud Computing Infrastructures

Cloud computing has emerged as a promising platform that grants users with direct yet shared access to computing resources and services without worrying about the internal complex infrastructure. We present a Predictive Elastic Load Management for Cloud Computing Infrastructures to achieve quality-aware service delivery in multi-tenancy cloud computing infrastructures. Our system dynamically captures fine-grained signatures of different application tasks and cloud nodes using time series patterns[1], and performs precise resource metering and allocation based on the extracted signatures[3]. Our system employs several mathematical algorithms like FFT, polynomial fitting, dynamic time warping algorithm, multi-dimensional time series indexing to achieve efficient signature pattern profiling and matching. To achieve more predictive resource control, we have also introduced machine learning techniques to predict future resource usage of tasks based on historical data.