Traffic Prediction and Resource Allocation: A Statistical Delay Bound Approach

摘要

The paper presents a predictive approach to network resourcernallocation techniques. The rationale of this work is to take advantagernof measurements to estimate future traffic behaviorrnby using linear, low-complexity and easy to implement predictionrnalgorithms. The main contribution of the research isrnthe derivation of a bandwidth allocation strategy, named SDBrn(Statistical Delay Bound), which guarantees a probabilisticrnbound on the delay experienced by packets traversing a networkrnresource. In order to assess the effectiveness of the proposedrntechnique, performance is compared to that obtainedrnby running a known algorithm which selects the service capacityrnaccording to a target resource utilization value. Resultsrnshow that, in spite of the simplicity of the adopted predictivernalgorithm, the proposed measurement based technique allowsrnto fulfill the project requirements and candidates for actualrnexperimentation into prototypal routers which support QoSrnmechanisms.