The grid: blueprint for a new computing infrastructure
The grid: blueprint for a new computing infrastructure
Data Mining: Concepts and Techniques
Data Mining: Concepts and Techniques
Hi-index | 0.00 |
This paper presents enhanced approach to profiling traffic in mobile packet services such as HSDPA. Deriving accurate and meaningful profiles of traffic generated by packet services can greatly improve dimensioning of the infrastructure for packet mobile networks. Traffic profiles are derived by clustering of daily aggregates of the traffic volume. In this work we propose a new definition of distance between the clustered vectors of daily aggregated traffic. This enhancement allows to derive clusters with desired characteristics in terms of both similar shape of the daily traffic profile and similar busy hour characteristics of each profile. The proposed method is used to obtain traffic profiles from several mobile networks in Europe and Asia. We discuss the differences in characteristics of profiles obtained as a function describing BTS in the context of load shape and its busy hour.