Currently, the HTTP protocol supports a lot of different applications, with distinct traffic intensities and patterns.
The ability to accurately map traffic flows to their corresponding application can be very useful in several operational
and management tasks, like service differentiation and personalization, network resources optimization, network
management and security. This paper discusses the possibility of performing the identification of HTTP-based applications
in an efficient way, using a classification approach based on the multi-scale analysis of the traffic flows: by performing
a wavelet decomposition, captured traffic can be fully characterized in terms of its time and frequency components, allowing
the identification of differentiating characteristics/behaviours that can be used to discriminate between different
applications. The results obtained by applying this methodology to traffic belonging to several Web-based applications
show that it is able to achieve good classification results, while being immune to some of the most important drawbacks
that limit the applicability of the most popular traffic identification approaches. Finally, the paper reviews the most relevant
traffic classification methodologies that have been published so far, including some recent patents.
Keywords: Application identification, multi-scale analysis, wavelet transform.Application identification, multi-scale analysis, wavelet transform.Application identification, multi-scale analysis, wavelet transform.
Rights & PermissionsPrintExport