Fractal Analysis of User Sessions Inter-Transaction Time in Social Networks

Online social networking sites like YouTube are among the most popular sites on the Internet. Understanding these behavior of user sessions is important, both to improve current systems and to design new applications of online social networks. In this paper, we analyzed numerous traces generated by a ubiquitous measurement technique from an edge network perspective to assess the inter-transaction times characteristics experienced. When viewed as time series data, inter-transaction times often exhibit long-range dependence manifested by Hurst parameter estimates greater than 0.5. Our results suggest that the inter-transaction times can be bursty across multiple time scales even at the microflow level, implying high performance variability during sufficiently long-lived application sessions. We anticipate that the quantification of such phenomena can enable applications to optimize and adjust their operation in care of potential performance degradation. These differences have implications for network capacity planning and design of next generation synthetic Web traffic.