Accurate online detection of bidimensional Hierarchical Heavy Hitters in software-defined networks

Software-defined networking (SDN) offers many opportunities to redesign architectures, protocols, and applications. However, SDN also demands revisiting traditional solutions in order to cope with new devices, such as the OpenFlow switches. Online detection of Hierarchical Heavy Hitters (HHHs) is an example, since traditional approaches are based on IP headers which are not always available in an efficient manner. In this paper we present a mechanism for online detection of bidimensional HHHs in networks based on OpenFlow switches. We evaluated our proposal in a emulator largely employed for rapid prototyping of SDN and we used both synthetic and real traces. Results show that our proposal has high accuracy in online detecting bidimensional HHHs based on source and destination IP addresses.