NEAT Project

NEtwork dATa plane measurements as first class primitive.

UKRI Logo

One poster accepted at ACM CoNEXT 2020

With DISCOvering the Heavy Hitters with Disaggregated Sketches, we propose an approach to flow monitoring that use fewer resources at switches than state-of-the-art solutions while increasing the heavy flows detection accuracy. This is achieved by disaggregating the computation of a (logically) centralized single big sketch into multiple small-sketches distributed across the flows’ paths.