Centralized and distributed algorithms are remarkable by their capacity of accurately estimating different types of statistics in a space and computational efficient way. The first part of this book is dedicated to the presentation of mathematical tools that will be needed in the remaining of the book. Then, a survey of the important models, and algorithms will be presented. Each of these algorithms will be accompanied by a theoretical and experimental analysis. Based on this material, two important problems will be studied: the detection of DDoS attacks and the construction of a uniform data stream out of arbitrary biased ones.