English

Optimal Tracking of Distributed Heavy Hitters and Quantiles

Data Structures and Algorithms 2008-12-02 v1

Abstract

We consider the the problem of tracking heavy hitters and quantiles in the distributed streaming model. The heavy hitters and quantiles are two important statistics for characterizing a data distribution. Let AA be a multiset of elements, drawn from the universe U={1,...,u}U=\{1,...,u\}. For a given 0ϕ10 \le \phi \le 1, the ϕ\phi-heavy hitters are those elements of AA whose frequency in AA is at least ϕA\phi |A|; the ϕ\phi-quantile of AA is an element xx of UU such that at most ϕA\phi|A| elements of AA are smaller than AA and at most (1ϕ)A(1-\phi)|A| elements of AA are greater than xx. Suppose the elements of AA are received at kk remote {\em sites} over time, and each of the sites has a two-way communication channel to a designated {\em coordinator}, whose goal is to track the set of ϕ\phi-heavy hitters and the ϕ\phi-quantile of AA approximately at all times with minimum communication. We give tracking algorithms with worst-case communication cost O(k/\epslogn)O(k/\eps \cdot \log n) for both problems, where nn is the total number of items in AA, and \eps\eps is the approximation error. This substantially improves upon the previous known algorithms. We also give matching lower bounds on the communication costs for both problems, showing that our algorithms are optimal. We also consider a more general version of the problem where we simultaneously track the ϕ\phi-quantiles for all 0ϕ10 \le \phi \le 1.

Keywords

Cite

@article{arxiv.0812.0209,
  title  = {Optimal Tracking of Distributed Heavy Hitters and Quantiles},
  author = {Ke Yi and Qin Zhang},
  journal= {arXiv preprint arXiv:0812.0209},
  year   = {2008}
}

Comments

10 pages, 1 figure

R2 v1 2026-06-21T11:46:57.343Z