Faster asynchronous MST and low diameter tree construction with sublinear communication
Building a spanning tree, minimum spanning tree (MST), and BFS tree in a distributed network are fundamental problems which are still not fully understood in terms of time and communication cost. x The first work to succeed in computing a spanning tree with communication sublinear in the number of e...
Gespeichert in:
Hauptverfasser: | , |
---|---|
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Building a spanning tree, minimum spanning tree (MST), and BFS tree in a
distributed network are fundamental problems which are still not fully
understood in terms of time and communication cost. x The first work to succeed
in computing a spanning tree with communication sublinear in the number of
edges in an asynchronous CONGEST network appeared in DISC 2018. That algorithm
which constructs an MST is sequential in the worst case; its running time is
proportional to the total number of messages sent. Our paper matches its
message complexity but brings the running time down to linear in $n$. Our
techniques can also be used to provide an asynchronous algorithm with sublinear
communication to construct a tree in which the distance from a source to each
node is within an additive term of $\sqrt{n}$ of its actual distance.
We can convert any asynchronous MST algorithm with time $T(n, m)$ and message
complexity of $M(n, m)$ to an algorithm with time $O(n^{1 - 2\epsilon} + T(n,
n^{3/2 + \epsilon}))$ and message complexity of $\tilde{O}(n^{3/2 + \epsilon} +
M(n, n^{3/2+\epsilon}))$, for $\epsilon \in [0, 1/4]$. Picking $\epsilon = 0$
and using Awerbuch's algorithm \cite{awerbuch1987optimal}, this results in an
MST algorithm with time $O(n)$ and message complexity $\tilde{O}(n^{3/2})$.
However, if there were an asynchronous MST algorithm that takes time sublinear
in $n$ and requires messages linear in $m$, by picking $\epsilon > 0$ we could
achieve sublinear time (in $n$) and sublinear communication (in $m$),
simultaneously. To the best of our knowledge, there is no such algorithm.
All the algorithms presented here are Monte Carlo and succeed with high
probability, in the KT1 CONGEST asynchronous model. |
---|---|
DOI: | 10.48550/arxiv.1907.12152 |