On generalized gossiping and broadcasting

Abstract

Abstract. The problems of gossiping and broadcasting have been widely studied. The basic gossip problem is defined as follows: there are n individuals, with each individual having an item of gossip. The goal is to communicate each item of gossip to every other individual. Communication typically proceeds in rounds, with the objective of minimizing the number of rounds. One popular model, called the telephone call model, allows for communication to take place on any chosen matching between the individuals in each round. Each individual may send (receive) a single item of gossip in a round to (from) another individual. In the broadcasting problem, one individual wishes to broadcast an item of gossip to everyone else. In this paper, we study generalizations of gossiping and broadcasting. The basic extensions are: (a) each item of gossip needs to be broadcast to a specified subset of individuals and (b) several items of gossip may be known to a single individual. We study several problems in this framework that generalize gossiping and broadcasting. Our study of these generalizations was motivated by the problem of managing data on storage devices, typically a set of parallel disks. For initial data distribution, or for creating an initial data layout we may need to distribute data from a single server or from a collection of sources.

Broadcasting in Heterogeneous Tree Networks

“…For the telephone model in an unweighted complete graph, Khuller et al [8] presented polynomialtime approximation algorithms for the single-source multicasting, i.e., |A| = 1, multi-source broadcasting problems, i.e., B = V , and multi-source multicasting problems.…”

“…In this paper, we consider the broadcasting problem in a heterogeneous tree network T = (V, E), where the weight w(u, v) of each edge u, v ∈ E(G), denotes the transmission time required. The broadcasting problem has been extensively studied for several decades [1,[6][7][8][9][11][12][13] due to the increasing demands of heterogeneous network of workstations [7,9]. The broadcasting problem is to find a broadcast center such that the broadcasting time from the broadcast center to all vertices in T is minimized.…”

New bounds on the minimum number of calls in failure‐tolerant gossiping

“…The gossiping problem and some variations of it have been extensively studied for several decades (see, for example, [3][4][5][7][8][9]). In the gossiping problem, first proposed by A. Boyd in 1971, there are n ladies, each of whom knows a unique message that is not known by any of the others.…”

Implementing Atomic Data through Indirect Learning in Dynamic Networks

“…Performance of a service implemented in this way depends on the prompt update of the local state at each node, hence requiring (logical) all-to-all communication, which can be quite expensive. The communication cost associated with all-to-all communication can be reduced by minimizing the number of bits in the message [2], or by limiting the communication by assigning to each sender a proper subset of the nodes to communicate with [12]. Such methods can lead to good results in static environments, however their utility is diminished in highly dynamic networks.…”