A Scalable Framework for Wireless Distributed Computing

Abstract
In this paper, a scalable framework for wireless distributed computing is proposed, in which the data shuffling load does not increase with the number of users in the network. The key idea is to utilize a particular repetitive structure of computation assignments at the users, in order to provide coding opportunities at both the users and the access point, which reduce the shuffling load by a factor that grows linearly with the number of users. It is demonstrated that the proposed computation assignments and coded shuffling schemes are optimal (i.e., achieve the minimum shuffling load) for both a centralized setting and a decentralized setting, by developing tight information-theoretic lower bounds.
View on arXivComments on this paper