English

Sparsifying Distributed Algorithms with Ramifications in Massively Parallel Computation and Centralized Local Computation

Data Structures and Algorithms 2018-07-18 v1 Distributed, Parallel, and Cluster Computing

Abstract

We introduce a method for sparsifying distributed algorithms and exhibit how it leads to improvements that go past known barriers in two algorithmic settings of large-scale graph processing: Massively Parallel Computation (MPC), and Local Computation Algorithms (LCA). - MPC with Strongly Sublinear Memory: Recently, there has been growing interest in obtaining MPC algorithms that are faster than their classic O(logn)O(\log n)-round parallel counterparts for problems such as MIS, Maximal Matching, 2-Approximation of Minimum Vertex Cover, and (1+ϵ)(1+\epsilon)-Approximation of Maximum Matching. Currently, all such MPC algorithms require Ω~(n)\tilde{\Omega}(n) memory per machine. Czumaj et al. [STOC'18] were the first to handle Ω~(n)\tilde{\Omega}(n) memory, running in O((loglogn)2)O((\log\log n)^2) rounds. We obtain O~(logΔ)\tilde{O}(\sqrt{\log \Delta})-round MPC algorithms for all these four problems that work even when each machine has memory nαn^{\alpha} for any constant α(0,1)\alpha\in (0, 1). Here, Δ\Delta denotes the maximum degree. These are the first sublogarithmic-time algorithms for these problems that break the linear memory barrier. - LCAs with Query Complexity Below the Parnas-Ron Paradigm: Currently, the best known LCA for MIS has query complexity ΔO(logΔ)poly(logn)\Delta^{O(\log \Delta)} poly(\log n), by Ghaffari [SODA'16]. As pointed out by Rubinfeld, obtaining a query complexity of poly(Δlogn)poly(\Delta\log n) remains a central open question. Ghaffari's bound almost reaches a ΔΩ(logΔloglogΔ)\Delta^{\Omega\left(\frac{\log \Delta}{\log\log \Delta}\right)} barrier common to all known MIS LCAs, which simulate distributed algorithms by learning the local topology, \`{a} la Parnas-Ron [TCS'07]. This barrier follows from the Ω(logΔloglogΔ)\Omega(\frac{\log \Delta}{\log\log \Delta}) distributed lower bound of Kuhn, et al. [JACM'16]. We break this barrier and obtain an MIS LCA with query complexity ΔO(loglogΔ)poly(logn)\Delta^{O(\log\log \Delta)} poly(\log n).

Keywords

Cite

@article{arxiv.1807.06251,
  title  = {Sparsifying Distributed Algorithms with Ramifications in Massively Parallel Computation and Centralized Local Computation},
  author = {Mohsen Ghaffari and Jara Uitto},
  journal= {arXiv preprint arXiv:1807.06251},
  year   = {2018}
}

Comments

This is a shortened version of the abstract. Please see the pdf for the full version

R2 v1 2026-06-23T03:03:49.262Z