Lecture Time & Place: Thursday 10:00-12:00, Ziskind 155
TA: Avi Cohen
Prerequisite: Sufficient comfort with both algorithm design & analysis, and basic knowledge in distributed algorithms (e.g., taking the Distributed Network Algorithms course of Prof. David Peleg).
New: Final assignment instructions
Course Schedule (tentative):
- (28/03) Lecture 01: Introduction (models, local and global graph problems), network decomposition and applications. Lecture-01
- Network decomposition and locality in distributed computation (Awerbuch, Luby, Goldberg and Plotkin, 1989)
- On the complexity of distributed network decomposition (Panconesi and Srinivasan, 1996).
- Lecture 04 of Mohsen Ghaffari (Distributed Graph Algorithms 6.S899, MIT)
- Chapter 9 in the Distributed Graph Coloring book of Barenboim and Elkin, 2001.
- (04/04) Lecture 02: Sequential and randomized network decomposition.
Exercise-1 (due May 2nd), Lecture-02- Low Diameter Graph Decomposition (Linial and Saks, 1993)
- Section 1.5.2 of Mohsen Ghaffari's notes (Principles of Distributed Computing S19, ETH)
- (18/04) Lecture 03: Randomized and sublogarithmic coloring algorithms Lecture-03
- Chapter 10 in Distributed Graph Coloring by Barenboim and Elkin, 2001
- The Locality of Distributed Symmetry Breaking by Barenboim, Elkin, Pettie, and Schneider, 2012.
- Notes by Mohsen Ghaffari, Section 1.7
- (25/04) No Class Due to Passover
- (02/05) Lecture 04: An Improved Distributed Algorithm for Maximal Independent Set by Ghaffari, 2016. Lecture-04, Exercise-2 (Due Monday 20/5)
- A more detailed description of the algorithm appears in Ghaffari's thesis (Section 3).
- A more detailed description of the algorithm appears in Ghaffari's thesis (Section 3).
- (09/05) No Class Due to Yom-Ha'atzmaut
- (16/05) Lecture 05: An Exponential Separation between Randomized and Deterministic Complexity in the LOCAL Model (Chang, Kopelowitz, and Pettie, 2017) New: Lecture-05
- (23/05): No Class Due to Lag B'aomer.
- (30/05) Lecture 06: Lovasz Local Lemma (LLL) and Distributed LLL New: Lecture-06
-
A constructive proof of the general Lovasz Local Lemma (Moser and Tardos, 2009)
-
Lectures 04 and 05 by Bernhard Haeupler (Algorithmic Superpower Randomization, CMU, Fall 2019)
-
For additional recommanded lectures notes, see here and here
-
- (06/06) Lecture 07: Improved distributed LLL and applications New: Lecture-07 Exercise-03 (Due June 30)
- Distributed Algorithms for the Lovász Local Lemma and Graph Coloring (Chung, Pettie and Su, 2014)
- Sublogarithmic Distributed Algorithms for Lovász Local lemma, and the Complexity Hierarchy (Fischer and Ghaffari, 2017)
- (13/06) Lecture 08: Intro to global problems in the CONGEST model. Minimum Spanning Tree and low-congestion shortcuts New: Exercise-04
- Lecture 10 of Principles of Distributed Computing 2019, ETH
- Lecture 6 and Lecture 7 of Theory of Distributed Systems, 2016, MPI
- MST Lower Bound, Chapter 24 in Peleg's Book
- (20/06) Lecture 9: Distributed minimum cut and connectivity decomposition
- Distributed Minimum Cut Approximation (Ghaffari and Kuhn, 2013)
- Lecture 11 of Principles of Distributed Computing 2019, ETH
- (27/06) Class Cancelled
- (04/07) Lecture 10: Communication complexity based lower bounds, guest class by Seri Khoury (UC Berkeley)
- Deterministic lower bound for set disjointness in the 2-party communication complexity model (Communication Complexity book by Kushilevitz and Nisan)
- Lower bounds in the CONGEST model via reductions to communication complexity
- Distributed Verification and Hardness of Distributed Approximation by Das-Sarma et al., 2011
- Lower bound for exact diameter computation based on Networks Cannot Compute Their Diameter in Sublinear Time by Frischknecht, Holzer and Wattenhofer, 2012
- Lower bound for 4-cycle detection based on On the Power of the Congested Clique Model by Drucker, Kuhn and Oshman, 2014
- (11/07) Lecture 11: Distributed Single Source Shortest Paths (SSSP)
Useful Material:
- Probabilities Inequalities: Handout by Avrim Blum and Anupam Gupta (15-859 Randomized Algorithms, CMU, Spring'11).
- Distributed Computing -- A Locality-Sensitive Approach, David Peleg, SIAM.
- Distributed Graph Coloring by Barenboim and Elkin, 2001