2023 Fiscal Year Annual Research Report

Bridging the gap between theory and practice for distributed graph algorithms

Research Project

Project/Area Number	21KK0204
Research Institution	Japan Advanced Institute of Science and Technology
Principal Investigator	シュワルツマングレゴリー北陸先端科学技術大学院大学, 先端科学技術研究科, 准教授 (20815261)
Project Period (FY)	2022 – 2023
Keywords	Distributed computing / Networking / Graph algorithms
Outline of Annual Research Achievements	We introduced a new framework for running message passing algorithms on programmable switches to enable fast reactions to network events in large networks, specifically P4 switches. We use our framework to improve over the state of the art for two fundamental networking problems: multicast and clock synchronization. We also a introduced the μ-CONGEST model where on top of the bandwidth restriction, the memory of nodes is also limited to μ words, in line with real-world systems. We provide fast algorithms and lower bounds in this new model, specifically for subgraph detection and streaming algorithm. Finally, we present new results for the well studied mini-batch k-means algorithm. This algorithm can be used in the field of networking for vector quantization and clustering of network traffic.