Abstract:This paper seeks to address the question of designing distributed algorithms for the setting of compact memory i.e. sublinear bits working memory for arbitrary connected networks. The nodes in our networks may have much lower internal memory as compared to the number of their possible neighbours implying that a node may not be able to store all the IDs of its neighbours. These algorithms are useful for large networks of small devices such as the Internet of Things, for wireless or ad-hoc networks, and, in general, as memory efficient algorithms. We introduce the Compact Message Passing(CMP) model;an extension of the standard message passing model considered at a finer granularity where a node can interleave reads and writes with internal computations, using a port only once in a round. The interleaving is required for meaningful computations due to the low memory requirement and is akin to a distributed network with nodes executing streaming algorithms. Note that the internal memory size upper bounds the message sizes and hence e.g. for log-memory, the model is weaker than the Congest model; for such models our algorithms will work directly too. We present early results in the CMP model for nodes with log^2-memory. We introduce the concepts of local compact functions and compact protocols and give solutions for some classic distributed problems (leader election, tree constructions and traversals). We build on these to solve the open problem of compact preprocessing for the compact self-healing routing algorithm CompactFTZ posed in Compact Routing Messages in Self-Healing Trees(TCS2017) by designing local compact functions for finding particular subtrees of labeled binary trees. Hence, we introduce the first fully compact self-healing routing algorithm. We also give independent fully compact versions of the Forgiving Tree[PODC08] and Thorup-Zwick's tree based compact routing[SPAA01].

Two Classes of Optimal Multi-Input Structures for Node Computations in Message Passing Algorithms

A Class of Benes-Based Optical Multistage Interconnection Networks for Crosstalk-Free Realization of Permutations.

Optimality of Message-Passing Architectures for Sparse Graphs

Better Necessary Conditions for Rearrangeably Nonblocking f-cast d-ary Multi-log Networks under Fanout and Crosstalk Constraints

Some Problems in Compact Message Passing

Polynomial-time Construction of Optimal Tree-structured Communication Data Layout Descriptions

Optimal low-latency network topologies for cluster performance enhancement

An Optimal, Strategy-Proof Scheme for Multi-Path Traffic Assignment in Non-Cooperative Networks

Efficient routing and message bounds for optimal parallel algorithms

Two Novel Structures Of Optical Packet Switching Nodes

Multilayer obstacle-avoiding x-architecture steiner minimal tree construction based on particle swarm optimization

Exact Algorithms for Multiagent Path Finding with Communication Constraints on Tree-Like Structures

Detection of Graph Structures Via Communications over a Multiaccess Boolean Channel

Multi-structure Objects Points-to Analysis

Light-Hierarchy: The Optimal Structure for Multicast Routing in WDM Mesh Networks

Timing-Driven Obstacle-Avoiding X-Architecture Steiner Minimum Tree Algorithm With Slack Constraints

A Time and Space Efficient Junction Tree Architecture

Towards Dynamic Message Passing on Graphs

Toward Efficient Distributed Algorithms for In-Network Binary Operator Tree Placement in Wireless Sensor Networks

A Dual-Port Access Structure of 3D Mesh-Based NoC

In the Search of Optimal Tree Networks: Hardness and Heuristics