This repository contains the dataset used for the paper "Low-Cost Search in Tree-Structured P2P Overlays: The Null-Balance Benefit" published at the 46th IEEE Local Computer Networks (LCN). We refer to the paper for a full explanation of the methodology used for generating the dataset.
Peer-to-Peer (P2P) networks are one way to create large-scale distributed systems. A single peer has only a limited view on other peers. Thus, efficient searching for other peers or their content is a key performance indicator. In this paper, we investigate the search efficiency in an m-ary tree-structured P2P overlay. While previous work aimed for balancing the maximum height of a node's sub-trees, we show that keeping the height balanced throughout the overall network – a property called null-balance – will increase search performance considerably. Simulations using the ns-3 discrete-event simulator show 50% better performance w.r.t. required routing hops in these null-balanced trees. Therefore, we develop algorithms that keep a tree null-balanced if a node joins or departures. I.e., we prevent the need for restructuring. As we show, the cost of our efficient structure-preserving algorithms is easily set off by a relatively small number of search operations.
The dataset is stored as a comma-separated values (CSV) file, using a semicolon (;) as a delimiter.
NumberOfNodes;Fanout2;Fanout4;Fanout6;Fanout8;Fanout10
1000;...
2000;
3000;
4000;
5000;
6000;
7000;
8000;
9000;
10000;
If you to cite the paper or this dataset for your research, please include the following reference in any resulting publication:
@INPROCEEDINGS{9525004,
author={Detzner, Peter and Gödeke, Jana and Bondorf, Steffen},
booktitle={2021 IEEE 46th Conference on Local Computer Networks (LCN)},
title={Low-Cost Search in Tree-Structured P2P Overlays: The Null-Balance Benefit},
year={2021},
volume={},
number={},
pages={613-620},
doi={10.1109/LCN52139.2021.9525004}}