A case study of DHT functionality: range query on DHT

Distributed Hash Table [1] (DHT) is a method to store and lookup data across distributed nodes. It provides a lookup service to locate a corresponding node for a given key using a hash function. DHT is scalable, fault-tolerant and widely used in P2P systems [1]. However, it does not inherently support range queries, since DHT is based on a hash function [2], In this project, a method to perform range queries on DHT is introduced with assumption of centralized index of all keys partitioned by tables. To verify and evaluate this method, a simple Chord [3] based DHT system was implemented using Erlang/OTP [4, p. XIII] [4, p. 263] [5], Experiments were performed using several machines. These experiments include measuring the performance under different conditions, comparison with a simple get method, memory usage of the range query method, and load balancing considerations. The experimental results show that the range query method outperforms exhaustive search using a simple get method; and the data load distribution was more balanced as the number of virtual nodes increases.