Weighted Load Balancing in Distributed Hash Tables

The rising amount of data in Internet of Things (IoT) and Wireless Sensor Network (WSN) scenarios motivates new computing paradigms like fog or edge computing. To reduce the amount of data sent upstream, in-network (pre-)processing is widely used, which demands for both compute and distributed storage capacities in highly constrained environments. This paper introduces a new way of using Distributed Hash Tables (DHTs) to create a distributed storage in P2P-networks. The main design goals are to introduce the lowest overheads possible and allowing for fair load balancing, even if nodes contributing storage capacities of arbitrary/different sizes form the network. A combination of an optimized bootstrap mechanism and a virtual node scheme that deploys a variable number of virtual nodes depending on a node's storage capacity yields success. An evaluation and comparison with state of the art work shows that the new method performs well in terms of load balancing while minimizing overheads introduced by newly introduced virtual nodes.