Research Area:  Fog Computing
With the rapid increase of the number of IoT devices, transmitting big amount of data from these devices to data centers which are far away will cause problems like high latency or network congestions. Fog Computing provides a better solution for Fog-enabled latency sensitive data services to place data on Fog nodes which are closer to the data generators. However, recent studies only focus on the data placement problem of placing one single data replica to the proper Fog node. Under the situation that there are several data consumers whose topology positions are different subscribing the same data, one single data replica cannot meet the latency requirement of all the consumers. Hence, we build a multi-replica data placement model iFogStorM for Fog Computing to formulate the problem of how many data replicas need to be placed on Fog nodes and how to optimize the data placement. Furthermore, we propose a greedy algorithm based data replica placement strategy, MultiCopyStorage, to reduce the overall latency. MultiCopyStorage uses a pruning method to filter the inferior solutions calculates the overall latency and chooses the solution with the minimum overall latency as the final solution. We conducted experiments on iFogSim, a toolkit for modeling and simulation of Fog Computing, evaluated the proposed strategy with the CloudStorage strategy, Closest Node strategy, iFogStor strategy, and two kinds of heuristic strategy, iFogStorZ, and iFogStorG. The experiment result demonstrates that MultiCopyStorage strategy reduces the overall latency by 6% and 10% compared to iFogStor and iFogStorG strategy respectively. Meanwhile, execution time of the MultiCopyStorage is less than the heuristic strategy, iFogStorG and iFogStorZ, which proves that the proposed strategy can support real-time scheduling.
Author(s) Name:  Tiansheng Huang, Weiwei Lin, Yin Li, LiGang He & ShaoLiang Peng
Journal name:  Journal of Signal Processing Systems
Publisher name:  Springer
Volume Information:  volume 91, pages1191–1204 (2019)
Paper Link:   https://link.springer.com/article/10.1007/s11265-019-1444-5