Predictive Replica Placement for Mobile Users in Distributed Fog Data Stores with Client-Side Markov Models

TL;DR

This paper introduces client-side Markov models to predict mobile user movement and data consumption patterns, enabling efficient data replication in fog environments with 35% improved data availability.

Contribution

It proposes variations of Markov models that run on clients to enhance data availability without global replication overheads.

Findings

Data availability improved by 35% at closest fog nodes.

Client-side Markov models reduce storage and communication overheads.

Enhanced prediction accuracy for mobile user behavior.

Abstract

Mobile clients that consume and produce data are abundant in fog environments and low latency access to this data can only be achieved by storing it in their close physical proximity. To adapt data replication in fog data stores in an efficient manner and make client data available at the fog node that is closest to the client, the systems need to predict both client movement and pauses in data consumption. In this paper, we present variations of Markov model algorithms that can run on clients to increase the data availability while minimizing excess data. In a simulation, we find the availability of data at the closest node can be improved by 35% without incurring the storage and communication overheads of global replication.