Exploiting edge computing for adaptive data update in Internet of Things networks

With the advent of the pervasive Internet of Things (IoT) era it is expected to have billions of entities simultaneously connected to the network, sharing heterogeneous data to support disparate applications. Such scenario will therefore open new challenges as for network management and information exchange rules. In this context, the increasing data volume may especially lead Cloud-based services to be suffering from overload and data traffic consumption increase when serving a huge number of devices. A potential approach to address this problem is to edge computing, including all those enabling technologies able to move large part of computing close to the data sources, proving several benefits in terms of latency reduction, bandwidth optimization and security [RGXZ17][MTPC19]. Another aspect impacting the performance is the optimization of the amount of data volumes transmitted by the IoT devices. This task is accomplished by specific data synchronization protocols and algorithms that are responsible for information exchange between devices and cloud. In this direction, we consider a decentralized IoT cloud framework where devices connect to the data center through an IoT gateway. Moreover, we present a mechanism for data synchronization that considers Octodiff, a well known tool for data compression, combined with an adaptive algorithm specifically tailored to limited, variable, IoT traffic volumes. By investigating the performance of the proposed architecture, we show how the traffic amount generated by IoT cloud-services can be conveniently reduced