In-Network Data Aggregation in ICN Demo paper

Internet-of-Things (IoT) devices, wearables, sensor networks and lately vehicles produce continuous streams of data. Distributing every data item to all interested applications can quickly bring the underlying network to its capacity limits. Moreover, applications are often only interested in the "big picture" derived from aggregates of this data. Hence, data aggregation becomes a necessary step in order to optimize bandwidth utilization. The standard approach to data aggregation is based on a centralized, application-specific method, with gateways collecting the data produced by the various devices and sending it to a central location, where data is aggregated and stored. Aggregation is performed with a specific network and/or application in mind, hence limiting the way this data may be used by other, current or future applications. This approach is also less suitable for applications and networks which are distributed in nature, like in the case of vehicle-to-vehicle communication. To cater for these use cases, we argue the need of in-network data aggregation, agnostic of data location and applications using it. In this paper, we explore an Information-Centric Networking (ICN) approach to in-network data aggregation, because of ICN's philosophy of breaking the dependence of data requesters on particular network hosts. Our demo will showcase how data aggregation can be performed by any ICN node and how it is matched to the requester's requirements, hence allowing applications to obtain personalized data aggregates while still using a generic-purpose networking stack.