Uploading and Replicating Internet of Things (IoT) Data on Distributed Cloud Storage

Abstract

The Internet of Things (IoT) phenomenon is creatinga world of billions of connected devices generating enormousamounts of data items. In particular, for purposes of highavailability and disaster recovery, replication of this data on cloudstorage needs to be implemented efficiently. To that end, in thispaper, we investigate the combined problem of uploading IoTdata from a set of sensor gateways and efficient replication ofthis data on distributed cloud storage. While other efforts doexist in this space, either they do not consider replication incontext of small size and high number of data items, which isinherent nature of IoT data or they concentrate on the access timeafter replication. Our solution assumes the existence of multipledistributed cloud data centers, called mini-Clouds, among whichdata can be replicated. We model our problem comprehensivelybased on various parameters such as effective bandwidth of theIoT network, available number and size of data items at eachmini-Cloud, and we present our problem as a collection of varioussub-problems based on subsets of these parameters. We provethat the exact solution to the problem is intractable, and wepresent a number of heuristic strategies to solve it. Our resultsshow that the performance of any heuristic is bounded by theread and write latency of mini-Clouds and the best we can do isoften 12 times the worst we can do for a given number of dataitems to be uploaded and replicated from the gateways to themini-Clouds in test setup.