Efficient pipeline architectures for Underwater Big data analytic

Big data are found in numerous underwater applications, such as Seismic Monitoring, Detection and tracking of marine objects applications, underwater multimedia systems, 3D maps of underwater sites, and many other applications. Such information is large in size and often generated continuously at fast rate. Current techniques used to collect data from underwater nodes relay on different approaches such as fixed wired link, AUV systems which usually incorporate RF waves, and wireless sensors networks. The first method is impractical in deep underwater sites while the second one suffers from high energy cost and short ranges. Underwater wireless sensors mitigate these previous limitations due to its cheap cost, limited power consumption and long distance coverage. Therefore as a technology, it has become a promising way to extract and collect underwater data. Nonetheless, underwater sensor networks suffer from low communication bandwidth and long propagation delay. Such factors limit the amount of transmitted data and reduce the efficiency for better bandwidth utilization. Therefore, it becomes crucial to develop systems that able to gather and process large underwater data sizes before or during the transmission stages (in-network processing). In this work, we develop a set of efficient pipeline architectures for Underwater Big data analytics. These architectures use pipelining to speed up data processing and transmission. The advantages of each architecture are highlighted as well as the operation of each one.