Proceedings of the 5th workshop on Data management for sensor networks

摘要: Sensor networks aim to offer unprecedented means of monitoring the physical world, thus enabling entirely new applications. Many areas science contribute research on sensor networks, for which reason many conferences exist that either cover or are devoted solely networks. The International Workshop Data Management Networks series, was inaugurated in 2004 and this workshop is fifth edition, stands out as a unique forum early innovative work data management fills gap in-between database other network areas. The scope DMSN'08 covers all important aspects management, including acquisition, processing, storage remote wireless networks; handling uncertain data; heterogeneous sometimes sensitive databases. resource-constrained, lossy, noisy, distributed, nature implies traditional techniques often cannot be applied without significant retooling. Challenges associated with acquiring, archiving large-scale, sets live also call novel techniques. inherently incomplete noisy further calls cleaning, inference, approximation. Finally, applications, collecting raises privacy security concerns require protection anonymization techniques. DMSN'08 received 20 submissions, each assigned three four members program committee. Based reviews discussions among committee members, 8 papers were accepted inclusion these proceedings presentation at workshop. The grouped into sessions, first in-network aggregation. With purpose reducing communication costs, Baljeet et al. study use tree topologies based dominating forwarding aggregation when computing MAX queries. Next, Cho propose new, so-called partial ordered capable exploiting spatial correlation readings performing top-k monitoring. Kontaki distributed solution d-hop k-data coverage query generalizes previously considered queries. The second session processing trade-offs involve energy. In particular, Tang Cao data-driven power framework, accuracy latency can traded improved energy efficiency. Trajcevski trading balancing consumption network. The third various complex processing. First, Mihaylov consider joins integration ad hoc streams systems. Evers associate time intervals rather than points then two Hidden Markov Models, where value may extend across multiple hidden states, context. Their focus inference algorithms models. Karpinski Cahill end by proposing stream-based language targeted specifically programming encompassing both sensors actuators. The concludes panel. panel, well-known researchers community present discuss specific applications technologies, well notable technological challenges posed As such, panel contributes setting directions field.