Institutional Repository
Technical University of Crete
EN  |  EL



My Space

Probabilistic data management for pervasive computing: the data furnace project

Garofalakis Minos, Brown Kurt P., Franklin Michael J., Hellerstein, Joseph, 1952-, Wang Daisy Zhe, Michelakis Eirinaios, Tancau Liviu, Wu Eugene, Jeffery Shawn R., Aipperspach Ryan

Full record

Year 2006
Type of Item Conference Publication
Bibliographic Citation M. Garofalakis, K. P. Brown, M. J. Franklin, J. M. Hellerstein, D. Z. Wang, E. Michelakis, L. Tancau, E. Wu, S. R. Jeffery and R. Aipperspach, "Probabilistic data management for pervasive computing: the data furnace project", in IEEE Data Engineering Bulletin, vol. 29, no. 1, March 2006.
Appears in Collections


The wide deployment of wireless sensor and RFID (Radio Frequency IDentification) devices is one of the key enablersfor next-generation pervasive computing applications, including large-scale environmental monitoring andcontrol, context-aware computing, and “smart digital homes”. Sensory readings are inherently unreliable andtypically exhibit strong temporal and spatial correlations (within and across different sensing devices); effectivereasoning over such unreliable streams introduces a host of new data management challenges. The Data Furnaceproject at Intel Research and UC-Berkeley aims to build a probabilistic data management infrastructure for pervasivecomputing environments that handles the uncertain nature of such data as a first-class citizen through aprincipled framework grounded in probabilistic models and inference techniques.