KRAFT: A Real-Time Active DBMS for Signal Streams

The applications of ubiquitous sensor networks require database system to support the following three functions in addition with conventional database functions. (1) continual event monitoring. Since control systems such as robots perform accurately, event monitoring must be executed in strict real-time. (2) signal processing. To recognize events in the physical world, sensor data must be processed by non traditional way such as similar sequence retrievals. (3) fast signal stream persisting. All of sensor data should be stored to consider the reason of illegal events after accidents or offline data mining. To support the requirements, we propose a new database system KRAFT. To realize (1), KRAFT controls user-level threads on FreeBSD KSE scheduler. To realize (2), KRAFT provides similar sequence retrieval operators. The operators' distance functions are Euclidean and dynamic time warping. To realize (3), KRAFT provides direct persisting, which does not execute the write ahead logging process. We describe preliminary results of experiments and show the performance of KRAFT.

[1]  Hideyuki Kawashima,et al.  RSV: sensor data viewer for human-robot interaction , 2007 .

[2]  Andreas Reuter,et al.  Transaction Processing: Concepts and Techniques , 1992 .

[3]  Hideyuki Kawashima,et al.  Improving Freshness of Sensor Data on KRAFT Sensor Database System , 2004, Multimedia Information Systems.

[4]  Wei Hong,et al.  The design of an acquisitional query processor for sensor networks , 2003, SIGMOD '03.

[5]  Frederick Reiss,et al.  Declarative Network Monitoring with an Underprovisioned Query Processor , 2006, 22nd International Conference on Data Engineering (ICDE'06).

[6]  Yong Yao,et al.  Network scheduling for data archiving applications in sensor networks , 2006, DMSN '06.

[7]  Michael Stonebraker,et al.  High-availability algorithms for distributed stream processing , 2005, 21st International Conference on Data Engineering (ICDE'05).

[8]  Hideyuki Kawashima,et al.  PORSCHE: A physical objects recommender system for cell phone users , 2006 .

[9]  Eamonn J. Keogh,et al.  Scaling up dynamic time warping for datamining applications , 2000, KDD '00.

[10]  Tetsuo Ono,et al.  Development and evaluation of an interactive humanoid robot "Robovie" , 2002, Proceedings 2002 IEEE International Conference on Robotics and Automation (Cat. No.02CH37292).

[11]  Sang Hyuk Son,et al.  Prediction-Based QoS Management for Real-Time Data Streams , 2006, 2006 27th IEEE International Real-Time Systems Symposium (RTSS'06).

[12]  Jörgen Hansson,et al.  Specification and management of QoS in real-time databases supporting imprecise computations , 2006, IEEE Transactions on Computers.

[13]  Ryan Newton,et al.  The Case for a Signal-Oriented Data Stream Management System , 2007, CIDR.

[14]  Krithi Ramamritham,et al.  Mutual Consistency in Real-Time Databases , 2006, 2006 27th IEEE International Real-Time Systems Symposium (RTSS'06).

[15]  Susan V. Vrbsky.  A Data Model for Approximate Query Processing of Real-Time Databases , 1996, Data Knowl. Eng..

[16]  Hideyuki Kawashima,et al.  Providing Persistence for Sensor Data Streams by Remote WAL , 2006, DaWaK.

[17]  Ying Xing,et al.  A Cooperative, Self-Configuring High-Availability Solution for Stream Processing , 2007, 2007 IEEE 23rd International Conference on Data Engineering.

[18]  S. Satake,et al.  Brownie: Searching Concealed Real World Artifacts , 2007, 2007 Fourth International Conference on Networked Sensing Systems.

[19]  Hideyuki Kawashima,et al.  MeT: a real world oriented metadata management system for semantic sensor networks , 2006, DMSN '06.

[20]  Brian N. Bershad,et al.  Scheduler activations: effective kernel support for the user-level management of parallelism , 1991, TOCS.