Providing Persistence for Sensor Data Streams with Temporal Consistency Conscious Wal

H. Kawashima, M. Imai, M. Toyama, and Y. Anzai (Japan)


Sensor Data Streams, Temporal Consistency of Sensor Data, Data Persistence, Temporal Consistency Conscious WAL


Sensor database systems need to provide both temporal consistencies of data and persistence to the incoming sensor data streams. For persistence, disk based logging methods have been widely used, however they are not applicable for this purpose because they are awfully tardy. In this paper, we propose the Temporal Consistency Conscious-Write Ahead Logging(T-WAL). The T WAL is a memory logging method and is a reļ¬nement of the Neighbor-Write Ahead Logging(N-WAL). To accelerate logging speed, the T-WAL utilizes temporal consistency constraints to transfer some records together, reveals under looser protocol than the two phase commit protocol and uses private network for logging to avoid congestion. As the looser protocol weakens the guarantee of logging successes, we incorporate a repair system and a checker system to complement the degree of the guarantee. Experiment results showed that the through put of the T-WAL is approximately 2.05 times greater than the N-WAL when the number of concurrent sensor data streams is 225.

Important Links:

Go Back