Improved Recovery Management and Routing in W-Grid, a Distributed Infrastructure for Effective and Efficient Multidimensional Data Management over Wireless Ad-Hoc Sensor Networks - Data Management in Cloud, Grid and P2P Systems - page 81

$ "" # !

"!

Fig. 5. Query APL in a Network with an Average of 8 Neighbors per Node

$!##!"

!

!#" !

Fig. 6. Query APL in a Network with an Average of 4 Neighbors per Node

Where lazy recovery is performed whenever a node could not solve a failure
situation with the active recovery. We present the results obtained with both
strategies.
In each Figure 7, 8 and 9 we compare W-Grid eciency with coordinates
dependencies against the W-Grid solution exploiting message broadcast. Here,
broadcast feature has been tried with different Time-To-Live (TTL) values and
obviously the ecacy, not the eciency, improves as TTL value increases. The
fifth curve represents an unlimited broadcast which has been simulated whenever
W-Grid could not be able to perform recovery. Figures show that almost every
time that W-Grid was not able to perform recovery, unlimited broadcast was
not able as well, meaning that preceding W-Grid version failed just because the
network was partitioned due to device failure.

"$!

#

Fig. 7. Recovery Success Ratio with an Average of 4 Neighbors per Node