TY - GEN
T1 - Wireless link scheduling under a graded SINR interference model
AU - Santi, Paolo
AU - Maheshwari, Ritesh
AU - Resta, Giovanni
AU - Das, Samir
AU - Blough, Douglas M.
PY - 2009
Y1 - 2009
N2 - In this paper, we revisit the wireless link scheduling problem under a graded version of the SINR interference model. Unlike the traditional thresholded version of the SINR model, the graded SINR model allows use of"imperfect links", where communication is still possible, although with degraded performance (in terms of data rate or PRR). Throughput benefits when graded SINR model is used instead of thresholded SINR model to schedule transmissions have recently been shown in an experimental testbed. Here, we formally define the wireless link scheduling problem under the graded SINR model, where we impose an additional constraint on the minimum quality of the usable links, (expressed as an SNR threshold Q). Then, we present an approximation algorithm for this problem, which is shown to be within a constant factor from optimal. We also present a more practical greedy algorithm, whose performance bounds are not known, but which is shown through simulation to have much better average performance than the approximation algorithm. Furthermore, we investigate, through both simulation and implementation on an experimental testbed, the tradeoff between the minimum link quality threshold Q and the resulting network throughput.
AB - In this paper, we revisit the wireless link scheduling problem under a graded version of the SINR interference model. Unlike the traditional thresholded version of the SINR model, the graded SINR model allows use of"imperfect links", where communication is still possible, although with degraded performance (in terms of data rate or PRR). Throughput benefits when graded SINR model is used instead of thresholded SINR model to schedule transmissions have recently been shown in an experimental testbed. Here, we formally define the wireless link scheduling problem under the graded SINR model, where we impose an additional constraint on the minimum quality of the usable links, (expressed as an SNR threshold Q). Then, we present an approximation algorithm for this problem, which is shown to be within a constant factor from optimal. We also present a more practical greedy algorithm, whose performance bounds are not known, but which is shown through simulation to have much better average performance than the approximation algorithm. Furthermore, we investigate, through both simulation and implementation on an experimental testbed, the tradeoff between the minimum link quality threshold Q and the resulting network throughput.
UR - https://www.scopus.com/pages/publications/70450187040
U2 - 10.1145/1540343.1540346
DO - 10.1145/1540343.1540346
M3 - Conference contribution
AN - SCOPUS:70450187040
SN - 9781605585239
T3 - FOWANC'09 - Proceedings of the 2nd ACM International Workshop on Foundations of Wireless Ad Hoc and Sensor Networking and Computing, Co-located with MobiHoc'09
SP - 3
EP - 12
BT - FOWANC'09 - Proceedings of the 2nd ACM International Workshop on Foundations of Wireless Ad Hoc and Sensor Networking and Computing, Co-located with MobiHoc'09
T2 - 2nd ACM International Workshop on Foundations of Wireless Ad Hoc and Sensor Networking and Computing, FOWANC'09, Co-located with MobiHoc'09
Y2 - 18 May 2009 through 18 May 2009
ER -