Izvorni znanstveni članak
https://doi.org/10.24138/jcomss.v3i3.250
Cross-layer performance control of wireless channels using active local profiles
D. Moltchanov
; Network and Protocols group, Institute of Communication Engineering, Tampere University of Technology, Tampere, Finland
Sažetak
To optimize performance of applications running over wireless channels state-of-the-art wireless access technologies incorporate a number of channel adaptation mechanisms. While these mechanisms are expected to operate jointly providing the best possible performance for current wireless channel and traffic conditions, their joint effect is often difficult to predict. To control functionality of various channel adaptation mechanisms a new cross-layer performance optimization system is sought. This system should be responsible for exchange of control information between different layers and further optimization of wireless channel performance. In this paper design of the cross-layer performance control system for wireless access technologies with dynamic adaptation of protocol parameters at different layers of the protocol stack is proposed. Functionalities of components of the system are isolated and described in detail. To determine the range of protocol parameters providing the best possible performance for a wide range of channel and arrival statistics the proposed system is analytically analyzed. Particularly, probability distribution functions of the number of lost frames and delay of a frame as functions of first- and second-order wireless channel and arrival statistics, automatic repeat request, forward error correction functionality, protocol data unit size at different layers are derived. Numerical examples illustrating performance of the whole system and its elements are provided. Obtained results demonstrate that the proposed system provide significant performance gains compared to static configuration of protocols.
Ključne riječi
wireless channels; non-stationarity; cross-layer performance control; system design; performance evaluation
Hrčak ID:
180586
URI
Datum izdavanja:
25.9.2007.
Posjeta: 700 *