Asymptotic Analysis of the Performance of DSL Systems

DSL is a family of technologies that provide internet access by transmitting digital data simultaneously with wired telephone service on the same telephone line (a copper twisted pair). The main limitation of modern DSL systems is ”crosstalk”, the interference caused by electromagnetic coupling of transmissions over other wire pairs.

In our work we derive an approximations to various performance measures, such as signal to interference plus noise ratio (SINR) and spectral efficiency, of a zero-forcing (ZF) receiver in a DSL system. These approximations are extracted by an asymptotic analysis, where we assume that the system size (i.e., the number of transmitters and receivers) grows. Through this method, random system parameters, which are quiet complex by their nature, can be replaced by deterministic equivalents, in order to achieve simple approximation terms. For this purpose, we rely on works in the field of large dimensional random matrices. 

We also provide numerical simulations which show that this approximations are accurate even for relatively small system dimensions.

 DSL is a family of technologies that provide internet access by transmitting digital data simultaneously with wired telephone service on the same telephone line (a copper twisted pair). The main limitation of modern DSL systems is ”crosstalk”, the interference caused by electromagnetic coupling of transmissions over other wire pairs.

In our work we derive an approximations to various performance measures, such as signal to interference plus noise ratio (SINR) and spectral efficiency, of a zero-forcing (ZF) receiver in a DSL system. These approximations are extracted by an asymptotic analysis, where we assume that the system size (i.e., the number of transmitters and receivers) grows. Through this method, random system parameters, which are quiet complex by their nature, can be replaced by deterministic equivalents, in order to achieve simple approximation terms. For this purpose, we rely on works in the field of large dimensional random matrices. 

We also provide numerical simulations which show that this approximations are accurate even for relatively small system dimensions.

* This research was carried out towards the Masters degree in Electrical Engineering at Bar-Ilan University, under the supervision of Dr. Itsik Bergel