Article Preview
TopIntroduction
Ideally, Optical Fiber should be the technology to be deployed to reaching the end-users (Fiber to the Home -FTTH-). However, the investment and maintenance costs are still prohibitive for some countries as indicated by Cioffi (2007).
Digital subscriber line (DSL) technology is currently the most widely deployed broadband access technology and will continue to play an important role during the coming years. In a previous report from Point-Topic by Vanier (2009), it is mentioned that the number of DSL subscribers adds up to ~65% of all worldwide broadband access technologies (being followed by ~21% of cable modem subscribers).
However, to cope with the bandwidth-intensive and a mixed set of quality-of-service (QoS) and quality-of-experience (QoE) requirements of the many emerging broadband applications and services (i.e., VoIP, triple-play services including HDTV, IPTV, video-conferences, etc.), it is essential to further improve DSL technology.
One of the major sources of performance degradation of current DSL systems is the electromagnetic coupling among different twisted pairs within the same cable bundle (also referred to as crosstalk). The presence of crosstalk transforms DSL systems into a very challenging multi-user multicarrier interference environment, where different users, i.e., modems, can significantly impact each other’s datarate transmission and the quality of the line; thus, having a direct impact on the final end-user experience.
Spectrum coordination management consists of coordinating the transmit spectrum of each modem so as to prevent (or substantially mitigate) the impact of crosstalk; nowadays, relatively new techniques like the different levels of DSM (single-user coordination, multi-user, vectoring, etc.) as proposed in the literature (Cioffi, 2007; Cioffi & Mohseni, 2004; Song et al., 2002; Ginis & Cioffi, 2002; Cioffi et al., 2006; Yu, Ginis, & Cioffi, 2002; Cendrillon, Moonen, Verlinden, Bostoen, & Yu, 2004; Cendrillon et al., 2006; Yu et al., 2004; Tsiaflakis et al., 2005; Papandriopoulus & Evans, 2009; Cendrillon & Moonen, 2005; Tsiaflakis et al., 2008) allow increasing dramatically the bitrates over copper lines, up to (or even higher than) fiber capacities.
One of the first proposed DSM algorithms (Yu, Ginis, & Cioffi, 2002) was iterative waterfilling (IWF) where it was demonstrated a significant performance improvement over traditional (let say, static spectrum management) multiuser DSL lines. This algorithm iteratively measures the aggregate interference received from all other users, greedily water-pouring their own power allocation. However, this approach disregards the impact on other users.
Cendrillon et al. (2004, 2006), proposed an Optimum Spectrum Balancing (centralized) algorithm, therefore assuming centralized control in a Spectrum Management Center (SMC). OSB applies dual decomposition (Yu, Lui, & Cendrillon, 2004) making use of Lagrange multipliers to enforce constraints that are coupled over the tones to find the optimal power allocation to a predetermined quantum and to lower the complexity from exponential to linear in the number of tones. As this algorithm considers the damage done to other modems, OSB avoids a selfish optimum, thereby significantly improving its performance over IWF. However, OSB remains difficult to implement over practical DSL lines though many attempts to reduce this complexity appear in Cendrillon et al. (2004, 2006).