Chapter 10. Spectrum Management of DSL Systems
The telco network was originally designed for the deployment of voice grade services. In these applications the subscriber line signals utilized frequencies below 4 kHz (approximately 200 Hz to 3.3 kHz). With the development of voiceband modems, this same voice channel was used to transmit bidirectional data up to 56 kb/s per ITU-T Recommendation V.90 [1]. Clearly, the 4 kHz band of the voice channel only utilizes a small portion of the subscriber line capacity. In the 1970s, the digital data system (DDS) was developed that transmitted symmetric data on the same twisted wire pair subscriber line, but the bandwidth was not restricted to the voice band. The line code was alternate mark inversion (AMI), and bit rates supported were 2.4, 4.8, 9.6, and 56 kb/s. Later in the mid 1980s, clear channel 64 kb/s access was defined where the actual bit rate on the subscriber line was 72 kb/s, which contained 8 kb/s of signaling overhead. A key difference to voice grade services is that the access for DDS required two twisted wire pairs for duplexing ; one wire pair transmits the downstream AMI signal and the other pair transmits the upstream signal [2]. Then in the 1980s, Basic Rate ISDN was developed. ISDN transports 160 kb/s on the access subscriber line using the 2B1Q (two binary one quarternary) line code, which is equivalent to 4-level pulse amplitude modulation (4-PAM). As we describe, ISDN is transmitted bidirectionally on a single twisted wire pair. Duplexing is performed using echo cancellation. Near-end crosstalk from other ISDN signals in the same cable limits the reach of ISDN [3]. In 1989, the study began on the feasibility of transmitting signals at a rate substantially higher than basic rate ISDN (160 kb/s). This study project became known as the high-rate digital subscriber line (HDSL). Although several modulation methods were considered for HDSL, the same 2B1Q line code was preferred by the industry for HDSL. The driving application for HDSL is the transmission of T1 service on loops without repeaters, and the performance objective was transmitting the1.544 Mb/s payload at carrier serving area (CSA) distances (9 kft of 26 gauge wire and 12 kft of 24 gauge wire) in the presence of near-end crosstalk from other like disturbers. To meet the performance objective, a dual duplex architecture was chosen where the 1.544 Mb/s payload was split into two parallel transmission channels, each carrying one half of the T1 payload. With the inclusion of overhead and payload, the bit rate transmitted on each wire pair of the dual-duplex HDSL is 784 kb/s. HDSL was thus targeted as a transport mechanism that would allow the operator an alternative for quickly deploying symmetric T1 grade service on two loops within CSA range without any repeaters. Provisioning of T1 service with the traditional repeatered AMI could take as long as two months if repeaters were required to be installed in the cable plant. Deploying the same T1 service using HDSL would reduce the provisioning time to as little as two to three days. T1 is a high-capacity, high-quality service targeted for business applications where HDSL is an operator option for provisioning T1 service [4], [5]. During the same time period, there was also interest from the phone companies to deploy video services to residential customers using the same subscriber line infrastructure. This is in response to the cable companies looking to deploy telephone services in their own coaxial cable infrastructure. Deployment of video service has an asymmetric transmission profile, namely, a high-capacity channel would need to be transmitted to the subscriber (downstream) whereas only a low bit “rate channel would need to transmitted back to the network (upstream) for program control. The application became known as the asymmetric digital subscriber line or simply ADSL. Video-on-demand (VOD) was the initial application driving ADSL. Unlike HDSL, which uses echo cancellation to duplex two fully overlapped upstream and downstream channels, the asymmetric nature of ADSL made use of frequency division duplexing (FDD) for transmission of the ADSL upstream and downstream channels. With FDD, the upstream channel uses a narrow band of frequencies for transmission of the low-speed upstream channel and a separate higher band of frequencies for the transmission of the downstream channel. In addition the two ADSL bands would be placed at frequencies above those of the voice channel so ADSL would support transmission of conventional plain old telephone service (POTS), a low-speed upstream channel for the transport of signaling data, and a high-speed downstream channel for the transport of digital video and signaling. ADSL achieves downstream bit rates far beyond the preceding DSL technologies; this was made possible by asymmetric bit rates that in turn made FDD practical. ADSL performance is further enhanced by the use of Trellis coding and Reed-Solomon coding. A key element of ADSL and it associated frequency division duplexing is that if each of the wire pairs in the cable is filled using the same upstream and downstream frequency bands, there is no self near-end crosstalk (SNEXT) present in either of the two channels because there is no overlap in the transmission of the upstream and downstream channels. With FDD only far-end crosstalk is present, which is significantly less disturbing than near-end crosstalk. Hence, if the cable plant could be managed to contain only FDD-based systems, then significantly higher bit rates (downstream or upstream) could be achieved than with echo- canceled systems. Unfortunately, it is difficult to manage or force such restrictions with competitive access in the cable plant, but limiting the amount of near-end crosstalk introduced in the cable plant is manageable. In the initial vision, ADSL was transporting approximately 1.6 Mb/s data downstream and 24 kb/s upstream to a distance of approximately 15 kft. The definition of ADSL was later expanded to include the transport of approximately 6 Mb/s downstream and 640 kb/s upstream. In the mid-1990s the driving application shifted from VOD to Internet access. The range of bit rates was still valid, where operators can provision different bit rates for a specified quality of service. The chosen modulation method for ADSL is discrete multi-tone (DMT), and the transmit power spectral density (PSD) for ADSL is defined in T1.413 Issues 1 and 2 and in ITU Recommendation G.992.1 [6, 7, 8]. An alternative modulation method (using carrierless amplitude and phase, CAP, modulation) was also deployed for rate adaptive ADSL (RADSL) systems prior to standards-based deployments. The corresponding PSD is defined in Committee T1 Technical Report No. 59 (TR-059) [9]. Note that RADSL uses the same upstream and downstream frequency bands, as does ADSL. Aside from the above standards-based services, there are proprietary or nonstandard systems, both symmetric and asymmetric, deployed in the same cable plant. Examples include multirate symmetric DSL (SDSL) that use 2B1Q and single carrier modulation methods. The bandwidth and spectral shaping of each modulation approach determines the degree of disturbance into other systems in the cable plant. There is also T1 AMI deployed in the distribution cable plant. This is known to be a very strong disturber into ADSL; so strong that it may even prohibit deployment of ADSL with any reasonable quality of service if encountered in the same cable at distances greater than 6 kft. In addition to the well-defined DSL service described earlier, there are other nonstandard systems deployed, and new, advanced DSL systems being developed. The impact of any newly defined DSL system on already deployed systems needs to be understood prior to providing large-scale deployments. If all of the systems, old and new, are to be deployed in the same cable, we first need to understand how they interfere with each other. This quantification of disturbance is called spectral compatibility. Spectral compatibility is the quantification of the level of disturbance that one type of DSL system has on other systems deployed in the same cable. A clear understanding of the interactions that the different types of systems have on each other in the cable is the foundation for managing the deployment of DSL systems in maintaining a specified quality of service. In the mass deployment of DSL services, we can expect the cable to be filled with many different types of signal spectra. Given the wide variety of signals in the cable, there can be a wide range of interference between the different systems such that a set of rules or guidelines are needed to manage the deployment of these systems to assure a minimum or specified quality of service. These rules for the deployment of systems in the loop plant are called spectrum management . The rules of spectrum management are based on a working knowledge of the spectral compatibility of the various DSL systems and an agreement on a desired quality of service. This chapter presents an overview of the spectrum management for the deployment of different DSL services in a single loop plant. Included are discussions on the fundamental concept of spectrum management per T1.417, along with background information leading to critical decisions in the spectrum management standard. |
| Top |
EAN: 2147483647
Pages: 154