| VDSL technology
resembles ADSL to a large degree, although ADSL must face much
larger dynamic ranges and is considerably more complex as a result.
VDSL must be lower in cost and lower in power, and premises VDSL
units may have to implement a physical-layer MAC for multiplexing
upstream data.
Line-Code Candidates
Four line codes have been proposed
for VDSL:
- Carrierless amplitude
modulation/phase modulation (CAP)--A version of suppressed
carrier quadrature amplitude modulation (QAM). For passive NT
configurations, CAP would use quadrature phase shift keying (QPSK)
upstream and a type of TDMA for multiplexing (although CAP does
not preclude an FDM approach to upstream multiplexing).
- Discrete multitone (DMT)--A
multicarrier system using discrete fourier transforms to create
and demodulate individual carriers. For passive NT
configurations, DMT would use FDM for upstream multiplexing
(although DMT does not preclude a TDMA multiplexing strategy).
- Discrete wavelet multitone (DWMT)--A
multicarrier system using wavelet transforms to create and
demodulate individual carriers. DWMT also uses FDM for upstream
multiplexing, but also allows TDMA.
- Simple line code (SLC)--A
version of four-level baseband signaling that filters the based
band and restores it at the receiver. For passive NT
configurations, SLC would most likely use TDMA for upstream
multiplexing, although FDM is possible.
Channel Separation
Early versions of VDSL will use FDM to separate downstream from
upstream channels and both of them from basic telephone service and
ISDN, as shown in Figure . Echo cancellation may be
required for later-generation systems featuring symmetric data
rates. A rather substantial distance, in frequency, will be
maintained between the lowest data channel and basic telephone
service to enable very simple and cost-effective basic telephone
service splitters. Normal practice would locate the downstream
channel above the upstream channel. However, the DAVIC specification
reverses this order to enable premises distribution of VDSL signals
over coaxial cable systems.
Forward Error Control
FEC will no doubt use a form of Reed Soloman coding and optional
interleaving to correct bursts of errors caused by impulse noise.
The structure will be very similar to ADSL, as defined in T1.414. An
outstanding question is whether FEC overhead (in the range of 8
percent) will be taken from the payload capacity or added as an
out-of-band signal. The former reduces payload capacity but
maintains nominal reach, whereas the latter retains the nominal
payload but suffers a small reduction in reach. ADSL puts FEC
overhead out of band.
Upstream Multiplexing
If the premises VDSL unit comprises the network termination (an
active NT), then the means of multiplexing upstream cells or data
channels from more than one CPE into a single upstream becomes the
responsibility of the premises network. The VDSL unit simply
presents raw data streams in both directions. As illustrated in
Figure , one type of premises network involves a star
connecting each CPE to a switching or multiplexing hub; such a hub
could be integral to the premises VDSL unit.
In a passive NT configuration, each
CPE has an associated VDSL unit. (A passive NT does not conceptually
preclude multiple CPE per VDSL, but then the question of active
versus passive NT becomes a matter of ownership, not a matter of
wiring topology and multiplexing strategies.)
Now the upstream
channels for each CPE must share a common wire. Although a
collision-detection system could be used, the desire for guaranteed
bandwidth indicates one of two solutions. The first invokes a
cell-grant protocol in which downstream frames generated at the ONU
or farther up the network contain a few bits that grant access to
specific CPE during a specified period subsequent to receiving a
frame. A granted CPE can send one upstream cell during this period.
The transmitter in the CPE must turn on, send a preamble to
condition the ONU receiver, send the cell, and then turn itself off.
The protocol must insert enough silence to let line ringing clear.
One construction of this protocol uses 77 octet intervals to
transmit a single 53-octet cell.
The second method divides the
upstream channel into frequency bands and assigns one band to each
CPE. This method has the advantage of avoiding any MAC with its
associated overhead (although a multiplexor must be built into the
ONU), but either restricts the data rate available to any one CPE or
imposes a dynamic inverse multiplexing scheme that lets one CPE send
more than its share for a period. The latter would look a great deal
like a MAC protocol, but without the loss of bandwidth associated
with carrier detect and clear for each cell.
|