- In trials, most CSPs achieved 100Mbps in downstream, at distances up to 400m.
- 40Mbps in upstream provides a strong differentiator for competing with cable.
- Vectoring equalizes performance across all lines (good and bad).
Communication Service Providers around the world are under intense pressure from both customers and government regulators to provide high-speed broadband as quickly as possible. While fiber-to-the-home (FTTH) remains the ultimate goal, the length of time it takes to deploy has left CSPs searching for a near-term alternative. Trials with more than 40 CSPs around the world have demonstrated that VDSL2 vectoring technology is a highly viable option.
The magic number: 100Mbps
The first thing every CSP wants to know is: “Can we really achieve 100Mbps”? That’s no surprise, since many of them are faced with fierce cable competitors who typically advertise very high speeds, although shared. Rates of 100Mbps allow CSPs to deliver a competitive service for the foreseeable future — and meet various national broadband targets, which typically call for 100Mbps by 2020. [caption id="attachment_3288" align="aligncenter" width="688"] Figure 1 — Downstream bit rate results from 27 trials
(Dark blue = prototype trials in 2010-11. Light blue = trials with commercial equipment in 2012-2013.)[/caption] Alcatel-Lucent has conducted more than 40 trials with various CSPs around the world. Figure 1 shows results from the first 27 trials, and demonstrates that 100Mbps is achievable over copper at distances up to 400m. These trials focused mostly on 300m to 600m loop lengths, reflecting the “sweet spot” for Vectoring gains. Because trial results cover both early prototype trials and commercial equipment trials and results improved significantly because of this, color coding is used to distinguish between the two. Some of the early trials, beginning in July, 2010, were capped at 100Mbps due to technology and prototype limitations. But the most recent trials, in 2012-13, achieved downstream bit rates between 100Mbps and 130Mbps. These improvements are partly explained by the switch from prototypes to commercial hardware and software. In addition, G.inp coding now provides an improved retransmission-based error-handling mechanism for VDSL2 & VDSL2 Vectoring which results in significantly lower overheads. By replacing the traditional Forward Error Correction (FEC) mechanism, G.inp avoids the FEC-associated fixed overhead of about 12 percent in typical cases, and G.inp only kicks in when there’s actually an error to correct. Wire diameter and cable type do have an impact on the results and account for most of the variation (between trials performed with commercial equipment) in Figure 1. In general, CSPs can expect to see very good vectoring gains whatever cable type they use. The good news is, vectoring delivers greater gain on poorer cable that is heavily impacted by crosstalk, because the crosstalk can be canceled by vectoring.
Upstream bandwidth beats cable
The second question CSPs ask is: “What about upstream bandwidth?” This is important for enabling more upstream-intensive services such as cloud applications or online storage and can also be a key differentiator in the competition with cable. So, given a downstream bit rate of 100Mbps, what upstream rates can CSPs expect with VDSL2 vectoring? Figure 2 shows upstream and downstream bit rates from fourteen CSP trials, representing loop lengths from 200m to 600m. Typical upstream performance was up to 40Mbps, for downstream bit rates at 100Mbps.Upstream bit rates benefit twice from VDSL2 vectoring. First, VDSL2 vectoring cancels the crosstalk in the upstream direction, which increases the bit rate. In addition, VDSL2 vectoring allows CSPs to relax their upstream power back off (UPBO) settings, which can boost upstream bit rates from the 20Mbps-range to the 40Mbps-range. This second benefit is illustrated in Figure 2 by the customer trial #7 performed at 400m. All the 20Mbps-range results in Figure 2 involve tests with traditional UPBO settings, while the 40Mbps-range results represent tests with UPBO relaxed. Using UPBO is standard in today’s VDSL2 networks, since it is necessary for managing upstream crosstalk. UPBO works by reducing the upstream transmit power on very short VDSL2 lines, so they don’t interfere with weaker upstream signals on longer loops. However, since VDSL2 vectoring eliminates crosstalk, UPBO settings can be relaxed.
Combining VDSL2 vectoring and bonding for increased rate or reach
CSPs with two pairs available for at least some of their subscribers are thinking about combining VDSL2 vectoring and bonding. Bonding gives CSPs a choice: they can double the bit rate for an existing subscriber; or they can deliver the same target bit rate over longer distances, thus reaching more subscribers. In customer trials with bonding plus VDSL2 vectoring, we achieved 200Mbps downstream and 50Mbps upstream at 400m. Bonding is largely a tactical solution, as it requires multiple pairs to each subscriber. But even with only 1.3, 1.5 or 1.8 pairs (on average) available per household, bonding can still allow additional, more remote subscribers to be connected to a cabinet. As a result, there is an increasing interest in bonding, even from CSPs who don’t have two pairs for every subscriber. And this interest is reflected in the wider availability of VDSL2 bonding CPEs.
The main benefit of VDSL2 vectoring is that it boosts bandwidth on copper lines – both downstream, and upstream. However, a second important benefit of VDSL2 vectoring is that it equalizes performance across lines. By cancelling cross-talk, it brings all lines very close to the single line rate. As such, VDSL2 Vectoring is a very “socialist” technology: poor lines gain the most, and all lines end up performing at more or less the same level. VDSL2 Vectoring delivers consistent and predictable performance across all lines (of similar length), by removing unpredictable cross-talk and making loop attenuation (and thus loop lengths) the determining factor. This allows operators to market and deliver very high bitrates to all subscribers - a significant advantage compared to ADSLx and unvectored VDSL2 technology, where line performance could vary greatly depending on the amount of cross-talk each line was subjected to.Figure 3 shows downstream bitrates across all lines in a 350m cable bundle. Without vectoring, downstream bitrates on the VDSL2 lines exhibit a significant spread – from 57Mbps to 88Mbps. With vectoring, all lines consistently reach between 101Mbps and 107Mbps. Vectoring will allow this service provider to increase his advertized bitrates from 55Mbps to 100Mbps (for users at up to 350m), whilst being able to maintain bitrates as take rate increases.
USING Zero-Touch Vectoring when mixing vectored and legacy VDSL2 lines
When introducing vectoring, many cable bundles will initially contain a mix of vectored lines and legacy VDSL2 lines. Operators need to take special care to avoid the impact from legacy VDSL2 lines on vectoring performance. Connecting VDSL2 lines to the vectoring system, and upgrading legacy VDSL2 CPE to vectoring-friendly or vectoring mode, removes any risk of interference from legacy VDSL2 CPE. However, not all legacy CPE can be upgraded to vectoring (or vectoring-friendly) mode. In addition, even if the CPE allows it, a network-wide CPE upgrade still presents an operational challenge and can take time – potentially delaying the introduction of vectoring services. Zero-Touch Vectoring technology automatically takes care of legacy VDSL2 lines – allowing operators to start rolling out vectoring, and to replace or upgrade the legacy CPE at their own pace. Customer trials demonstrate that Zero-Touch Vectoring works. Figure 4 shows downstream bitrates in a bundle with 8 legacy VDSL2 lines with Zero-Touch Vectoring enabled (shown in red) and 8 vectoring lines (shown in blue). Thanks to Zero-Touch Vectoring, the vectoring gain can be maintained in downstream, with all 8 lines achieving bitrates of 110Mbps at 300m. In addition, we again see the equalizing effect of vectoring: while there is significant variation in the VDSL2 line performance, the vectoring performance is consistent. Zero-Touch Vectoring does not limit bitrates on legacy VDSL2 lines, as is the case with DSM-based approaches.
Building on the promise
VDSL2 vectoring delivers as promised: 100Mbps and more downstream, with up to 40Mbps upstream – and all lines (of similar length) deliver the same high speeds. These rates provide DSL service providers with strong differentiators in their competition with cable, while enabling them to reach the broadband targets defined in various national broadband agendas and plans. With more than enough bandwidth available for the foreseeable future, the most pressing challenge is now mass deployment of VDSL2 vectoring. Zero-Touch Vectoring removes one of the big hurdles to deploying vectoring by automatically handling legacy VDSL2 lines and delivering 100Mbps on vectoring lines even when legacy VDSL2 lines are present in the binder. Editor's Note: The authors would like to thank Paul Spruyt, Bruce Orr and Courtenay Treleaven for their contribution to this article. To contact the authors or request additional information, please send an e-mail to firstname.lastname@example.org.