VDSL2 vectoring can boost copper speeds up to 100Mbps and beyond, but it is complicated and impractical with physical sub-loop unbundling. Bitstream access is the easy alternative to maintain vectoring gains in multi-operator deployments.
Vectoring and unbundling in a nutshell
There is considerable confusion in the market regarding deployment of vectoring in an unbundled or multi-operator environment. In some cases, the issues have been downplayed, while in others they have been exaggerated.
Therefore, when assessing the applicability of vectoring for a network, it’s important to know which unbundling scenarios pose problems and which do not. In brief, vectoring is fully compatible with the widely used Asymmetric Digital Subscriber Line (ADSLx) Local Loop Unbundling (LLU) from the Central Office (CO). However, in a very high-speed digital subscriber line 2 (VDSL2) Sub-loop Unbundling (SLU) scenario, significant vectoring gains cannot be guaranteed unless all VDSL2 lines in the cable are controlled by the same vectoring system.
To maintain the benefits of VDSL2 vectoring, the best solution is to have a single operator deploy VDSL2 from a remote location (for example, a cabinet), control all the lines, and offer bitstream access to other operators. Today’s “next-generation bitstream access” includes the flexibility and control needed to differentiate service offerings. And none of the alternatives that have been proposed, such as alien crosstalk cancelation and dynamic spectral management, can reach bit rates similar to vectoring. Vectoring across nodes is technically feasible but would require that all access providers use digital subscriber line access multiplexer (DSLAM) equipment from the same vendor – and it introduces operational complexity.
Vectoring requires control over all lines
Vectoring enables higher bit rates over copper by canceling crosstalk between all VDSL2 lines in a telephony cable. This is achieved through the continuous monitoring of crosstalk coupling within a cable, and the real-time generation of “anti-noise” that cancels out this crosstalk between all pairs. As a result, much higher bit rates can be achieved. Vectoring also makes bit rates more predictable because vectoring makes every line – whether heavily impacted by crosstalk or not — perform as if it were the only active line in the cable, that is, without crosstalk. Figure 1 shows VDSL2 bit rates as a function of loop length without vectoring and with vectoring.
For best results, the vectoring system must be able to monitor and control all VDSL2 lines in the cable. In addition, these lines all need to operate in vectoring mode or in vectoring-friendly mode.
The crosstalk from any uncontrolled VDSL2 lines, referred to as “alien” lines, will remain uncanceled, reducing vectoring gain (Figure 2). The more alien lines there are in a cable, the lower the gain. The exact impact is hard to predict and varies from line to line. For example, if the strongest disturbers for a “victim” VDSL2 line are uncontrolled, the loss will be significant. On the other hand, if the uncontrolled lines are only weak disturbers for a victim line, the loss will be limited.
The effects of physical loop unbundling
Where regulations require physical unbundling, providers may not be able to control all lines in a cable. To understand how this affects vectoring, it is important to distinguish between two different types of unbundling. The first is ADSLx unbundling, which is typically LLU from the CO. The second is VDSL2 unbundling, typically SLU from the cabinet. Both types are illustrated in Figure 3.
ADSLx LLU – small impact. ADSLx LLU at the CO is widely used today. When fiber to the node (FTTN) deployments start, ADSLx lines from CO from multiple operators will co-exist with VDSL2 lines from the cabinet in the distribution cable beyond the cabinet. This scenario does not pose a problem: ADSLx unbundling from the CO only has a small effect on vectoring gains because most of the gain is realized above the ADSLx spectrum.
In fact, the challenges related to mixing ADSLx from the CO with VDSL2 from the cabinet have long been solved by shaping the VDSL2 Power Spectral Density (PSD) in the overlapping frequency band. This technique, known as Downstream Power Back-off (DPBO), is a standard operating practice in VDSL2 deployments. Vectoring relies on the same standard VDSL2 mechanism to minimize impact between ADSLx and vectored VDSL2. Typically, ADSLx unbundling from the CO will result in a decrease of 10 percent or less in vectoring bit rates (Figure 4).
VDSL2 SLU – unpredictable, potentially high impact. VDSL2 unbundling can cause a significant decline in vectoring gains, because crosstalk between VDSL2 lines from different providers and terminated on independent DSLAMs cannot be cancelled. As the VDSL2 lines use the same frequency spectrum, there is no way to avoid crosstalk through PSD shaping without significant bit rate loss.
As a general rule, VDSL2 SLU is not widely used today. As fiber moves closer to the end user, collocation of VDSL2 equipment from different vendors at the same remote location becomes less practical, and the business case becomes less attractive. Even in the few cases where it is applied, there is typically only a single operator offering VDSL2 services from that remote location, again allowing an upgrade to vectoring.
As shown in Figure 5, with VDSL2 SLU the actual impact on vectoring gains is unpredictable, but is likely to be significant. It depends on the number of lines controlled by other operators and on the amount of crosstalk generated by those lines (which is determined by the relative positions of the different lines in the cable section).
Solving the vectoring/unbundling challenge
The ability to wholesale vectored lines is key to ensuring fair competition and to complying with regulatory obligations. The obvious remedy is to offer bitstream access. That way, all VDSL2 lines remain under the control of a single access provider, and all operators (the owner of the VDSL2 lines and the wholesale customers) can enjoy the full vectoring gain.
But are any other solutions available besides bitstream access? Over the past couple of months, various solutions have been suggested; however, there is no “silver bullet.” All these proposed solutions have significant drawbacks. Here’s a closer look at the possibilities:
- “Ignore the problem and take whatever gains you get.” Technically, it is possible to apply vectoring in a scenario with VDSL2 SLU and simply be content with whatever bit-rate gains one happens to get. However, with this approach, the gains become totally unpredictable. One might get 20 Mbps extra on a VDSL2 line with low alien VDSL2 crosstalk and zero gain on another line of similar length with dominant alien crosstalk. And the gains may change at any moment, whenever conditions change. For example, adding a new VDSL2 subscriber — or moving a subscriber to another operator — can cause significant bit rate changes on other lines in the same cable. Because the extra bit rates cannot be guaranteed, customers won’t be willing to pay more.
- Use alien crosstalk cancellation. This approach is sometimes referred to as “SuperMIMO” (multiple input multiple output) and only works if multiple pairs are available per subscriber. It offers useful gains only when the number of “alien” VDSL2 lines is very limited (a single line or a few lines maximum). These limitations render SuperMIMO useless in practical deployments with SLU.
- Use dynamic spectral management (DSM) to mitigate crosstalk between VDSL2 lines from different operators. DSM can reduce interference between VDSL2 lines in the same cable. But it can only provide substantial gains on one set of lines at the expense of a reduction in achievable bit rate on other lines. In general, in an unbundled environment DSM cannot offer bit rates comparable to vectoring. In addition, DSM raises the practical and regulatory issue of who would control the central DSM entity, managing lines from different operators and deciding on the bit rate trade-offs between lines.
The solution: Bitstream access
When all factors are considered, bitstream access offers the most feasible approach with proven results. It is important to understand that the concept of bitstream access has evolved significantly, offering enhanced flexibility and control compared to initial definitions. New capabilities include multicast and quality of service (QoS) control that give wholesale customers better means to differentiate their services from competitors. Such next-generation bitstream access solutions are sometimes also referred to as “virtual unbundling”.
In many countries, regulators are reviewing unbundling obligations in the context of vectoring. After market analysis in Belgium, the national regulator (BIPT) decided to withdraw the obligation for VDSL2 SLU. This 2011 decision was based on the recognition that SLU would render vectoring ineffective and a study by Analysys Mason showing the extremely limited economic viability of SLU in Belgium. The debate is ongoing in other countries.
Vectoring across DSLAM systems from different access providers
The Alcatel-Lucent vectoring solution allows crosstalk to be canceled across lines that are connected to collocated mini-DSLAMs, allowing an operator to add extra nodes as the take rate increases. While conceptually this could also be used to vector across DLSAMs from different providers in a multi-operator deployment, this would pose serious challenges. There is currently no standard describing the vectoring interface between such systems, and differences in DSLAM architecture and implementation make standardization unlikely. As a result, “Multi-System Vectoring” can only work if the different service providers deploy systems from the same equipment vendor. Moreover, Vectoring across DSLAMs owned and managed by different operators also introduces significant operational complexity, e.g. for software upgrades, troubleshooting and guaranteeing QoS. Bitstream access avoids these challenges.
Deployment in multi-operator scenarios
Vectoring can be used effectively in multi-operator scenarios. It is fully compatible with the widely used ADSLx local loop unbundling (LLU) from the CO, which reduces vectoring gains only slightly. On the other hand, VDSL2 vectoring gains cannot be realized unless all VDSL2 lines in the cable are controlled by the same vectoring system. Therefore, only a single operator should deploy VDSL2 from a given remote location, control all the lines — and offer bitstream access to other operators. Vectoring across nodes is technically feasible if the different operators agree to deploy the same DSLAM equipment, but complicates operations.
Today’s “next-generation bitstream” or virtual unbundling enables everyone to enjoy full vectoring bit-rate gains, along with the flexibility and control needed to differentiate service offerings. None of the alternative solutions, such as alien crosstalk cancelation and dynamic spectral management, can deliver bit rates similar to vectoring.
To contact the authors or request additional information, please send an email to email@example.com.