Two schools of vectoring thought – Use board or system level?
There has been a great deal of discussion recently regarding vectoring, a technology that can help extend the lifespan of today’s copper networks by boosting their performance to near fiber speeds.
Historically, copper has been widely deployed globally, much more so than fiber. Vectoring is an important development since the copper infrastructure has always been susceptible to crosstalk–the electromagnetic interference that occurs between adjacent copper pairs. This ‘noise’ degrades the signal that is being transmitted through the wire and ultimately negatively affects the speed and reach of copper networks.
For telcos that are not yet ready to make the move to fiber, the crosstalk limitation puts them at a distinct disadvantage when it comes to deploying the high-speed broadband services that consumers want, and which the cablecos and alternative service providers are more than happy to provide.
That’s where vectoring comes in. The technology mitigates crosstalk interference and thereby improves the rate and range of VDSL2 to fiber-like performance. Vectoring levels the playing field for telcos, enabling them to prolong the life of existing copper networks, deliver even the highest-speed broadband services and stay competitive as they make the transition to fiber.
Problem solved, right? Sorry, not so fast. There are currently two opposing schools of thought when it comes to implementing vectoring in the field. While some telecommunications equipment vendors are touting board-level vectoring systems, others are extolling the benefits of system-level vectoring solutions.
We’ll break it all down and examine several real-life deployment scenarios and provide a head-to-head comparison of the two approaches.
Board-level vectoring
Some vendors have taken a board-level approach to vectoring. In this approach, the VDSL2 line card houses a vectoring engine that controls the vectoring activity on the ports of the card.
This method has a few drawbacks. First, vectoring requires significant processing power and memory resources, and when it is allocated from the card, traffic performance can suffer if the number of ports deployed is not reduced. Second, all pairs in a feeder cable must be connected to the same card. Since feeder cables typically contain 100 or more pairs, this approach is highly impractical. If the pairs are not connected, the vectoring engine misses out on crucial measurements and coding that ensure the efficiency of the vectoring operation. This can also have opex implications; when customers are added or deleted there is significant cable management activity that must take place.
Consider the following deployment scenarios, which illustrate some of the shortcomings of board-level vectoring systems.
Scenario 1: High take-up rate
In this scenario representing a high take-up rate, 100 end-users subscribe to a premium service that is enabled by vectoring, as depicted in Figure 1. Assuming the use of a 48-port VDSL2 line card, board-level vectoring would be enabled only if 48 pairs were utilized in each 100-pair feeder, resulting in a waste of 52 copper pairs.
Scenario 2: Low take-up rate
In this scenario, which demonstrates a low take-up rate, 40 end users subscribe to a premium service enabled by vectoring. Twenty subscribers are from one neighborhood (and feeder cable), and twenty are from another neighborhood, as shown in Figure 2.
Initially, there is no problem managing both cables from a single card. However, as time goes on and new subscribers are added, as illustrated in Figure 3, there are not enough free ports on the card to satisfy demand. A new line card is installed to connect the additional customers, but vectoring no longer works efficiently.
This situation requires that a technician be sent out to the street cabinet to perform cable management activities. The technician realigns the cables with the line cards, but now there are unused line card ports, as shown in Figure 4. The end result is unnecessary capital expenditures.
System-level vectoring
System-level vectoring is in sharp contrast to board-level systems. In this case, the vectoring engine resides on a dedicated service card that controls the vectoring activity of all the VDSL2 line cards (See Figure 5). This enables any card to be connected to any pair in any feeder, eliminating cable management concerns and saving opex.
Vectoring–Board or system-level?
A typical system-level architecture consists of a DSLAM shelf for street cabinet deployment, four 64-port VDSL2 line cards, a network interface card and a vectoring engine card (See Figure 6). This type of any-to-any configuration is well suited to the deployment of dynamic, high-speed broadband access services.
It all comes down to performance
As previously established, vectoring at the board level potentially presents both opex and capex challenges. An even more critical issue, however, is that the VDSL2 lines in the cable going from the street cabinet to the subscribers cannot be controlled through board-level vectoring. This limitation means that only marginal performance gains are possible, normally a fraction of what can be achieved with system-level vectoring.
To illustrate the performance gap between system- and board-based vectoring, Figure 7 compares the operation of the two approaches using a 0.5 mm cable with 50 lines and 2 line cards.
The graph shows that while system-level vectoring is capable of achieving near single line performance, board-level vectoring is able to provide only a trivial improvement over no vectoring at all.
Throwdown, vectoring style!
The following table provides a head-to-head comparison of the key attributes related to system-level and board-level vectoring:
And the winner is…
Today, wireline operators face considerable challenges from cable companies and alternative providers alike, and they need every competitive advantage they can get. That’s why they are welcoming with open arms new technologies like vectoring which can help them keep pace with the competition by extending the range and speed of their existing copper plant.
But telcos must tread carefully when deploying vectoring solutions, as not all implementations are created equal. Huge performance disparities can exist between system- and board-based systems, and these can have a dramatic impact on bottom-line capex and opex.
When considering a vectoring solution, operators must look at key attributes including: crosstalk noise cancellation, cable management requirements, port density and usage, and pay-as-you-grow investment.
A system-level vectoring architecture can meet the needs of most next-generation networks. This approach suits perfectly the FTTC/N deployments of street cabinets with active VDSL2 DSLAMs, as it enables operators to fully control all copper terminated at the DSLAM. The FTTC/N deployment is taking off rapidly because it enables wireline operators to convert their passive street cabinets into a perfect springboard towards a next-generation network.
While communication service providers continue to extend their fiber reach, they continue to face major issues when it comes to providing fiber-based services in the very last part of the access network. This is where board-level vectoring is well suited, as FTTB (fiber to the building) enables these providers to primarily target apartment buildings and offer the residents real broadband services.
About the Author
Ariel Caner is Product Marketing Manager at ECI Telecom.
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
