XG-CABLE for HFC networks

Transcription

1 New technology enables full duplex transmission for 10 Gb/s symmetrical services on existing HFC infrastructure Technology White Paper XG-CABLE, a breakthrough innovation from Nokia Bell Labs, gives cable and multiple system operators (MSO) the potential to provide 10 Gb/s symmetrical services over their existing hybrid fiber-coaxial networks. XG-CABLE effectively doubles the capacity of HFC networks by enabling full duplex transmission with extremely wideband echo cancellation (1.2 GHz) while eliminating the interference that once made full duplex impossible. This proof of concept technology could be made compatible with legacy DOCSIS networks and validates CableLabs Full Duplex DOCSIS 3.1 proposal. This is significant since it addresses upstream constraints that have long been the Achilles heel of DOCSIS technologies, just as cable operators face the challenge of growing demand for cloud services, HD video uploads, real-time gaming, live streaming video and virtual or augmented reality. 1 Technology White Paper

2 Contents Demand for symmetrical services 3 Current limitations of DOCSIS 3 Boosting DOCSIS upstream capacity 3 Solving the full duplex challenges 4 Deployment scenarios 6 XG-CABLE performance 7 Considerations for deployment 8 Choosing the right ultra-broadband option 9 References 9 2 Technology White Paper

3 Demand for symmetrical services The last few years have seen an explosion in ultra-broadband technologies and services. Many consumers now take for granted downstream broadband speeds of tens or hundreds of megabits per second (Mb/s), with Gigabit and multi-gigabit services being offered in many markets. With the increasing popularity of cloud services, HD video uploads, real-time gaming, live streaming video and virtual or augmented reality, demand is increasing rapidly for ultra-high upstream services. While cable operators have always been competitive in providing high-speed download rates, they have been unable to provide symmetrical or high bandwidth upstream services due to the technical restrictions of DOCSIS technology. However, upstream services are increasingly seen as a key differentiator for both residential and business services. Current limitations of DOCSIS All DOCSIS technologies use Frequency Division Duplexing to separate the upstream and the downstream traffic, respectively assigning upstream and downstream traffic to the lower and the higher parts of the frequency spectrum. Today s HFC networks use a so-called low-split, with upstream bandwidth being limited to 5-42 MHz (USA) or 5-65 MHz (most of Europe), while downstream starts at 54 MHz (USA) or 85 MHz (most of Europe) to occupy the rest of the available bandwidth. This split is hardwired into all of the active components in the HFC network with diplexer filters (in the cable modem termination system [CMTS], cable modem [CM], but also in the amplifiers in the outside plant). This split is responsible for the limited upstream data rates in current HFC networks. Although more recent DOCSIS technologies (3.0 and 3.1) increase the maximum upstream frequency to 85 MHz ( mid-split ) or 200 MHz ( highsplit ), even the high-split option in DOCSIS 3.1 limits the upstream/ downstream ratio to 1:10. This 1 Gb/s of upstream capacity is restrictive when you consider the large number of subscribers that could be sharing this capacity and when compared to the latest passive optical network (PON) technologies used in competing fiber-to-the-home (FTTH) networks. Boosting DOCSIS upstream capacity In 2014, Nokia Bell Labs began looking at ways to improve upstream capacity in HFC networks by enabling full duplex transmission. Separately, CableLabs announced in February 2016 that they had begun work on a project to develop a full duplex version of DOCSIS Technology White Paper

4 CableLabs acknowledged that a number of technical challenges remained in making the technology viable for deployment. These challenges included the echoes and interference created by sending and receiving on the same frequencies at the same time. In May 2016, Nokia announced that it had overcome these challenges with a proof of concept technology called XG-CABLE which combines full duplex transmission with advanced echo cancellation techniques. Nokia demonstrated that XG-CABLE can achieve symmetrical 10 Gb/s data rates over 100 meters of coaxial drop cable (RG6) using 1.2 GHz of spectrum. XG-CABLE builds on Nokia s extensive knowledge of transmission technologies and Bell Labs achievements with XG-FAST, which in 2014 set the world record for transmission speeds over traditional copper lines [1]. Solving the full duplex challenges Echo cancellation Instead of splitting the available spectrum into dedicated upstream and downstream bands, XG-CABLE uses the full 1.2 GHz spectrum for both at the same time, i.e. full duplex. Using the spectral efficiency of DOCSIS 3.1 as a reference, each 100 MHz of bandwidth yields approximately 800 Mb/s downstream and upstream, giving a total capacity of 10 Gb/s in each direction. A simple analogy for full duplex is a conversation between two people with each speaking at the same time. While this effectively doubles the capacity to communicate, the problems quickly become apparent. As one person speaks, they hear themselves at the same time as they hear the other person. In a network, this is referred to as an echo. Applying this analogy to the HFC network, the transmitter can be the cable modem (CM) or the CMTS. When transmitting, the port interfacing to the outside plant causes an initial echo. A second echo is created when the transmit signal is reflected back from the outside plant. Combined, these two conditions can create a large echo which exceeds the strength of the useful signal. As such, no communication is possible without echo cancellation. To cancel echoes, XG-CABLE uses similar techniques to those in Nokia s market leading vectoring solution. The principle is comparable to noise cancelling headphones. XG-CABLE measures the echo experienced by the transceiver and properly compensates for it. The performance of the echo cancellation approach used in the XG-CABLE proof of concept is illustrated in Figure 1. 4 Technology White Paper

5 Figure 1. Echo cancellation depth Echo cancellation depth (db) Frequency (MHz) The figure shows the echo cancellation depth, in logarithmic scale, over the entire band of interest (from 5 to 1200 MHz). The echo cancellation depth is defined here as the difference between the signal coming out of the transmitter and the noise floor experienced at the end of the receiver chain. XG-CABLE achieves an echo cancellation depth of between 70 and 80 db over almost the entire bandwidth, spanning more than 1 GHz of spectrum. To reach the 10 Gb/s symmetrical target over typical fiber-to-the-last-amplifier deployments (FTTLA, see below), the echo cancellation depth would need to exceed approximately 80 db over the entire band. Full duplex interference management Another challenge with full duplex is the interference between concurrent signals from different cable modems in close proximity, e.g., cable modems served by the same tap or splitter. These nearby cable modems can cause interference with each other when upstream signals occupy the same frequencies as the downstream signals. In this case, the signal from the transmitting cable modem will be strong due to the close proximity and limited cable plant traversed. These signals can then interfere with the downstream signals that nearby cable modems receive. Downstream signals are weak in comparison due to the high attenuation incurred on the passive network between the CMTS and the CM. In the figure below, the downstream signal of cable modem 1 is subject to interference from the upstream signals of other cable modems. 5 Technology White Paper

6 Figure 2. Full duplex interference CMTS DS Tap A Tap B US US CM1 CM3 CM5 CM7 Sa Sb Sc Sd CM2 CM4 CM6 CM8 To extend our analogy, imagine person A talking to person B. When person C, who is standing much closer to person B, also starts talking, it is hard for person B to discern what person A is saying. This full duplex interference will be managed using central coordination dictated by the XG-CABLE CMTS. The CMTS will jointly manage upstream and downstream transmissions from all cable modems in such a way as to mitigate the impact of the full duplex interference using dynamic allocation of time slots, frequency slots and transmit power based on rate demands. It is based on the identification of interference groups, which are determined by the XG-CABLE CMTS. The interference groups are defined as follows, based on the impact of the upstream signals of one modem on the downstream signals of another modem: Cable modems in the same interference group impact each other (the upstream signal of any cable modem in the group impacts the downstream to all the other cable modems in the same group) Cable modems in different interference groups do not disturb each other (the channel losses between the modems are sufficiently high) In the figure above, e.g., cable modems 1 to 4 and cable modems 5 to 8 could represent two different interference groups. The XG-CABLE CMTS will take these interference groups into account in its overall scheduling, making sure that cable modems within one interference group are managed in such a way that they do not impact on each other, for example by assigning disjointed time intervals. Deployment scenarios We considered two different deployment scenarios for XG-CABLE: fiber-tothe-last-tap (FTTLT) and fiber-to-the-last-amplifier (FTTLA). Fiber-to-the-last-tap A FTTLT deployment consists of an optical overlay network to the last tap, essentially yielding a homes-passed optical network, homes-connected coaxial network (see Figure 3). 6 Technology White Paper

7 Figure 3. Fiber-to-the-last-tap deployment Private property Public property Tap Cable (coax) Fiber The connection between the XG-CABLE node at the tap location and the home is then a point-to-point coaxial drop cable, typically of RG6 or RG11 type, and would allow for reverse powering (as used in the more recent DSL technologies like vectored VDSL2 and G.fast). Fiber-to-the-last-amplifier A FTTLA deployment would consist of an optical overlay network to the last amplifier. The connection from the XG-CABLE node to users would then be a point-to-multi-point connection, as is the case in today s networks. Fiberto-the-building deployments, with a shared point-to-multi-point coaxial infrastructure inside the building are considered to be a subset of FTTLA deployments, with similar characteristics. The XG-CABLE node would typically be forward powered (from the network over the legacy feeder coaxial cable, as is the case for powering amplifiers today) or locally powered. Since the FTTLA section is point-to-multi-point, a legacy migration and coexistence scenario needs to be in place when deploying XG-CABLE. Figure 4. Fiber-to-the-last-amplifier deployment Fiber backhaul Tap CMTS CM CM XG-CABLE performance XG-CABLE was tested in the above FTTLT and FTTLA deployment scenarios using 1.2 GHz of signal bandwidth (the bandwidth typically supported by DOCSIS 3.1). The tests were carried out under laboratory conditions but using network elements and cabling comparable to a typical HFC network. In a FTTLT HFC network topology, XG-CABLE was able to deliver 10 Gb/s symmetric data speeds over a point-to-point 100m coaxial drop cable (RG6 type) 7 Technology White Paper

8 For a typical FTTLA HFC network topology, XG-CABLE was able to deliver 7.5 Gb/s of symmetrical data speeds Figure 5. XG-CABLE results 25 Net data throughput (Gb/s) FTT-last-tap (point-to-point coax) FTT-last-amplifier (point-to-multipoint coax) Downstream Upstream Further echo cancellation improvements and increased achievable rates are expected in future iterations of XG-CABLE, with the aim of achieving 10 Gb/s symmetrical data rates in FTTLA configurations. Considerations for deployment Any significant change in technology requires new investment in the network. However, XG-CABLE, like Full Duplex DOCSIS, has been designed with existing outside plant in mind and implementing XG-CABLE is potentially comparable with a move from DOCSIS 3.0 to DOCSIS 3.1. By aligning XG-CABLE with CableLabs proposals for Full Duplex DOCSIS 3.1[2], backwards compatibility is assured and different generations of DOCSIS equipment would be able to co-exist on the network. For example, it would not be necessary to replace an installed base of DOCSIS 3.0 cable modems, only switching out those modems where users have subscribed to new XG-CABLE services. Current passive HFC plants are typically limited in bandwidth due to the bandwidth limitations of taps and splitters present in the network. As is the case for DOCSIS 3.1, taps and splitters in the outside plant that do not support 1.2 GHz of bandwidth would need to be replaced in order to achieve the full potential of XG-CABLE. Since full duplex operation requires fully overlapping spectra for upstream and downstream signals, both the forward and reverse line amplifiers would need to amplify the full spectrum in order for them to support full duplex operation. As such, XG-CABLE is only suitable for deployments in passive HFC networks (last amplifier or beyond). This requires fiber to be pushed deeper into the network. However, many operators are already pursuing a deep-fiber strategy. 8 Technology White Paper

9 Choosing the right ultra-broadband option Nokia and CableLabs are working together to align XG-CABLE with the Full Duplex DOCSIS 3.1 proposal. While a commercial solution is a few years away, XG-CABLE will potentially increase the competitiveness and, therefore, the lifespan of HFC networks by enabling ultra-broadband symmetrical services over existing HFC infrastructure. While FTTH is now commonly accepted as the ultimate end goal for both cable and telco service providers, XG-CABLE gives cable operators more options in how to meet the growing bandwidth demands of customers, either upgrading existing HFC infrastructure or deploying FTTH where it makes economic sense. Nokia Bell Labs Consulting can help cable operators develop the business cases for these scenarios. References [1] W. Coomans, R.B. Moraes, K. Hooghe, A. Duque, J. Galaro, M. Timmers, A.J. van Wijngaarden, M. Guenach, and J. Maes, XG-FAST: The 5th Generation Broadband, IEEE Communications Magazine 53(12), pp , Dec [2] February 16, 2016 Nokia is a registered trademark of Nokia Corporation. Other product and company names mentioned herein may be trademarks or trade names of their respective owners. Nokia Oyj Karaportti 3 FI Espoo Finland Tel (0) Product code: PR EN (May) Nokia 2016 nokia.com