The Cable Operator’s Road to 10G

The demand for high-speed data service bandwidth has continued to grow in recent years, accelerated in many unexpected ways by the COVID pandemic. Much of this new demand is the result of unprecedented shifts in working and schooling from home, coupled with the augmentation of several existing trends. Anticipating this new growth trajectory is the subject of cable operators’ current broadband network planning.

The specific drivers of tomorrow’s bandwidth growth are difficult to predict, but we see the following trends playing a large role in industry strategies for years to come:

1. Continued shift from traditional cable TV to streaming services
2. 3D, virtual reality, augmented reality services
3. Faster downloads of large files (especially for new, synchronized gaming releases)
4. Lower latencies (driven by higher bandwidth capacities)
5. Competition between service providers marketing multi-gig bandwidth capacities

These demands of the future will undoubtedly move operators towards the ultimate adoption of many 10G-oriented technologies as they plan for their next-generation networks.

Note: Within this article, any technology that provides a bandwidth service level agreement that is greater than 1 Gbps and less than 10 Gbps is viewed as a 10G-oriented technology. All of these technologies may be utilized on the “road to 10G.”

HFC augmentation paths

Although the HFC network has existed for many years, cable operators are only beginning to recognize its full capabilities. Many of its latent potential will be ripe for exploit over the next five to ten years, as operators phase in upgrades to their HFC plants.

While each operator will utilize a slightly different path, most will follow the progression described below. It’s important to note that some may opt to change the order and skip or delay certain steps, but the resulting path will still be a logical one to follow.

Step 1 — Node segmentation and node splitting

For most operators that require a quick and simple augmentation of their effective per-subscriber bandwidth capacity, the node segmentation (for segmentable nodes) or the node split (for non-segmentable nodes) is the obvious path. Both approaches effectively increase the depth of fiber—moving the HFC plant down the directionally-correct path towards an ultimate PON solution in the deeper future. The only potential challenge with these approaches is that they tend to be more costly on a per-subscriber basis as the number of subscribers per node becomes smaller.

Step 2 — DOCSIS 3.1 OFDM and OFDM enablement

This step involves merely enabling more spectrum using the DOCSIS 3.1 OFDM (downstream) and OFDMA (upstream) capabilities that were deployed over the past five years. Transitioning from SC-QAM downstreams with ~6 bps/Hz spectral efficiencies to OFDM downstreams with ~8 bps/Hz spectral efficiencies or higher results in a 33% increase in bandwidth capacity for those transitioned channels. Transitioning from SC-QAM upstreams with ~4 bps/Hz spectral efficiencies to OFDMA upstreams with ~8 bps/Hz spectral efficiencies or higher results in a 100% increase in bandwidth capacity for those transitioned channels.

Step 3 — DOCSIS 3.1 mid-split or high-split enablement with 1.2 GHz downstreams

This step requires operators to make changes to active equipment within their HFC plant. The enablement of mid-split operation with the upstream operating to 85 MHz or high-split operation with the upstream operating to 204 MHz will provide increased upstream bandwidth capacity when compared to the 42 MHz or 65 MHz splits of the past. In particular, 85 MHz splits will likely support ~500 Mbps upstream service level agreements (or higher), and 204 MHz splits will likely support ~1.2 Gbps upstream service level agreements (or higher). However, this requires changes to filters within the already-deployed amplifiers and nodes. In areas where downstream congestion is expected to be problematic, operators can enable 1.2 GHz downstreams while also enabling the mid-split or high-split operation.

Step 4 — Low latency DOCSIS enablement

This is a relatively new option in the field of DOCSIS capabilities for the HFC network. It is the array of tools used to ensure low-latency transport for latency-sensitive services on the HFC plant. For many operators, this step will help ensure that gamers will continue to experience good service even if congestion occurs and will likely result in the activation of this feature in upcoming CMTSs and CMs.

Step 5 — DOCSIS 4.0 ultra high-split enablement

This is a futuristic move to DOCSIS 4.0 ultra high-split operation, which permits the  operator to run the upstream spectrum to 300 or 396 or 492 or 684 MHz. This transition can permit up to ~5 Gbps of upstream bandwidth capacity, however it may require more extensive changes to the actives (amplifiers and nodes) within the HFC plant.

Step 6 — DOCSIS 4.0 full duplex (FDX) DOCSIS enablement

This is also a futuristic move to DOCSIS 4.0 with the enablement of FDX operation, which permits running the downstream spectrum and the upstream spectrum on top of one another in a selectable region within the FDX band (between 108 and 684 MHz). This transition can permit the downstream to support up to 10 Gbps while the upstream supports ~5 Gbps. However, it may also require more extensive changes to the actives (amplifiers and nodes) within the HFC plant.

Step 7 — DOCSIS 4.0 extended spectrum DOCSIS (ESD) enablement

This step is an alternative to the previous one with a move to DOCSIS 4.0 with the enablement of ESD operation, which permits running the downstream spectrum above the upstream spectrum, but the downstream spectrum is permitted to operate up to 1794 MHz. This transition can also permit the downstream to support up to 10 Gbps while the upstream supports ~5 Gbps. However, it may also require more extensive changes to the actives (amplifiers and nodes) and passives (taps) within the HFC plant.

Reclaiming bandwidth from video

An important complement to the above steps is the implementation of strategies to reduce the RF bandwidth that is assigned to traditional digital TV delivery. In today’s networks, 50-70% of the precious downstream bandwidth is used for video delivery. As operators move to smaller service group sizes, and consumers move from traditional to streaming services, this is an increasingly inefficient allocation of capacity. There are four options that operators can pursue for the recovery of some or all of this bandwidth within the context of traditional linear and on-demand service offerings:

1. Exit the “traditional” video space completely. This would free up all of the bandwidth, but also eliminate a significant subscriber revenue stream.
2. Complete transition to ABR IP video delivery. This also frees up all of the RF bandwidth assigned to traditional delivery. It requires investment to create the end-to-end ABR-based IPTV service with all of the features associated with a TV service, including local ad insertion as well as regulatory features such as Emergency Alert. On the subscriber side, it requires all subscribers to have set-top devices capable of receiving the new IP delivered services, and potentially the implementation of an operator-managed in-home networking technology such as MoCA or Wi-Fi.
3. Convert MPEG-2 services to MPEG-4. This frees up about half of the RF bandwidth assigned to video. On the subscriber side, it requires subscribers to have MPEG-4 capable set-tops, which may lead to the need to swap out some older devices. On the network side, it requires investment in encoding infrastructure and more significantly in some markets, investment in ad insertion infrastructure to replicate the zone based local ad insertion that exists today. Recently, new ad insertion approaches have emerged that can create the MPEG-4 broadcast streams using the same platform that also provides ad insertion on ABR IP services. This approach enables the operator to focus on building a new IP-based ad insertion capability that will carry them into the future, and to bring new value-added capabilities to the traditional digital video platform.
4. Implement switching technology. This ensures that the only video carried over the network is that being actively consumed. This switched digital video (SDV) technology is well proven and used by multiple operators today. The beauty of SDV is that it can enable a more graceful sunset of traditional digital video delivery in that as subscribers move to IP services, fewer streams are required in the switched pool. So, the bandwidth assigned to video will naturally ramp down over time. In addition, the benefits of node segmentation are magnified by SDV because smaller service groups mean fewer eyeballs watching TV which means fewer streams are required in the switched pool. Finally, SDV is also compatible with the shift to MPEG-4 described in option 3, amplifying the gains of improved compression.

Conclusions

From the lists above, it should be clear that operators will have many technologies and many options from which to pick as they migrate forward towards 10G operation in the future. Each operator will choose its own unique path, but all will be utilizing some form of these technologies in the coming decade. Regardless, it is important to define and execute a bandwidth reclamation strategy for video and to plan the overall evolution of the network in a holistic manner.

This blog was originally posted to Broadband Library in February 2021