Advances in DOCSIS Keep Technology Relevant, but Multiple Variants and Concentrated Ecosystem Create Challenges

By Andrew Spivey | 3Q 2024 | IN-7489

Registered users can unlock up to five pieces of premium content each month.

Log in or register to unlock this Insight.

Over the past 2 decades, U.S. cable Internet Service Providers (ISPs) such as Comcast and Charter have delivered broadband to millions of U.S. households through cable networks which were originally installed for pay TV. Cable remained a competitive broadband access technology throughout this period thanks to successive advancements in the Data Over Cable Service Interface Specifications (DOCSIS) standard, developed by the CableLabs research laboratory and contributors from across the cable ecosystem. All major U.S. cable ISPs today are currently operating networks supporting the DOCSIS 3.1 standard, but it is widely acknowledged that these will struggle to remain competitive against the growing fiber footprint being laid by the likes of Verizon and AT&T, which has recently seen a boost from the US$42.45 billion Broadband Equity, Access, and Development (BEAD) program. DOCSIS 3.1 is also unable to underpin the advanced Wi-Fi performance possible with the introduction of Wi-Fi 7 and the 6 Gigahertz (GHz) spectrum. The fact that cable ISPs need to upgrade their DOCSIS 3.1 networks is not disputed, but what is less clear is exactly what they should upgrade to. Should they invest in DOCSIS 4.0, and if so, what variant? Perhaps the incremental update that DOCSIS 3.1+ offers would be more appropriate? Or maybe it makes the most sense to abandon cable altogether and migrate to fiber? This ABI Insight explores these upgrade options that face cable ISPs, analyzes the key trends in the DOCSIS chipset and Original Design Manufacturers (ODM) ecosystem that will influence their decisions, and ends with some actionable recommendations for both ISPs and ODMs of DOCSIS Customer Premises Equipment (CPE).

The current stock of DOCSIS 3.1 modems deployed by ISPs across the United States offer download speeds that max out at 5 Gigabits per Second (Gbps), upload speeds that are capped at 1 Gbps, and latencies that can only reach 10 Milliseconds (ms). Although these attributes are sufficient for basic applications requiring only fast download speeds such as video streaming, they are not adequate for advanced applications that demand high upstream speeds and low latencies, such as video conferencing and Virtual Reality (VR). To address these limitations, a handful of the major U.S. ISPs are now beginning the process of migration to DOCSIS 4.0 for its improved 6 Gbps upstream capacity and 5 ms latency, as well as a maximum download speed of 10 Gbps. Last year saw CableLabs begin certifications of DOCSIS 4.0 modems, the first wave of DOCSIS 4.0 CPE shipments from equipment vendors including Vantiva, and the world's first residential DOCSIS 4.0 deployment from ISP Comcast. The year 2024 has seen the gradual acceleration of the DOCSIS 4.0 rollout, and ABI Research is forecasting that DOCSIS 4.0 CPE shipments will climb steadily over the coming years from a total of 0.7 million this year to 5.4 million in 2028.

There are two approaches to DOCSIS 4.0, each with its own pros and cons. The Extended Spectrum DOCSIS (ESD) variant supports expanding the compatible spectrum to 1.8 GHz to attain faster speeds, whereas the Full Duplex (FDX) method allows both downstream and upstream traffic to flow over the same 1.2 GHz section of spectrum for symmetrical speeds. The DOCSIS 4.0 standard supports both ESD and FDX, although ISPs can be expected to concentrate resources on one option for cost effectiveness. Indeed, all the major cable ISPs upgrading to DOCSIS 4.0 have expressed their preferences, and the different choices are creating a divided landscape in the market. Comcast, the largest U.S. cable ISP, is taking the FDX route, whereas Charter Communications, the second largest cable ISP, has opted for ESD. Other cable ISPs that appear committed to the ESD method include Cox, Liberty Global, and Rogers Communications.

DOCSIS 4.0 isn't the only option on the table though. For ISPs that need to enhance their existing cable network, but don't have the funds or time for a full DOCSIS 4.0 upgrade, there is a new alternative known as DOCSIS 3.1+ (or DOCSIS 3.1 Enhanced, DOCSIS 3.1 Extended, DOCSIS 3.1E, or BoostD 3.1, depending on who you ask). This pushes DOCSIS 3.1 to its limits to help achieve 8 Gbps downstream and 1.5 Gbps upstream. This is achieved via several relatively inexpensive upgrades to the existing DOCSIS 3.1 network, including a software update to the Cable Modem Termination System (CMTS), which enables more than 2 Orthogonal Frequency Division Multiplexing (OFDM) channels to be bonded together by currently deployed DOCSIS 3.1 modems. Some approaches also involve installing DOCSIS 4.0 modems into existing DOCSIS 3.1 networks to again enable additional OFDM channels to be bonded. There are two distinct potential customer types for DOCSIS 3.1+. The first group consists of those that see the technology as a short-term stopgap between DOCSIS 3.1 and DOCSIS 4.0. Tier One ISP Charter is taking this approach, as it deploys DOCSIS 4.0 modems within its existing DOCSIS 3.1 network in the middle phase of its 3 phase plan for rolling out DOCSIS 4.0 across its entire network by 2026. The second group consists of those that consider DOCSIS 3.1+ compelling enough that they will adopt the technology and forgo DOCSIS 4.0. Midco, a regional cable ISP based in the Great Plains region, has confirmed that it will deploy DOCSIS 3.1+ instead of DOCSIS 4.0, and will most likely look to invest in the rollout of fiber thereafter.

Given that ISP strategies are still in flux, and that they may abandon cable altogether in favor of fiber, DOCSIS chipmakers understandably have apprehensions over developing a new technology for which future demand is not certain. To control this risk, market-leading DOCSIS chipset vendor Broadcom mandates that all ISPs and CPE ODMs that wish to access its DOCSIS 4.0 technology must enter into Joint Development Agreements (JDAs) with the company. These JDAs require ISPs/ODMs to divulge their full DOCSIS 4.0 plans to Broadcom and commit to set order volumes, a measure that helps Broadcom to accurately forecast future demand and ensure a healthy Return on Investment (ROI). ISPs that have signed JDAs with Broadcom include Comcast, Charter Communications, and Cox Communications, and ODMs include Vantiva. Alongside Broadcom, the only other provider of DOCSIS 4.0 chipsets is MaxLinear, with its Puma 8 chipset that was made available in 4Q 2023. The DOCSIS 4.0 go-to-market strategy of MaxLinear, which only entered the DOCSIS market following its acquisition of Intel's home gateway division in 2020, is markedly different from Broadcom's in that it doesn't require restrictive JDAs. Instead, MaxLinear is open to working with any CPE ODM, which, to date, includes Askey, Hitron, Sercomm, and Vantiva (through the CommScope product line that Vantiva recently purchased). Aside from the JDA, another major contrast between the two vendors is their approaches to the different variants of DOCSIS 4.0, with Broadcom offering a solution that supports both ESD and FDX, whereas MaxLinear's Puma 8 supports ESD only.

Cable Internet is, by far, the most dominant broadband access technology in the United States today, representing 68.1% of current fixed broadband subscriptions, and despite the rapid increase of fiber subscribers (climbing 487,000 in 1Q 2024, compared to a decrease of 169,000 for cable), cable will remain the predominant access technology through 2030, when it will still constitute 53.5% of fixed access subscriptions. The continued relevance of DOCSIS Internet in the United States is good news for the main DOCSIS CPE ODMs, which include Arcadyan, Askey, Hitron, Sagemcom, Sercomm, Ubee Interactive, and Vantiva. Yet, these ODMs need to be strategic in how they approach the future of DOCSIS if they wish to retain market share in what is still a declining market.

When developing for DOCSIS 4.0, it is recommended that CPE ODMs should focus on developing CPE that supports just one variant of DOCSIS 4.0. This is because cost efficiency concerns mean that cable ISPs have and will continue to prioritize either the ESD or FDX approach to DOCSIS 4.0, not both simultaneously. Considering this, there is little value to add by designing CPE that can support both variants, because the benefit that such CPE will offer (flexibility to enable ESD and FDX from the same equipment) will be irrelevant to ISPs. ISPs will also be resistant to pay the higher cost of dual-variant DOCSIS 4.0 CPE if they only intend to leverage one variant. At the same time, given the not insignificant interest in DOCSIS 3.1+, ODMs should develop solutions for this market, too. If they choose to do so, chipset solutions are available from both Broadcom and MaxLinear. The former has introduced the BCM3392, a DOCSIS 3.1 chipset that supports the bonding of 4 OFDM channels (up from the standard 2 for DOCSIS 3.1), which importantly is not restricted by a JDA. MaxLinear, on the other hand, has confirmed that its Puma 8 chipset can also be used for DOCSIS 3.1+.

Turning our attention now to cable ISPs, it is recommended that they devise a long-term strategy for their network, planning not just the next upgrade, but future upgrades, too. For those considering fiber, which requires considerably more Capital Expenditure (CAPEX) than an update to an existing cable network does, it is important that ISPs choose the optimal Passive Optical Network (PON) technology and the right timing for their upgrade, in order to ensure they can future-proof their network and lock-in greater long-term ROI. 25G and 50G PON are the latest standards available today, and although they currently come with a high price premium, economies of scale mean that the cost will fall dramatically once large-scale deployments commence in earnest. These will begin around 2027, and so for some cable ISPs planning a fiber transition in the near-term, instead of installing XGS-PON today, it may make more economical sense for them to implement a low-cost DOCSIS 3.1+ cable upgrade in the near-term to extend the life span of their cable network, and then in 2 to 4 years commence the rollout of 25G/50G PON once the prices have fallen.

That said, ISPs shouldn't assume that the DOCSIS story ends with DOCSIS 4.0. Proposals for the next advancements to the standards (most likely going by the name of DOCSIS 4.1 or perhaps DOCSIS 5.0) have already begun. Some advocate that DOCSIS 4.1 should merge both the ESD and FDX approaches of DOCSIS 4.0 to achieve upstream speeds of 10 Gbps. Another suggestion is extending the DOCSIS spectrum out to 3 GHz to deliver a whopping 25 Gbps download speed. If this latter approach were to be realized, cable would once again see an update, which would keep it competitive against fiber, helping to keep DOCSIS relevant for many years to come.