It’s a Stretch – Expanding The Upstream Via Mid Splits and High Splits
In the Advanced Upstream blog series we will explore how cable operators have traditionally managed upstream bandwidth capacity planning, how market dynamics and technical innovation are driving changes, and where things are likely headed in the coming years. In this entry we will discuss the “More Hz” approach – or how to make more spectrum available to add upstream carriers.
More Hz Introduction
Back in the early days of two-way plants, the easiest solution to increasing upstream capacity was to simply add another carrier or two. Fast forward to today, and in most of the 42MHz world and some of the 65Mhz world, there is no such thing as vacant spectrum for adding carriers. During the early days of the pandemic when upstream capacity was the most scarce, operators were adding 3.2MHz and even 1.6 MHz carriers right against the diplexer and deeper into the junk band to squeeze out a few extra Mbps. This was clearly not a desirable or even adequate solution, so attention turned to mid-split and high split architectures.
What is a Mid-Split, What is a High-Split?
The DOCSIS 3.0 specification calls out an upstream frequency range of up to 85MHz, which is now defined as the Mid-Split. Some operators looked into the cost/effort required to implement this, but few went beyond the pilot phase as the benefit just wasn’t worth the effort. Pre-2020, the upstream simply wasn’t the sore tooth demanding attention. DOCSIS 3.1 introduced the concept of a 204MHz upstream, or the High-Split. This approach is more disruptive for reasons discussed below, but since it gives 5x more usable spectrum vs 42MHz plant implementing it can actually defer the need for upstream-driven node splits for many years. It is also required to deliver 1Gbps upstream speeds that are on many operators service offering roadmaps. Like everything in life, the benefits of increased spectrum available for carriers doesn’t come without costs and challenges.
- CPE Compatibility – Only CPE with 204Mhz diplexer are capable of taking advantage of the extended spectrum. Activating 204MHz doesn’t require that all CPE (including legacy STB’s) be replaced however assuming that the downstream is vacated up to 258MHz.
- Downstream Spectrum Loss – As stated above, spectrum up to 258MHz must be vacated to accommodate the 204MHz upstream plus guard band. Operators with 860MHz or 1GHz plant can usually accommodate this through analog reclamation and/or increased video compression, but 750MHz and below plants may need downstream expansions to maintain capacity.
- Actives and Diplex Filters – This is the elephant in the room, every active must be touched as well as some passives to install 204MHz-capable gear and diplexers. There are several approaches to getting this done with minimal customer disruption, but details are beyond the scope of a blog.
- FM Interference – FM interference is especially troublesome in the downstream today as FM signals are always in the air, and can be quite strong near transmission sites. Some operators simply vacate the downstream spectrum around powerful local stations to get around this. With high split networks the FM band moves to the upstream, where one leak can take out the entire upstream vs just a handful homes as is common on the downstream. Conceding valuable upstream spectrum is a much less attractive option, and excessive power from FM interference can tax the upstream laser power budget.
- Signal Leakage – Many operators monitor downstream signal leakage both to find shielding weaknesses that can allow ingress into the plant and in some countries to meet governmental mandates for aeronautical band purity. With high-split architectures the aero band moves from the downstream to the upstream, forcing changes in how leakage detection is done. Many alternative techniques are being proposed, with detection of OFDMA Upstream Data Profile (OUDP) test bursts that DOCSIS 3.1 CPE are already capable of transmitting in the upstream being the leading candidate. Modern FPGA-based signal leakage detectors are essentially highly-sensitive software defined radios, so transitioning to detecting different leakage signatures should be possible.
- Return Path Monitoring/Ingress Troubleshooting – Ingress is worst at the low end of the return spectrum, but operators must make sure that return path monitoring gear and field meters are 204MHz-capable to see the entire upstream. This is especially important in the FM band.
- OFDMA Required for High Split – DOCSIS 3.1 only specifies OFDMA support above 85MHz, so there is no benefit to high split architectures unless accompanied by OFDMA. See HERE for more details on OFDMA
Whew – that’s a lot of challenges, but for many networks expanding the upstream spectrum available through mid/high splits combined with one or both of the other primary options makes great financial and operational sense. Whether used as a stop-gap to defer node splits while DAA solutions continue to mature, deployed as a path to 1Gbps services, or any other reason mid/high splits will most certainly play a role in the advanced upstream plans for many operators around the globe.