Every time a new smartphone comes on the market, I ask myself, should I buy it or not?  It has cool new features, but my current model does everything I need.  Do I really want to spend more money and replace my existing hardware? But, I really like the cool new features. It’s a dilemma.

What if there was a way to keep your current smartphone and still get the cool new features?  So, same performance, but you don’t have to invest in a new device. Yeah, the new iPhone has better bragging rights, but most people (especially parents who have to buy multiple phones) would probably choose the software update path. It’s fast, easy and the kids can start using instantly.

Let’s apply that same question and decision-making process to the network that delivers service to the smartphone. Is building a new fiber plant the only way to support LTE/5G small cell deployment? What if there was a better, less costly and faster way? What if there was a software update that made the existing wireline plant just as good as a new fiber plant.  Same performance, but less cost and less time to market. Yeah, fiber sounds cool, but so does faster time to market and lower capital investment.

The wireline plant I am referring to is the cable operator’s hybrid fiber coax (HFC) plant and the software update is the Low Latency Xhaul (LLX) feature for DOCSIS. More on the technology in a moment. First, who needs this and why?

The Business Opportunity

There are at least two distinct classes of customers in service provider – telco/mobile operators like Verizon and AT&T, and cable operators like Charter and Shaw. Both serve the same customers, namely you and me. Both sell the same services – data, voice and video. The mobile operators already sell mobile services and over 50% of the cable operators sell mobile services, either as an MNO or MVNO, and that percentage is quickly increasing. Both also have a common challenge, and that is how to build out the next generation of the mobile network that involves densification with small cells.

As a radio’s operating frequency go up, the coverage radius of the radio goes down. Today, a macrocell site may cover an area of several miles while a small cell site may only cover a few hundred feet. That means more cell sites are needed for small cell deployment.  To cover the same area of a macrocell, it may take as many as 20 to 50 small cells (I will cover that calculation in a future blog).  The upside is that smaller cell radio also increase the capacity of the network as the number of radios increases and the number of users per radio decreases.

For the cable people in the audience, this is somewhat analogous to the capacity gained from node segmentation in an HFC network. If the HFC plant moved from a cascade of node plus 5 amps (N+5) to a cascade of node plus 0 amps (N+0), there could be 10 to 20 new service groups replacing the previous single service group. In this scenario, node segmentation would provide 10 to 20 times the capacity.

Now, if the ARPU per user is relatively constant, then the cost of deployment and operation of the next generation small cell network has to be on par or less than the cost of deployment and operation of the current macrocell network. If there are 20 to 50 times the number of radios, that means that the cost of the small cell deployment and operation has to be 2% to 5% of the cost of a macrocell deployment. That is a world of difference in economics! How can we make that possible?

The Mobile Backhaul over DOCSIS Proposition

In a conventional macrocell installation, there are three fundamental CAPEX and OPEX requirements:

  • A cell site has to be built or leased
  • Power has to be installed and paid for monthly
  • A fiber backhaul has to be installed and paid for.

If an existing HFC plant is used, then these three fundamental hurdles of installation are almost free.

  • Free site: use either strand-mount or home-install
  • Power is there: both the HFC plant and the home are powered
  • Backhaul is ready: both HFC plant and the home have access to a high-speed DOCSIS network.

This means the 2% to 5% goal may be feasible.

Now, if the small cell had the cost equivalent of a cable modem, then the net CAPEX and OPEX costs are virtually eliminated. You could massively overbuild a geographical area with small cells just to achieve coverage. If a tree or a hill is in the way, just deploy more small cells and connect them to the HFC plant.

Network Bandwidth

A question that often comes up is about the network bandwidth. If there are 50x more cell sites, where cell site haves 5x the throughput, don’t you need 250x the backhaul capacity? No, you don’t. The small cells are focused on connectivity, not bandwidth accumulation. The net bandwidth within a serving area is constant. It depends on the people and devices. Connectivity determines which service provider gets those bits.

Business Models

This line of thinking leads to at least three business scenarios.

  1. Wholesale, where the telco rents space, power, and backhaul from the cable operator, but brands the service under the telco’s name. This is the fastest way for a telco to achieve coverage in an area where they do not have a fiber footprint.
  2. MVNO (Mobile Virtual Network Operator), where the cable company rents the spectrum and markets the service.
  3. MNO (Mobile Network Operator), where the service provider owns and operates the service.

According to the NCTA, the HFC plant passes 93% of the households in the USA. By re-using that plant, time-to-service could be reduced from months to days, and from say $50,000 to maybe $500.  That is something to seriously consider.

An Interesting Transition

For the telco operators to build out a next generation wireless plant, they will need a good wireline plant. The HFC plant owned by the cable operators is a great practical choice. That implies that the telco operators need the cable operators.

For the cable operators to build out a next generation wireless plant of their own, they need a background network that has broad coverage so they can incrementally add wireless bandwidth at the most concentrated points and over time work towards the least concentrated points. The telco operators have that background network today in the form of LTE. That implies that the cable operators need the telco operators.

These facts have not gone unnoticed. Cable operators have provided fiber backhaul for telco macrocell sites for years. And now one of the larger cable companies in the world is Vodafone.

Dual Networks

It is useful to think in terms of a dual network structure that could be referred to as a foreground and background network.  The background network may be a legacy network that has broad coverage but lower bandwidth, and the foreground network has less coverage but is more concentrated and has higher bandwidth. The two networks can be owned and operated by different companies. The role of the networks could be reversed depending on your viewpoint. But ultimately, the foreground network probably contains the majority users, revenue, and investment when compared to the background network, despite which one is actually faster.

For example, in mobile today, the LTE network is the foreground network and the previous 3G network is a background network. 5G will be built out soon, and it will start out as a background network. At some point when there is enough 5G, the 5G network will become the new foreground network and LTE will become the background network.

From a cable operator’s viewpoint, renting or owning LTE based on existing macrocells is the ideal foreground network to start with. LTE is predominantly in the 700 MHz band with some additional AWS spectrum at 1700 and 2100 MHz. As the new CBRS spectrum at 3.5 GHz is built out with small cells, it will form a new background network. Wi-Fi may also be part of that background network. Then, at some point when there is enough CBRS and Wi-Fi, the LTE network becomes the new background network.

Wireline Network is the Foundation of Wireless

In summary, you cannot build a great wireless plant without a great wireline plant. Building a new fiber plant with today’s economics but 50x the radios could be a business disaster. However, by re-using an existing HFC plant that already has sufficient bandwidth, powering, and geographical locations, CAPEX and OPEX costs can be dramatically reduced.

There are three more secondary factors to seal the deal. Each of these are optimizations rather than requirements. These are: ample and scalable bandwidth, low latency services, and network timing. We will talk about these and the technical details of low latency xHaul (LLX) in the next few blogs.

In the meantime, check out the CableLabs blog and the CableLabs LLX website. Want to hear more? Join me on October 8th for the webinar: Densifying the 5G Mobile Network with DOCSIS Low Latency Xhaul.  I’ll cover how LLX works in technical detail and how you can deploy small cell radios as easy as a DOCSIS cable modem. Cisco will be demonstrating LLX backhaul over DOCSIS at the Cisco booth at the SCTE Expo 2019. And, presenting a joint white paper on LLX from CableLabs and Cisco.  See you in the Big Easy!


John Chapman

Cisco CTO Broadband & Fellow

Cisco’s Cable Access Unit