Jeff White, head of business development and strategy at Fixed Networks Cable, discusses how cable operators can address the demands of an increasingly digitally driven world.
As the broadband market continues to evolve, reams of analysis on current and future trends are emerging, providing insights into consumer behaviour and the ‘next big thing’. But dial down the details of each report and the findings can be simplified to one overriding requirement: The need for more.
In today’s digitally driven world, operators need additional bandwidth to meet the unprecedented demand for over-the-top video, a more responsive network for real-time applications like video games and faster upload speeds to match changing application and consumption patterns. Consumer demand for digital entertainment and applications is on the rise and shows no sign of abating. New devices and services are constantly being introduced, impacting every aspect of our daily lives.
To address these evolving demands, better and faster networks are needed, especially for cable operators, who are having to upgrade their existing Hybrid Fibre-Coaxial (HFC) networks to stay competitive, better meet evolving end-user needs and meet the growing demand for gigabit services.
Keeping up with demand
The average consumer is likely unaware of the strain their increased digital appetite is putting on networks. However, if adequate network capacity is not available, it can quickly manifest into a customer service issue.
To address this, multiple system operators (MSOs) are looking at fibre-to-the-home (FTTH) as a long-term strategy, but it will likely take decades before reaching the end goal. In the interim, to keep up with cable access capacity demands, cable operators must effectively overcome two critical challenges:
- Improving the fidelity (i.e. signal quality) in the outside plant (OSP)
- Increasing capacity density in the headend to reduce excessive space and power consumption
To boost network performance and bandwidth, cable operators are looking across a wide array of options ranging from fundamental approaches such as node splits, deep fibre, and spectrum reclamation to adopting new technologies, specifications, and network architectures such as CCAP, DOCSIS 3.1, distributed access, and virtualisation techniques.
The established approaches can’t stand alone
There are a few established approaches to increasing network capacity. One is the node split. This tactic involves identifying a heavily burdened node that is serving say 500 customers, and supplementing it with a second node so that each node is now serving only 250 customers. Fibre is often drawn closer to the home as node splits are implemented.
Another method is QAM channel reclamation where MSOs are turning off analog channels, moving to IP video, and reducing QAM video channels by moving to more advanced video compression techniques. All of which allows MSOs to reclaim the now-available spectrum for data.
A more recent development — with rollouts just beginning — is DOCSIS 3.1, the next-gen version of the DOCSIS specification, which aims to make multi-Gbps speeds possible over the existing HFC. DOCSIS 3.1 increases bandwidth in two ways: Adding capacity by broadening the spectrum used (initially up to 1.2GHz) and increasing efficiency to deliver more bits per hertz. Among the techniques used to do this are wider channels (up to 192MHz) and higher order modulation schemes (up to 4096 QAM).
DOCSIS 3.1 can deliver speeds of up to 10Gbps downstream and 1Gbps upstream. However, to make effective use of DOCSIS 3.1, many cable operators will need to make changes to their outside plant, likely pushing fibre deeper and increasing the number of nodes. Additionally, customer premise equipment will need to be replaced with DOCSIS 3.1-compatible modems.
While these approaches may provide some benefit, they require better fidelity in the outside plant which in turn demands more space and power in the hub and headend.
CCAP is not enough
The Converged Cable Access Platform (CCAP) was designed to address the challenge of increased capacity density in the headend. The CCAP is a very large, cable-specific piece of hardware that sits in the headend and combines traditional video (EQAM) and data (CMTS) platforms into a single box.
The idea for the CCAP came at a time when the demand for QAM-based video-on-demand and data were increasing at a similar rate. Since that time, and due to the unforeseen and sudden shift to IP video, demand for QAM-based video has declined while demand for data has skyrocketed, making the CCAP poorly suited for current traffic patterns.
Additionally, eliminating the combining equipment, which was one of the objectives of CCAP, requires the existing EQAM equipment be ripped and replaced with corresponding CCAP line cards. Since the move to IP video will bring an end to QAM-based video, most cable operators don’t want to incur the additional rip and replace cost. Consequently, most CCAPs are being deployed as dense CMTS’s (i.e. data only) with little densification benefit.
Even in those cases where QAM-based video is integrated into the CCAP, it can only go so far, not providing significant rack space or power-savings, as the CCAP itself is still a sizeable and power-demanding piece of equipment. Furthermore, CCAP does nothing to increase the outside plant fidelity.
Distribution is the first step
One significant initiative many MSOs are looking at to resolve the OSP fidelity is a new network architecture called a Distributed Access Architecture (DAA). DAA takes the DOCSIS physical layer (PHY) and distributes it out to the node. Placing the PHY in the node replaces the analog optical signal from the headend to the node with a digital 10Gbps Ethernet connection. The standard RF signal that existing cable modems and set-top boxes look for now begins in the node, very close to the end user. This is an enormous improvement, delivering not only ten-times more fibre efficiency, but also significant CapEx and OpEx savings on analog optics and maintenance/support costs, respectively.
However, distribution only solves half the problem. A DAA still requires a large, cable-specific piece of hardware in the headend – a CCAP-core – to house the DOCSIS MAC. Legacy cable vendors are typically making use of their existing CCAP platforms, with the DOCSIS PHY removed, for that purpose. While the CCAP-core further reduces the power demands in the headend, it does not do as much as today’s technology is capable of. Enter virtualisation.
Less is More – Viva Virtualisation!
Virtualisation makes use of Software Defined Networking and Network Functions Virtualisation techniques to perform functions in software that have historically required hardware. Virtualisation delivers enormous savings as it uses off-the-shelf servers, which are more cost, space and power efficient than purpose-built hardware. It also provides much greater flexibility and agility with respect to service creation.
In the case of the cable access network, virtualising the DOCSIS MAC in conjunction with a DAA completely eliminates cable-specific headend hardware. Because the MAC is simply software, it can be implemented on servers in the headend or even on the node itself. By running the MAC on the node, the headend is left with only standard data centre switches and routers, and servers acting as the virtualisation controller. Properly implemented, a virtualised controller also can support multiple access technologies (e.g. 10G EPON, Point-to-Point Ethernet) on the same controller and network.
A cable network architecture leveraging both DAA and virtualisation fully addresses both the outside plant fidelity issue and the headend capacity density issue. The headend-to-node link eliminates the use of analog optics, pushing IP closer to the customer and supporting 10Gbps Ethernet. The space and power-savings in the headend represents a 7x and 8x reduction compared to traditional CCAP. This approach enables MSOs to cost-effectively support gigabit services on their existing HFC while providing a seamless evolution to an all-fibre, all-IP network.
Picking the right path
When deciding which upgrade route to take, operators must take into consideration a multitude of factors, including the competitive landscape, the modernity of their hybrid fibre/coax networks, the mix of potential customers they are looking to serve and whether they intend to operate primarily in their existing market or expand their footprint.
In some cases, a combination of strategies might be appropriate. But with the need for speed ever-increasing, a decision on how to best meet this demand and remain competitive is needed – and quickly.