Remember the Headend
The data crunch is here, so naturally the topic of one gigabit billboard speeds is still top of mind for broadband providers around the world. As a result, we have explored at length the challenges that operators face in preparing the outside plant to deliver these speeds.
These challenges include: How far do we push fiber deeper into the network to meet subscribers' service and application expectations? Does the node split and segmentation model continue to work with changing standards, policies and procedures? Where does remote PHY fit into the CATV network's evolution? Under what conditions is the ROI right to deploy fiber all the way to businesses and homes?
When considering the best approach for the outside plant, it’s also important to remember the headend in your network transformation strategy. Even the most efficiently deployed outside plant network can miss the operator’s service targets if the headend services are not in place. So, for every change in the outside plant, there needs to be an equally significant change at the headend. Sounds simple enough... until you consider the inherent challenges that come with the headend's physical constraints, including footprint limitations and cooling and powering requirements.
These constraints are among the reasons why we are hearing less about full-spectrum transmitters and more about network functions virtualization (NFV) and virtual cable modem termination system (vCMTS) platforms. Both approaches help deliver bandwidth where and how a subscriber wants it while also moving some of the headend functionality into the outside plant where there's more room to grow. Consequently, today’s headend is looking less and less like a headend and more and more like a data center.
Inside the headend, the traditional RF cabling and quadrature amplitude modulation (QAM) equipment are giving way to gigabit Ethernet switches and virtual machines running on traditional server platforms. This headend re-imagining affects cabling and connectivity because operators are essentially replacing the heavy coax infrastructure with high-density fiber optic cabling, cross-connects and dense wavelength division multiplexing (DWDM) capabilities.
What makes this migration daunting for many cable engineers is the already capacity-constrained environment and the need for uninterrupted service to existing subscribers during the changeover. Fortunately, dense but modular fiber management solutions can help. "Dense" in this case can be defined as a tenfold increase in connectivity at the headend -- a scale suitable for supporting today's Fiber Deep deployments, as well as any future migration toward remote PHY and/or fiber-to-the-home.
Let's look at three categories of usability and density improvements that enable headend transformations without service interruptions:
Traditionally, cabling in the headend has adhered to the classic interconnect architecture. Cables “interconnect” to a main distribution frame (MDF), which then provides access to the outside plant nodes. MDFs are then segmented by how many fibers are needed for each node -- whether that’s six or eight fibers, or more.
Today, operators must look toward an integrated architecture where they may choose a cross-connect structured cabling platform to help support additional demarcation and flexible moves, adds and changes (MACs). Although demarcation additions and MACs may not be common in a traditional HFC environment, the expansion of business-class services and possibility of new service capabilities in the outside plant (such as small cell backhaul) will require a common and consistent cabling architecture.
Integrated cable assemblies
Having seen my fair share of headends, I can assure you that you're not alone if the area between your bays is a tangled jumble of jumpers. At first, running jumpers between bays starts off as a temporary installation practice but then inevitably turns into standard operating procedure. The resulting jumper cluster is an eyesore, and even worse, the very picture of inefficiency in a network with little tolerance for practices that can compromise service.
Doing it right doesn't have to be hard, though. Integrated cable assemblies can connect services from origin to termination point easily, while enabling easy expansion through the use of parking lots within the passive frames -- no untangling required.
Dense, modular DWDM
With migration to remote PHY, operators are now able to consider tightly spaced DWDM channel plans. But deploying 48-channel muxes throughout your headend is not the answer. Instead, deploy just what is necessary today with a slight eye toward expansion.
Even if it’s not in the forecast, future expansion shouldn't require a re-cabling with this modular approach. Just make sure modular capacity expansion is available and easily achievable.
Network migration at the headend doesn't have to be daunting -- it's doable. Even without impacting service delivery to existing subscribers, it’s possible to expand your network's services within existing congested footprints in perfect alignment with the work being done in the field.
Once complete, the payoff is visible. You will see less coax and more fiber in the headend with ten times greater connectivity and more space to expand. The transformed platform will be able to take on capacity to enter new markets, such as wireless backhaul, and grow your base of business-class subscribers.
— Kevin Bourg, Optical Network Architect, Market Development, Corning Optical