Meeting residential bandwidth demand
User requirements will drive next-generation topologies
Many cable network operators with limited fiber assets are working to capture additional revenue from the growing demands of the core residential network. To support continued residential bandwidth growth, MSOs must look for innovative design solutions based on proven technologies that can obviate the need for costly overbuilds, support rapid service velocity, and allow for flexible and scalable networks.
Fortunately, there are system design approaches that can increase the overall performance of existing deep-fiber networks in a simple, easy and cost-effective manner. These approaches are adaptable to various MSO network architectures and applications, and can also support advanced features that improve network reliability at a fraction of the cost of more expensive architectures.
Operators should evaluate the future bandwidth requirements of their residential subscribers. That will help determine which architectures support those bandwidth requirements. We can then show how implementing various combinations of passive optical components in the outside plant will enable flexible networks that meet those requirements.
Figure 1: Downstream bandwidth requirements for all IP delivery.
1. Residential bandwidth growth (traffic modeling)
For this analysis, a traffic model was created to predict the upstream and downstream bandwidth needs of the residential subscriber out to 2012. The method used to calculate downstream residential bandwidth needs was based on bandwidth per application and number of digital streams.
Bandwidth per application. An important factor that will drive residential bandwidth requirements is changes in the type and number of applications people will use over the Internet. Applications such as Web pages, video, music, software downloads and pictures will drive the need for greater downstream bandwidth and upstream bandwidth.
Video will be an important application for the residential subscriber. Standard definition television (SDTV) and high-definition television (HDTV, 1080i and 1080p) all require video compression to reduce the size of the data stream so they can be deployed over an access infrastructure.
Number of digital streams. The next area analyzed in the traffic modeling was the number of digital streams that can be delivered to the subscriber. In the most basic sense, the first three that come to mind are data, voice and video, but other factors can increase the amount of streams past these three. Some of the factors include:
• Number of available TVs
• Number of phones
• Number of computers with a broadband Internet connection
• Number of digital video recorders.
An algorithm was used and a graph was plotted using all these factors to determine bandwidth. Using this type of calculation, two different types of users were plotted, the “power user” and the “average user.” This is plotted in Figure 1. Improvements in compression were taken into account, but the analysis shows that by 2012 the power user will require over 100 Mbps of downstream bandwidth and the average user will require over 40 Mbps of downstream bandwidth. (A power user is defined as a first adopter of new technologies and would expect the best bandwidth. The average user is defined as a late adopter of new technologies and would be satisfied with standard bandwidth.)
Figure 2: Downstream bandwidth requirements for voice and data only.
If the video is delivered by other means such as RF, then the bandwidth curves for the power user and the average user will be different. By subtracting out the video portion of the chart in Figure 1, the adjusted graph in Figure 2 is plotted. Figure 2 illustrates that if video is RF, then by 2012 the power user will require about 80 Mbps of bandwidth. The average user will need close to 20 Mbps of bandwidth.
Using the existing infrastructure and driving the optical fiber deeper into the network will be very important to the MSO. If the MSO stays with RF video, the DOCSIS platform will be the workhorse to deliver the upstream and downstream residential data requirements. With DOCSIS 3.0, channel bonding and smaller node pockets, the DOCSIS platform will be able to deliver better downstream and upstream rates to residential subscribers. This is shown in Figure 3.
2. Determining which architectures support future bandwidth requirements
For this analysis, both technical and business considerations were taken into account. An appropriate planning period of five years (2008 through 2012) was chosen. The considerations to be discussed include technical, deployment costs and operational expenditures (opex).
Technical. As stated above, bandwidth requirements are a function of application. Of course, applications always arise that consume available bandwidth. Furthermore, considerations for commercial services need to be included with those of residential. Commercial customers often require both voice and symmetrical data services.
Both HFC and passive optical network (PON) FTTH can handle the bandwidth requirements over the chosen planning period. HFC, however, will require extensive modification to handle the bandwidth toward the end of the planning period. HFC modifications will include but are not limited to node splitting, upgrading to the latest version of DOCSIS, switched digital video and upgrading to 1 GHz.
Figure 3: Based on 30 percent penetration x 50 percent on-line
during peak hours x 50 percent simultaneous download/upload during peak hours.
Even with these modifications, commercial services often require direct fiber links to accommodate small and medium businesses as today’s cable modem is limited in the upstream. These upstream bandwidth limitations have forced many MSOs to standardize new builds to 250 homes per node and begin splitting to 125 homes per node. PON FTTH does not have these same upstream bandwidth limitations but has been, until recently, difficult to integrate into existing HFC networks.
Several electronics manufacturers have available or are in the process of developing and offering D-PON (DOCSIS PON) solution sets that give the MSO the advantage of utilizing their existing RF, QAM, DOCSIS network with today’s PON standards.
Deployment costs. Initial deployment costs are a primary factor in choosing which architecture fits the need. It is always cheaper to make upgrades and changes to existing plant than rebuild unless the effects of your upgrade have a limited financial life. Then, the operator must consider the consequences of spending money on upgrades versus spending that same money on next-generation technology. The cost of deploying FTTH in the greenfield is approaching parity with the cost of deploying greenfield HFC.
Figure 4: FTTH – cost per home connected.
The deployment costs of HFC are flat and well known. FTTH costs in years past were high, relative to HFC, but continue to fall over time. Figure 4 shows how the cost per “connected home” has fallen in FTTH. The HFC plant must deploy all electronics before turning on the first customer. The FTTH plant only deploys an ONT when a customer takes service and new OLTs after service saturation. Labor and actives comprise the vast majority of the plant deployment cost as shown in Figure 5. By only having to deploy electronics where a customer has taken service, the initial deployment costs are the labor and passive components. According to Kagan Research, the December 2006 penetration rate for MSOs was 58.8 percent of the homes passed, yet 100 percent of the electronics were deployed in the HFC network.
Opex. Operational expenditures are the ongoing yearly costs of plant operations after the initial install. When considering profitability, two components become obvious: labor and power issues.
Figure 5: Deployment costs.
Source: Corning Cable Systems
Today’s HFC plant is approximately 70 percent coax and 30 percent fiber in terms of plant mileage. The MSO need only look at the number of fiber technicians versus the number of coax technicians employed to see that, on average, fiber requires much less labor per mile than coax to maintain. This is intuitively obvious since the coax portion of the plant has active components and power supplies that must be maintained at standard levels.
The difference in opex between FTTH and HFC in terms of power is striking. The FTTH PON is totally passive. Power supplied to the ONT is often taken from the subscriber’s house so the only power costs are the headend. The power consumption of an HFC plant has become an accepted yearly drain on financial budgets. Power consumed from the local utility grows every year and the price of power increases. Powered items (including power supplies) are susceptible to heat, lightning and moisture; thus the maintenance costs for labor, repair and replacement of these items must be included in the overall cost comparisons.
3. Enabling flexible networks through passive optical components
Preconnectorized assemblies in both HFC and FTTH have proven their value by increasing plant flexibility and lowering deployment and maintenance costs. These products can be used to start-up a system in a fraction of the time required to complete a system build component-by-component. The labor savings gained from eliminating repetitive and time-consuming fiber preparation is greater than any additional material cost. These same labor benefits also decrease the downtime risks associated with the lack of fiber management and support longer operational lifetimes as there is significantly less product handling involved.
Table 1 shows an example of the differences in resource requirements for installing a preconnectorized system versus traditional splicing in an FTTB
From a project management standpoint, preconnectorized plug-and-play cable and hardware solutions provide the most efficient system to install.
Preconnectorization promotes a higher level of support in servicing current customer requirements and future growth. Due to their expanding role in various access applications, modular outdoor plug-and-play solutions are becoming increasingly commonplace and easier to specify and order. Having fewer parts also aids in simplifying inventory management by reducing the number of individual parts ordered and stocked. The burden and cost associated with ordering and maintaining consumables over the duration of the installation, as well as the associated tool kits or other specialized equipment needed for field termination, is also lowered.
Table 1: Comparison of labor requirements by method. Source: Corning Cable Systems
Because no preparation is needed for the cable ends, no special training is required. The potential that the system start-up will be delayed as a result of installation-related errors is greatly reduced. Performance-wise, field-assembled installations will not produce a product with the same consistent, premium, 100 percent tested and guaranteed performance as assemblies built by a high-quality manufacturer.
Form follows function. Understanding future bandwidth requirements forces the fundamental change in architecture. The next five years will require bandwidth beyond today’s HFC capabilities. While some modification of the outside plant can extend the life of HFC a few years, capital investment is better utilized deploying the long-term solution of fiber. Preconnectorization of the fiber plant not only reduces the high cost of labor during deployment but will save money in the long term through faster maintenance and faster customer service. Functionally, tomorrow’s outside plant needs to be high bandwidth, flexible, low opex and low maintenance at an installation price comparable to today’s HFC. The form of that plant is preconnectorized fiber.