Save Green with AFAN
AFAN solutions offer the potential for significant opex savings
For cable operators, opex is the gift that keeps on giving. But rarely does it give back. All-fiber access networks (AFANs) offer a good news story on two fronts.
All-fiber solutions require less power than their hybrid fiber/coax brethren, with the potential to save cable operators millions in opex dollars. This comes largely from the elimination of amplifiers and, in the case of an all-passive AFAN, nodes.
And when taking into account the impact of headend and subscriber equipment, such as an RFoG optical network unit (ONU), the total power consumed by the network is also less, making AFAN a green technology, as well. So it is possible to both save green and be green with AFAN. To understand how this works, we can analyze the power consumption in HFC and AFAN networks, the impact of take rates and who pays for the power consumed.
NETWORK POWERING CONCEPTS
HFC network powering is dependent in large part on its design and construction. Traditionally, the HFC network is an “upfront build,” meaning the major elements (nodes, amplifiers, power supplies, coaxial cable and taps) are installed upfront, even though there may be few or no initial customers on the network. However, the HFC network is completely turned on and powered by the cable operator’s opex dollars, regardless of take rate.
AFAN networks can be deployed in two ways. The most economical is a fully passive access network – “direct feed” – with no active (powered) devices between the hub and/or headend and subscribers. In this case, there is a slight increase in power at the hub/headend for additional lasers/receivers. In the field, the only powered device is the subscriber ONU, which is: a) powered by subscriber dollars, and b) scaled directly with take rates.
AFAN networks may also use an active device in a serving area, the need for which is driven by distance (greater than 20 km), available fibers (less than needed to support a 1 x 32 ratio) or a need to use a higher split ratio (1 x 64). This approach introduces a powered device into the field, but the network is still considered an all-fiber solution. Because the “nodes” in this approach can typically serve as many as 256 subscribers over distances of 10 to 20 km, power consumption remains much less than that of an HFC system with multiple nodes or a cascade of amplifiers. Analysis
The easiest way to compare HFC and AFAN is to examine typical power consumption levels for each network approach. In this article, we do this based on kilowatt-hours (kWh) to see the differences between approaches. Using local power costs, kWh can then be converted to dollars. For the analysis presented here, a national average electrical power cost of 10 cents per kWh is assumed.
The chart in Figure 1 details the assumptions built into the analysis of both HFC and AFAN networks.
• Because AFAN allows scaling in 1 x 32 groups, headend power is shared with this neighborhood and others. Headend power accounts for downstream lasers (EDFAs) and upstream receivers.
• An HFC power supply is dedicated to the neighborhood analyzed, containing one node and a cascade of amplifiers.
• Subscriber power does not include set-top boxes, cable modems and other devices that would be common to both HFC and RFoG/AFAN deployments. Here, the HFC network is assumed to drive a coax house splitter (without a house amp), and AFAN is assumed to have an RFoG ONU.
• Any variations in HFC node/amplifier powering associated with take rate are assumed to be small and are not considered in this analysis.
Using the input parameters described above, each network was analyzed for total power consumed across take rates from 0 to 100 percent. Total power consumed includes headend power (for transmit/receive equipment), power for field equipment (such as nodes) and power at the subscriber’s premise. Power consumption for AFAN was broken down to show the network power (cable operator’s responsibility) and R-ONU power (subscriber’s responsibility).
Figure 2 shows the results of the analysis in graphical form. The yellow line across the top of the graph represents the total power consumed (headend and node/amplifier) over the course of a year for the 250-home analysis area for a typical HFC deployment. The line is flat because the network requires basically the same power, regardless of the number of subscribers in the service area. The cable operator funds this entire amount from opex dollars.
AFAN (using RFoG technology) is represented by two lines on the graph. The green line at the bottom is the power consumption at the headend. In the analysis, it is actually more than the HFC headend power because more receivers and lasers/EDFAs are required to feed each operating splitter group in the serving area. It is ramped up gradually as take rate increases to show that more and more headend ports are dedicated to this neighborhood. The greater amount spent in headend power is far overshadowed by savings in the outside plant equipment. In this deployment, there are no actives in the field until the subscriber’s premise is reached (the R-ONU). Since R-ONU power is funded by the subscriber, the cable operator is only responsible for the green line, which ranges from 2 to 14 percent of the yellow HFC line. The blue line represents the cable operator’s baseline power (headend) plus the incremental power consumed as subscribers (R-ONUs) are added and the take rate increases. Therefore, the subscriber-paid portion – the lion’s share – of the power is the difference between the green and blue lines. In this example, RFoG power consumption only reaches parity as take rate nears an uncommon 100 percent level. At traditional MSO take rates of around 50 percent, the power consumed by the AFAN network is about 54 percent of that of the traditional HFC network, and the cable operator’s portion is about 7 percent of that of HFC.
VARIATIONS ON A THEME
Naturally, there will be variations on the AFAN theme. For example, what if a remote hub device is needed due to the distance to the serving area being greater than the 20 km reach common to RFoG and other technologies? In this case, the headend resource is similar to that needed for HFC, and the remote hub, which accepts optical signals and sends them on to the subscriber, must also be powered. So the remote hub must be modeled because, in effect, the receivers and EDFAs that were in the headend are now in the field. The difference is that for the same 250 homes passed, there is only one powered device in the field, compared with the HFC node and its cascade of amplifiers. This changes the picture by raising the baseline that is powered by the cable operator, becoming approximately one-third that of HFC (versus 7 percent above). Despite this increase in base power consumption, the total power consumption for AFAN (including R-ONUs at the subscriber premise) does not reach parity with HFC until a 70 percent take rate is achieved.
Another variation is to consider different densities. For example, assuming the same geographic area in our example now has 300 homes passed that are served from the same node, the same HFC power now passes more homes. In AFAN, the headend value changes only slightly, but the total power from subscribers goes up. With the same take rate, there simply are more of them. In this case, parity is achieved at an approximately 75 to 80 percent take rate. The inverse is also true: Reducing density in the same geographic area to 200 homes means the same HFC power is spread over the same plant miles with fewer homes passed. For AFAN, the baseline declines very slightly, while there is less power due to fewer subscribers. Fiber experiences virtually no impact on power consumption unless a remote hub becomes necessary. Modeling at 200 homes in the same area shows that AFAN power consumption remains below that of HFC at all take rates and does not reach parity.
WHAT DOES IT MEAN?
For the cable operator’s opex budget, significant savings can be achieved with AFAN. Looking back at Figure 2, the baseline for AFAN at a 50 percent take rate is about 600 kWh per year for headend power, while HFC requires about 8,000 kWh in total. At 10 cents per kWh, the power reduction is worth $740 annually over 250 homes passed. For an operator with 1 million homes passed, that translates into a savings of nearly $3 million per year. Over a period of 20 years, this would add up to almost $60 million. While this analysis covers just power-related savings, deployment experience over the last six to seven years shows additional savings from increased reliability due to fewer active devices and fewer trouble calls. In addition, routine maintenance items such as leakage testing and network balancing are eliminated.
In addition to the operator’s savings, the fact that AFAN solutions generally require less power than HFC networks at typical take rates makes them an overall “greener” solution. Opex savings for the operator are in alignment with a smaller carbon footprint for the environment. The green footprint of AFAN does not stop at power consumption.
Eliminating truck rolls for routine CLI testing and troubleshooting reduces vehicular pollution and operating costs and allows technicians to focus on other priorities. Fewer truck rolls for trouble calls also reduces pollution and costs.
Because AFAN can operate without active devices in the field, the need for backup power generation equipment and fuel expenses can be eliminated.
At a time when national bandwidth targets are being set to endure well into the future, these savings are available in a platform that also offers extremely highperformance capabilities.
While HFC networks will continue to be the mainstay of cable operators’ networks for some time to come, AFAN solutions offer the potential for significant savings in opex costs. As new builds and major rebuilds occur, operators can now leverage AFAN solutions. Since the same passive AFAN solution will support RFoG, GPON and EPON technologies, cable operators can choose the solutions most compatible with their current networks and enjoy the opex savings. At typical take rates, AFAN power consumption is well below that of HFC networks. Less power and less maintenance add up to reduced recurring costs for the cable operator and a much smaller impact on the environment, too.
Author Mark Conner wishes to thank Shawn Esser, Fred Slowik and Dean Stoneback of Motorola for their contributions with writing this article.