Prospective RC Broadband Project Addendum

More Project Information

Here is more information developed since the earlier Broadband Report, or included from that to help this report be independently useful without needing a copy of that too. All of this will be copied to the project website soon.

No Utility Pole Fiber Deployment

Frontier and SuddenLink (indirectly) sued to challenge the legality of HB3093 section 4, claiming the simplified make-ready pole access rules conflicts with FCC rules and infringes on corporate property rights. Meanwhile the FCC is considering a rules change, anticipated to be some variant of one-touch make-ready, to reduce broadband deployment costs. So use of this provision is in limbo, as are projects dependent on utility poles.

Also, the prospective project in the Broadband Report did not account for utility pole rental costs, roughly $7 to $10 per pole per month. Instead it aspired to come to some arrangement with power companies for low or no rental fees, in exchange for use of the co-op network to deploy smart grid hardware and/or for volunteers helping to maintain utility line right of way. Now its all but certain First Energy would not agree to either offer.

Buried cable is lower cost long term, when accounting for pole rent, storm damage, reduced cable service life from environmental exposure, and squirrels chewing on them.

More Fiber Bandwidth, Lower Cost

There were several significant technology improvements over the past few years, with hardware coming out now and over the next few years. This was missed in the Broadband Report due to a focus on case studies, which are historical in nature. Three will have a significant influence on this project.

Laser Modulation: Until recently low cost fiber transponders used Intensity Modulation with Non-Return to Zero (NRZ) signaling, put simply, “dim light is a 0, bright light is a 1”, up to 10 billion dim-bright 1-bit “symbol” cycles a second (1GHz ~ 1Gbps, 10GHz ~ 10Gbps). This plateaued for years, reducing the cost of everything up to 10Gbps. Now its being replaced with Pulse Amplitude Modulation, it supports larger symbols, of 4bit, 8bit and 16bit today, 32bit in years, and above. Unfortunately this introduced errors, so there is higher error correction overhead, e.g. 1GHz x 16bit is ~11Gbps, instead of 16Gbps, or 10Gbps with old 10Gbps hardware. Likewise 10GHz x 16bit is ~120Gbps, instead of 160Gbps, but platform to supply 100Gbps is new, expensive. Benefits:

Faster, Lower Cost FTTx: Expect 10Gbps PON ports serving a larger number of 1Gbps links, while also supporting 10Gbps links. Lower cost than an equivalent 4 to 10 ports on old hardware, but higher cost in low density areas where the high split ratio cannot be achieved, worth it for 10Gbps support.
Faster, Lower Cost Backbones: Expect faster backbones, probably 100Gbps (4x 25Gbps), this is now what 10Gbps was a decade ago (and 10Gbps is now what 1Gbps was a decade ago). 200Gbps (4x50 or 8x25) and 400Gbps (8x50) transponders are available, but much more than an incremental increase in cost. 100Gbps per channel transponders exist but are very expensive.

Error Correction: Forward Error Correction uses bandwidth to correct errors, its required for higher bit rate modulation but can also be used to increase range in combination with an Optical Amplifier. Dispersion Correction compensates for dispersion without needing to use corrective optics (e.g. fiber coil with opposite dispersion shift).

Optical Amplifiers: Erbium Doped Fiber Amplifier, a 980nm or 1480nm booster laser injected into a short length of special fiber causes light passing through in 1550nm range to get brighter. Seems like this can turn an otherwise dark strand into a power line, used to boost power on other strands where needed. The booster laser doesn’t carry data so the only issue is optical power, not signal loss. This can increase the range of PON lines and backbones.

Increased Range: Combined benefit of the error correction and optical amplifiers is greater flexibility in network engineering. The Broadband Report noted 10 mile range (16km), which is actually 20km with a large dB loss budget, now 20 miles (32km, really 40km) should be attainable. This means fewer roadside cabinets needing power and more flexibility in selecting POP locations, such as by community benefit rather network topology, which is critical if a POP is a new community center.

“Coherent Optical” Technologies: Next generation fiber optics, all very expensive today but will allow the co-op to transform its network in the coming years. Optical switches allow for directly routing fiber channels, without optical-electrical-optical conversion, and routing around cut lines. Modulation, polarization, error correction, signal processing, and better processors will increase bandwidth 10x then 10x again, making 10Gbps residential service with 1Tbps backbones practical. Ranges will increase 10x for direct P2P links, less for PON due to noise.

Wireless Deployment

Wireless strengths are wide area broadcast, mobility, lower upfront deployment costs, and low temporary deployment costs. Fiber strengths are higher bandwidth, longer service life, it can snake around terrain obstacles, immunity to interference, fiber may cost less over decades than successive generations of wireless in low-power point-to-point links (our usage). Wide area broadcast is generally not an option for the types of deployments needed to bring service to our rural areas. We have cell coverage over most of the county but anecdotally most of the rural residents have no service at home due to terrain, so for long range fixed-wireless to work here everyone would need to put up masts to get above terrain.

The general plan is the co-op will deploy fiber as far as feasibility and funding allow, then temporarily deploy wireless until funding is available to replace that with fiber. Here are some notes on what that might look like in specific scenarios, and the ugly yet effective alternative of surface running cable as demonstrated by Frontier.

Clusters of Houses (early): Ideas for early wireless deployment are focused on clusters of houses within line of sight of each other. Providing them early temporary service can done with either a house to house mesh (lower cost shorter range), or a mast near the fiber to a mast near the house cluster (higher cost longer range). Expected service life is a few years, but may remain long term if residents buy the equipment.
Lone Houses in Curved Hollows: Seems every little hollow in Roane has (or had) an old farm house, some are back a mile, and many with enough curve or hills to block wireless line of sight to the public road. These are our worst cases, and anecdotally around a tenth of our rural households. Providing them service requires either a mast on a ridge top plus a mast at the house, or a daisy chain of repeaters on utility poles, or a zigzag of repeaters on posts on the sides of the hollow. Its all expensive, so much so surface running armored fiber is both less expensive and probably more reliable than repeaters.
Ring Around A Hill: Seems larger hills tend to have houses on all sides, spread out so none have line of sight with the others, and tight to hill sides or hollow walls. Providing them service is another geometry problem ~ hilltop mast plus a mast at each house, or hilltop mast plus repeater set out from each house, taller masts at each house to link to neighboring houses, or a ring of repeaters around the hill. Its all expensive, again, surface run armored fiber is both less expensive and probably more reliable.
Remote House Clusters: There are old remote villages, and split up family farms, with a tight cluster of houses but nothing else for miles out to a public road. Providing them service requires either a pair of masts with a wireless backbone then local wireless links to each house, or the same masts with local FTTH. This is cost effective, reliable, and can provide gigabit service, but surface run fiber may cost less than the masts.

Note, while its suggested repeatedly, we probably will not “surface run” cable, but will happily sell it to people that decide they would prefer to do that instead of a wireless mast (with a signed liability release). Surface cable should be fine as long as its not run over by vehicles, hit by mowers, and is protected from direct sunlight ~ forest, edges of fields, inaccessible stream banks opposite roads, etc.

Early Broadband Access

Currently rural residents have high speed internet access via libraries and wifi in the parking lots of a few businesses in Spencer. These have usage limits, limited schedules, may not provide broadband speeds, and have significant travel cost for outlying residents. To help these residents as early as possible, the following can make better service more accessible, before network construction begins, and maybe before a co-op is officially created:

Community Centers: Upgrade existing community centers with broadband, computers, and wifi hot-spots. Organize programs, staffed by volunteers, for these centers to remain open after regular business hours. Grants and loans are available to fund setup costs, and build community centers where needed, but it all starts with community organizing.
Wifi Hot-spots: Work with anchor institutions and area businesses to increase wifi availability. Work with co-op members to host hot-spots for use by other members.
Co-op Offices: Open an office as early as possible with broadband, a wifi hot-spot, a small common area with some PC’s, a private room with teleconferencing equipment, printers, etc. Charge minimal fees to cover costs, request donations of both money & equipment to upgrade facilities. Recruit techs to offer tech services to members, and refurbish donated equipment. Use this first office as an incubator to refine this model then replicate it around the county, in places without community centers able to host these services.

These community centers and later co-op offices (not the first) will serve as Points Of Presence (POP’s) with a lot of fiber, network hardware, backup power, and support for network construction & ongoing maintenance.

Volunteers & Service Credit Overview

Summary: Co-op members are required to volunteer 20 hours a year, or pay a $20 per hour non-volunteer fee ($400 per year). Work over 20 hours is paid with $10 per hour service credit, plus expense reimbursement. During normal operations this is cost neutral with fees covering volunteer costs. During network construction this is expanded into a form of local micro financing, with interest on carried credit balances (e.g. half bank rate), in the hope more members will volunteer a lot more hours, and that people will invest in the co-op via buying service credit. Ideally this will result in a local market within which service credit can be used like a quasi money. If more volunteers are needed, add a personal expense stipend or minimum wage. Once the startup bank loans are paid off the service credit can be cashed out.

Member Volunteers: A common way for cooperatives to fulfill labor needs is to require members to volunteer a certain number of hours per year and charge fees for not volunteering. Those fees are used for operating expenses of volunteers, and to reward members which volunteer beyond their annual requirement, and/or for paid labor on remaining work needed. This model is most common in small rural agricultural cooperatives, but should do well in regular operations of a rural county broadband co-op.

Labor Needs: The largest obstacle to building a broadband network is cost, the largest part of that cost is labor, and a lot of that labor is of various types suitable for volunteers. With a community centered approach, labor needs may include local coordinators (managers), local deployment planning (survey teams), local work crews, local user groups (tech support & education), and more. County-wide volunteer roles include, assisting staff, various non-professional roles, or support communities without enough volunteers. The labor needed during construction is beyond annual requirements of members, and its unlikely the co-op can afford to replace it with regular paid labor.

Labor Financing: Extend the volunteer paradigm to fulfill the co-op’s needs by offering to pay members for “volunteer type work” with credit on the co-op’s services. There are historical examples in old WV mining towns, with company script (credit) for pay that was only usable in company stores – the problems relating to this practice were because the companies were not employee owned cooperatives with democratic management.

Local Micro Financing: This service credit system is a local direct micro financing mechanism. The benefits of this type of financing are it keeps money local, reduces finance costs, gives members direct control. Being tied to work limits availability to volunteers and staff. Selling service credit to residents allows them to invest and support the co-op, even if they don’t have time to volunteer.

Quasi Money: This co-op company credit would be transferable and is of value to most county residents. With support of several local merchants, and a payment gateway, this credit could be turned into a local quasi money. Its backed by co-op & faith in its membership – in effect, backed by the community’s faith in itself.

Service Credit Usage: Credit can only be used on regular service fees. Cannot be used on non-volunteer fees, taxes, or installation loan payments. For example, if a bill were $100 service, $40 install loan, $34 non-volunteer, $6 tax, $180 total, with $1000 in credit on the account, the minimum money payment due would be $80.

Service Credit Management: The service credit redeemed per co-op billing cycle is an income-side loan payment, this risks causing a cash flow shortage. Some tools to manage this risk include, projections & budgeting reserve funds, an opt-in delayed credit usage program, member advisories requesting members to pay bills with money, selling more credit to people who will pay bills with money & buy back credit from people that pay with credit, credit paid work stoppage, short term bank loans to cover short falls, membership vote to impose credit usage limits on those most able to pay money on a limited emergency basis. Service credit must be inviolable up until the threshold of insolvency. Once all startup bank loans are paid off, the co-op will pay off service credit.

Proposed Numbers: Figures suggested in the earlier Broadband Report were, $10 per hour service credit (time sheet) plus reasonable documented operating expenses (figured $5 per hour). Paid staff included the same as an offset to pay – e.g. instead of $15 per hour its $10 per hour plus $10 service credit per hour, plus any other benefits. The member volunteer requirement is 20 hours a year, they only receive service credit over 20 hours, as membership dues this is valued at $200 per year. The non-volunteer rate is double, $20 per hour, up to $400 per year, paid monthly. Members without service and that do not participate would be offered a lower “member in name only” sort of status, forgiving the $400 per year fee and suspending most member benefits.

Extra Pay: There are concerns people will not be able to afford deferred compensation and the co-op will not be able to attract enough volunteers during construction. In this case, consider adding a personal expense stipend for “reliable regular” volunteers performing above & beyond the relatively low expectations. Consider adding minimum wage for a more experienced group between regulars and entry level professionals. This extra pay would need to be short term only, long term cost is intended to be balanced against people paying the $400 fee.

Super Co-op Service Credit: The super co-op would issue the credit and manage it throughout the region. Member co-ops would swap their credit for super co-op credit, and use that to pay people, sell, or spend. Super co-op would maintain a larger cash pool and bank credit to ensure overall liquidity, like regional treasurery.