The Palo Alto Fiber to the Home Trial

The Palo Alto Fiber to the Home Trial A Work in Progress Ken Poulton Palo Alto Fiber Network

Topics • Background • The Palo Alto Fiber to the Home (FTTH) Trial – – Network Plan Finance Marketing Lessons Learned • Challenges and Opportunities – Inexpensive Fiber to the Home – Open Networks 2

The Setting: Palo Alto • Mostly affluent • Adjacent to Stanford University – Long-standing emphasis on education • City-owned utilities (electricity, gas, water) • High proportion of engineers and other Silicon Valley types • Palo Alto Internet Exchange (PAIX) • City-owned dark-fiber ring (completed in 1997) 3

Existing Palo Alto Fiber Ring: Route Map Underground Overhead 4

The Setting - Problems • City-owned dark-fiber ring – Fell short of (unrealistic) cost recovery goals • Subscriber-owned local cable TV company – Long history of financial problems – Being sold to AT&T • City government – Very cautious staff and city council – “Process-Rich” (i. e. , slow) 5

Origins of the Palo Alto FTTH Project The Naïve Question: “Why can’t I just hook up that fiber to my house? ” -- Residents The Visionary Statement: “Fiber to the Home is not a question of if, but when. ” -- Brian Reid 6

Palo Alto Fiber Network • • Volunteer organization of FTTH enthusiasts ~200 members ~20 members working actively on FTTH Functions: – Education of city staff, city council, the public – Providing networking expertise to staff – Organizing the political effort – Assisting with marketing effort 7

Palo Alto Fiber Network Goals • True high-speed network access for everyone – Scalable (user choices and future expansion) – Affordable • Open Network – Promote competition – Support a variety of services – Encourage local content providers and services • A fiber connection to every building in the city – Homes, schools, businesses 8

A Simple Plan • Build a small trial first • Focus on a data-only network • Use inexpensive, proven, off-the-shelf technologies – 10/100 Mb/s Ethernet – multimode fiber • City to build and own the network • Hire an existing ISP for operations • Cable TV, phone, other services in the long run – may be added sooner if they help the economics 9

Why Fiber to the Home? The biggest Internet bottleneck is the Last Mile to the home. • Use is growing: More users, more uses, more frequent use. • Richer content: graphics, sound, video, bloat-mail. Fiber can deliver an unbeatable price/performance ratio. 10

Consumer-Level Internet Access Technologies vs. Year Per-user limit with existing wires Any technology will need considerable new infrastructure investment to go much beyond 2 Mb/s per user. But only FTTH allows inexpensive further upgrades. 11

Cost of Construction vs. Year of Construction The construction cost factors favor doing it right now. 12

Why Now? • FTTH construction cost no longer dropping rapidly – Electronics now only 5% of system cost for 10 Mb/s • The market is ready – Telephone modems have reached the 56 Kb/s limit – Users starting to move to medium-speed (~1 Mb/s) services • There is a window of opportunity – Most attention focused on squeezing out the last dregs from existing copper infrastructure – FTTH is a natural monopoly - it will be uneconomic to duplicate this infrastructure. Þ The first provider to build an open Last-Mile fiber infrastructure in a given area will be the last for decades 13

Palo Alto FTTH Trial Topology 14

Typical Pole to Home Wiring Home installation is similar to cable modem but uses fiber. 15

FTTH Trial Costs 10 Mb/s Service to 100 Homes Construction Cost Estimate: $630 per home passed +$830 per home connected (+$380 for 100 Mb/s) Total: $350 K Operations Cost Estimate: $7/month for physical maintenance $25 -50/month for Internet Access, support, ISP services 16

Per-User Capital Cost vs. Participation Rate Target for Trial: 24% 17

What are the Building Blocks of a Network? • Customers – Residential, Commercial, Academic, Civic, Special Interest • Services – e-Mail, Content, Web Hosting, e-Commerce, Education • Internet Access Competition Possible – Internet Access and Transport • Network Operations – Routing, Traffic Control, Security, Billing, Customer Support • Physical Infrastructure Natural Monopoly – Wiring, Poles, Easements, Splices, Switching Equipment 18

Trial ~100 homes Who Does What? City-Wide • Customers – Residential, Commercial, Academic, etc Any ISP • Services Single IAP/ Network Operator • Internet Access City • Physical Infrastructure – e-Mail, Content, Web Hosting, e-Commerce Any ISP Qualified IAPs – Internet Access and Transport • Network Operations – Routing, Security, Billing, Traffic, Support – Cables, Poles, Easements, Switch Sites For the trial, use a single IAP/Network Operator to be cost-effective. 5, 00026, 000 homes Network Operator City/ Private City-wide system will be an open network. 19

Financial Model For the Trial • Ownership: – City builds and owns the physical network – City chooses an ISP to provide operations and services • City Council’s Financial Choices: – Recover construction costs within 5 years – Subscribers commit to repay 2/3 of the cost before construction • Resulting Offer to Residents: – $1200 installation fee – $45/month to city + $25 -50/month to ISP – 2. 5 -year commitment to the service 20

Marketing Results • August ‘ 98 Survey: – A single utility-bill insert and a few ads – Yield: a 4% city-wide signup rate in just 4 weeks (compared to 4% use of cable modems in 4 years). – 19% in two areas with neighbor-to-neighbor campaigns • September ‘ 99 Trial-Area Signup: – 2 letters from the city (with ISP rates still not certain) – Yield: 9 -15% signup rate • December ‘ 99 Trial-Area Signup: – Will have ISP rates defined and neighborhood campaign – Goal: 24% participation 21

Timeline for FTTH in Palo Alto • • • Fiber Ring built Begin Trial construction Begin Trial operations First evaluations Decide on a city-wide system – – 1997 $2 M Q 1 ‘ 00 $0. 4 M Q 3 ‘ 00 Q 1 ‘ 01 2001 Is FTTH worth doing? Should the city be involved? Should private partners be involved? Can sufficient political will be mustered? • Deploy city-wide system 2002 $25 M 22

Hurdles (and how we passed them) • Right-of-way ownership à Have the city be the builder and owner of the network • Need for fairness among neighborhoods à Expand the focus from one neighborhood to citywide survey • Negative, incorrect, initial staff report à Wrote a technical and budgetary plan using staff numbers • Council desire for zero financial risk to the city à Small trial, users commit up front to pay for the system • Not enough subscribers without firm ISP cost numbers àGoing back to finish signups after ISP signed up 23

Lessons Learned • We found huge grass-roots enthusiasm for FTTH – High speed – Open network – New services • City ownership is very attractive to residents • City ownership is scary to city staff and council – Educate citizens, city staff and representatives. – Advocates must remain engaged with city staff. • Cities do not run on Internet time 24

Fiber Choice: Current Costs vs. Long Term Flexibility • Lowest cost today: – neighborhood switch sites serving ~1000 homes, distances up to 2000 meters – multimode fiber (cheaper splices and converters) – data only (10 Mb/s Ethernet, 100 Mb/s in a few years) • But we may need single-mode fiber for CATV and 1000 Mb/s over >500 meters – Single-mode splices and media converters add ~$1000 per subscriber to the current costs – CATV electronics add another ~$1000 25

Technical Opportunities Taking FTTH From Attractive to Irresistible • Cheap (~$50) single-mode fiber media converters for Ethernet and CATV • Cheap wave-division multiplexing components • Pole-mountable, non-air-conditioned switches and media converters – could be within 500 m of homes, so multimode is enough 26

Open-Network Challenges • Technical Implementation – No clean solution to multiple-ISP network yet • Network Business Model – So far: Ownership = Control = No competition – Opening monopoly networks via regulation ineffective 27

Open-Network Paths • Public Ownership + Most direct way to ensure an open network – Risk to public funds – Political battles to get started • Private Ownership + Can move more swiftly + Existing networking expertise – No proven model that benefits from maintaining openness • Public/Private Partnerships + Could have the best of each – No proven models yet 28

Summary • FTTH is coming sooner or later; sooner is better. • Open networks are a major benefit to the public. – FTTH is a natural fit for open networks – Public vs. private ownership choices • Room for innovation to make FTTH more competitive. • We found lots of support for publicly-owned FTTH in Palo Alto. • We hope to demonstrate viability of FTTH in the coming year. 29

Thanks to. . . • Palo Alto Community Network – For starting the discussion • Brian Reid – For vision and inspiration • Residents of the Community Center Neighborhood – For leading the way • Palo Alto Fiber Network – For volunteers supporting FTTH throughout the city • Palo Alto City Council – For funding • City of Palo Alto Utilities Department – For doing the work and taking the heat 30

References • Palo Alto Fiber Network site: www. pa-fiber. net – Major Contributors: Mike Eager, Ken Poulton, Peter Allen • City of Palo Alto FTTH site: www. cpau. com/fth • Slide: “Do U. S. Homes Really Use the Internet? ” – See www. cyberdialogue. com/isg/timeline/forecast. html for results of a FIND/SVP survey estimates and projections. This corresponds to the following government survey: – • • See www. ntia. doc. gov/ntiahome/net 2/charts. html for details on the “The Digital Divide, Net II” survey released 7/28/98 by the National Telecommunications & Information Administration. Slide: “Why Fiber? ” – Snapshot from 11/99 of each listed service provider’s price structures. Slide: ”Palo Alto Fiber Backbone Route Map” Slide: “Typical Pole to Home Wiring” – Source: City of Palo Alto Utilities Slide: ” FTTH Trial Costs” Slide: “Cost of Construction vs. Year of Construction” – Source: City of Palo Alto Utilities, ‘Fiber To The Home Trial Cost Estimates. ’ – Analysis: Ken Poulton, ‘Palo Alto Fiber To The Home Trial Technical and Budgetary Report. ’ (www. pa-fiber. net) Slide: “What are the Building Blocks of a Network? ” Slide: “Palo Alto Fiber Network Trial “ – Source: Peter Allen 31

Palo Alto FTTH Network Phases • Build the Backbone (1997) • FTTH Trial (Q 3 2000) $2 M $0. 4 M – Refine cost estimates and design – Measure user satisfaction, participation rate – Make recommendations for a city-wide system • City-wide Rollout $25 M – Market competition – New services 32

Consumer-Level Internet Access Technologies vs. Year with existing wires Any technology will need considerable new infrastructure investment to go much beyond 2 Mb/s per user. But only FTTH allows inexpensive further upgrades. 33

Detail of Costs for Services 34

Expected Results of the Trial • • • Demonstrate that FTTH is practical and pays for itself. Refine the construction and operational cost models. Work out operational details and user support. Measure user satisfaction and willingness to pay. Enable new applications that are currently bandwidth-starved. • Increase awareness, demand financial justification for a city-wide FTTH system. • Reduce uncertainties and risks of a city-wide FTTH system. 35

Why Ethernet? • It’s the standard - used in most offices in the world – 10 Mb/s is the least expensive kind of network now – Familiar to all ISPs • It’s easy to upgrade later – Many companies are creating new Ethernet products – 100 Mb/s will be cheap in 3 years, 1000 Mb/s in ~8 years • It meets the whole spectrum of data service needs - now and into the future – 10 Mb/s provides enough speed for >90% of home uses – 100 Mb/s option can support virtually any use today – Room to grow as demand grows 36