Michael Silverton spoke on his Stanford thesis "Social Informatics of the Information Superdriveway", deploying residential community fiber optic networks. His interest in this question started in 1991. What is it going to take to deploy residential networks? There are a number of proposed solutions to deploying high speed access to homes. How will this affect Palo Alto? Is there one best solution or will there be a matrix of solutions.? A recent article by Jesse Berst on Ziffnet predicted that in 2000 80% of users will still be using 56K or slower modems unless there are significant changes in deployment. Even Berst's best estimate was 40% using 56K modems in 2000. There are 3 behaviors that have to be considered. There is a mounting desire for more and faster access, there is a mounting consumer interest that hasn't been recognized yet, and there is a difference between monopolistic and competitive markets. These things have to be grappled with. We need to get residential users tied to Internet I, before aggressively deploying to Intercom 2. Moving off [Stanford] campus, the drop in speed is very noticeable. He's looking for 10 to 100 Megs/sec. access. Getting there will move closer to fiber to the curb, which is considered by many to be the holy grail.
His objective is to deploy a working EdCom (educational community) model in single family neighborhoods. EdCom International is a business idea he has that is similar to Smart Valley. He wants to put together a valid model. Major interest is single family neighborhoods, not businesses or places like hospitals. In a hospital there is just 1 building to wire. Single family units present a particular type of problem, and they need to be addressed. It could showcase the Palo Alto fiber ring and model universal service.
Right now we have fiber to the bus stop. It is an incredible accomplishment in itself. The city has implemented the strategy and expects to recover costs by leasing access. Cable Co-op offers high speed cable access but the cost is twice as expensive as Time Warner Roadrunner in San Diego. There also are other entities that are dealing with the issue, such as PAN (Public Access Network). The computers in City hall have been broken most of the times he visited.
Jay Thorwaldson and Harry Saal noted that PAN stopped working on those computers. They were demonstration operations only. PAN no longer is maintaining the system.
Michael was very impressed with Plugged In located in East Palo Alto. He was directed to the site by asking at random while driving in the area, so people in EPA are aware of the operation. He found it very good. A young man showed him around and helped him out. They seem to have great publicity. There are many other groups such as SV-PAL that are trying to help out. For example they are looking at how to get 9600 baud modems to people.
[Current residential networks are too slow and too expensive to set up.] Waiting for pages to load you waste time. The longer it takes to load the less time you have to do useful work. There will be market pressures to change the residential data environment, for example landlords in apartment houses could exert market pressures to install high speed lines to be able to attract tenants and charge extra for the service. There will be cost and technological pressures to provide the high speed access. Organizations like Smart Valley have done an excellent job getting systems started and educating people.
Harry asked if availability of high speed is the major factor limiting use of residential networks. It isn't a major factor, but it is an underestimated factor. It isn't the speed itself but what it enables to be done. People get very impatient just using the network. Waiting 10 seconds for data to load can be far too long. Some people are very happy using the Internet only for e-mail. Others use it less because they don't have high speed. Bob & Jay both gave examples of wanting higher speeds. The cable modem makes a huge difference in ability and desire to download big files. Nevertheless some people will be very happy with current speeds. Stan Smith noted that there is a price-performance tradeoff. People who want speed will pay for it, those who don't care won't.
Michael noted there are 2 issues in that - what people want and what they want to pay for. He wants to talk about lost opportunities with the present analog network - reading morning news is harder, so it reduces ad opportunities, slower searching and posting jobs and resumes also loses ad opportunities, school-home partnerships, new kinds of cooperation between students, teachers, parents. Telecommuting opportunities are reduced and are unequal. People with faster Internet access are more likely to get telecommuting jobs than those with slow service. There are lost opportunities from slow networks. Productivity is lost. Markets for network products and services are underserved. Many interactive opportunities such as on-line games are served inadequately, losing major opportunities for revenue and serving people. There is undetected demand for products and services that isn't being met fast enough. If bandwidth is out there people will suck it up. On-line games are a big market. There are groups of 11 and 12 year-olds participating in on-line games, setting up web pages, and spending lots of time and money on-line.
We got to our present state by the astounding technical skill of the early network pioneers, developers, government support, following the path of least resistance. How did we get where we are in Palo Alto? A City telecommunications study that was just implemented. The City took a low risk, reasonable, incremental approach. Moores' Law also applies to the Internet. The industry is not prepared for the explosive growth on the Internet. Uses and capacity will continue to grow enormously. The City can take several approaches - wait and see and don't spend anything at the risk of being wrong, let someone else do it, or develop a real model.
The Palo Alto telecommunications study cost $656,000 on the fiber ring, or less than $27/house or less than $13 per capita. Fiber links to the house were estimated at $700/user in 1996. That rivals the $500 network computer. Amortized costs over 18 years is just $21/house. Actual costs depend on number of users and state of the municipal data network to be upgraded.
Michael's recommendation is to build the information Superdriveway. Costs are a barrier but they can be justified if we want it enough. A universal data service can be modeled. Maybe the best way is with some sort of non-profit inter-industry consortium, or some combination of public and private sectors. Such a model would allow all participants to share unavoidable costs, benefits, blame and credit.
There are a number of unanswered questions, many of which began with this group. What speed access do people really want? What will they pay for various services? Should it be high speed both ways or 1 way? Should access be via TV or computer? How can advanced telecommunications improve education, reduce commuting, make the area more competitive, improve the quality of life and empower the people by allowing high-capacity 2-way interaction? These all are very good questions and it will be very interesting working together to find the answers.
The impact of competition wasn't mentioned much. That could be important.
Bob Moss noted that Pac Bell made it clear that they want to keep the classic telco model. SBC has been very successful in buying politicians in Texas, and they will make the same efforts in California where our politicians are equally purchasable, although maybe at higher cost (higher cost of living here). Telcos won't offer high speed access any time soon, especially after they pulled out of San Jose. DBS has very limited success here. Cable Co-op now is offering high speed access, and hooked up 20 users the first 20 days, and had 11 waiting installation. Another 5 got the high speed 1-way modems. This is without really announcing it. The idea is to avoid having a big surge of people asking for high speed modems all at once. It is better to ramp up slowly and learn how to do it correctly before getting too many users. They are just starting the promotions for cable modems. The idea is to walk before we run.
Stan noted that the superhighway already is here and is being used extensively today. People want the service, but the question is what do they need and what are they willing to pay. It isn't a technical issue but a cost-benefit issue. The question is how is it going to develop into the network. The technical details can be worked out later.
Keith Cooley disagreed. The technical aspects will drive both performance and cost and have to be taken into account. The technical decisions made now will have influence for the next 20 years. All services will be migrating to the Internet - video, telephony, everything.
A participant said his grandmother doesn't want to use the Internet. Kathryn Martinez said that when she is a grandmother she sure will want to use it, and not planning now means it won't be available then.
Stan Smith said the issue is how will it evolve, and for whose benefit?
It's a social issue, not a technical issue. Think if this was 1987 and someone said we want to dam the Hetch Hetchy and bring water to the peninsula. Think of all the EIRs, all the controversy. Would we end up doing it? So we did it and we have water. But if LA hadn't done it for their water would San Francisco have done it?
Michael noted it's also a competitive issue. It can be important for a Pac Bell to co-operate with others to spread the risk. Once the initial problems are worked out there will be a rush to skim the cream. The difference between information superhighway and Superdriveway wasn't clear enough. There already is access to the superhighway by major users. The Superdriveway is to homes. A Cable Co-op model had fiber to 500 home nodes, and there was the issue of where do we put the ugly boxes, do we put them underground, etc. Fiber reduces that need and makes it look better and better. But the Superdriveway is a local network, where the same neighborhood is tied together as with ethernet, like universities. When he was reading the Wall St. Journal on line it was in the graduate student townhouses, where they had a LAN, and it was a neighborhood network. Sending big files back and forth, they stayed local and didn't clog the external switches.
Harry said the cable modem he has provides a LAN connection which goes to the Internet. We need to talk about high speed networks, but we also need to talk about what people actually will do with them.
Bob noted that the cost structure for cable modems was done to make it more economically practical, keep from underpricing it so too many people rush to hook up at once. The design of 500 home nodes is intended to allow expanding service with fiber to fewer homes, or fiber to the curb as demand warrants. If lots of people locally want high speed access it can be broken down as much as needed. It's very feasible to upgrade service as demand warrants.
Imsong pointed out that if more people use high data rates at the same time the usable speed goes down. In order to keep high data rates the number of users per hookup must be limited which increases costs. It's an economic issue.
Harry said it isn't a natural monopoly. There are several ways of getting data to the home now. The cable modems use a local ethernet setup now.
Jay said the important point is to allow the ownership and use to be decentralized, not in the hands of a few. We have that now, but it may not last.
Paul Pease said we are having a program Sept. 25 to discuss the issues before city candidates. What are we supposed to be getting across? How can we provide examples of costs? How do we convince people that it is both desirable and economically feasible?
Marvin Lee noted we already have 12 co-sponsors for the program, including the Palo Alto Weekly, without major efforts to sign up organizations. People are interested in the subject. Tom Passell noted that the wireless people couldn't give firm cost estimates on setting up systems. Michael said he can get some figures for installing residential LANs and try to have them next month.
Jay wants to get a packet of information available for the meeting on the 25th. Mary Jo has people working on that.
Harry said there are lots of details on finances beyond the raw cost to pull wires. What are the wires connected to, who provided Internet access, what content is provided, for what cost? The problem is serving single family homes, not high rise apartments or offices.
Stan asked how the right of ways will be obtained, how many people can be tied together, how do they connect elsewhere?
Michael thinks they can get small groups of neighbors together for local LANs and not need public right of way.
Mary Jo said the city can help with the access issue if there is a co-operative program.
Jay said we have to demonstrate the advantages of high speed access. We have to get across that we can't do it ourselves, the city fiber is vital, but now how do we connect from the fiber ring to the home? We can border on technical information by providing background information, but not make it a major part of the formal presentation. Some physical neighborhoods are spread out and hard to connect. There also is a neighborhood of interests. We don't want a global village, we want lots of separate neighborhoods, tied by interests.
Harry said a model for data service is to let anyone who wants to tie into the dark fiber at a significant cost. The market can determine if they want ISDN, cable modems, wireless, etc. Should investment be on the core infrastructure or the last mile?
Jay said a significant lesson from the "fiber ring" was that it took a collaborative process to get it built. The city almost certainly would not have taken the initiative on its own, and in fact had real concerns about it. The Technical Advisory Committee initially had a strictly technical approach without considering the broader community aspects. Citizens and the business community alone could not have achieved it. Now the question is how do we connect the high speed broadband access to all the homes and businesses that need or can use it? Perhaps a similar process can work for the "last mile" connection. The mode used should be transparent. Use whatever approach is most effective. Use the technology and show how we can get there from here. Show that it is both desirable to go there and feasible to get there.
Paul said we have to show that it works and is practical. He had no volunteers to help with PR yet. Bob warned that if we give cost estimates people will tend to consider them real and quote them in the future. His experience is that initial cost estimates are always wrong.
There will be no formal meeting in Sept. The Sept. 25th forum will replace the regular meeting. The committee will meet as required to finalize plans ad the program.
Meeting adjourned at 7:00 PM.
Submitted by Bob Moss