Community Communications Infrastructure
by Brian Reid, Director, Network Systems Laboratory, DEC, presented to PA-ComNet on June 11, 1998; minutes by Bob Moss.  Slides.  See also NY Times Article.

Brian Reid spoke on the Community Communications Network. He lives on Waverly within 62 meters of the city's fiber ring, but he can't use it yet, although he would love to. He introduced 2 of his co-workers who also were present. He is very proud that created the first real time on-line voting report, for the 1994 California elections. The State thinks they may be able to do it this year. It takes them 4 years to finally implement things like that.

Technology has changed history in 3 ways - violence and war, communications and transport, and food. To change things we can make the old way harder to continue (force), make the new way easier to use (seduction) or educate people to do it better (information). He gave some examples of technologies that have changed the world - negative numbers allowed banking and general commerce, stirrups allowed mounted warriors and more effective armies, Roman roads created better means of communications and trade and made the military more effective, telegraph provided fast and cheap information transfer, railroads provided the same for goods and people, electronic banking and ATM machines made commerce more efficient and more widespread. Inter networking will change the world more than anything else. People who work with inter networking don't really understand it because they think of it as technology, not how it is used.

Networking is about having control over the source, destination and manner of communication. Inter networking is different - it allows allocation of blame, finding an advantage, etc. The Internet is just a means of joining other networks together, and exists only to benefit users. It is all about relationships and agreements about what people want to do.

Phone companies won't put equipment in other phone company offices because of distrust. The internetworking world is different. There are Internet exchanges where competitors hand off data to each other and where they are located together. Palo Alto's Internet exchange is the biggest in the world. It used to be second biggest, but in the last few weeks it passed the former #1 exchange. Japan just opened an Internet exchange. Before that they had to send data to Palo Alto for distribution, even to others in Japan.

To engineer things properly there has to be customer feedback. If the government is involved that feedback is lost because government doesn’t feel the same need to get user reactions as businesses do.

Internetworking is what enables computer networks to talk to each other. Internet 2 and other supposed changes in the backbone and data transmittal are still inter networking. It is not a major new technology.

There are 3 important trends in communications - deregulation of power and telecommunications will allow many companies to get into and compete in new and different businesses. Computers are getting better and cheaper - they are everywhere. Computerization is expanding everywhere. Everything is connecting to everything else on a scale never seen before. It all will end up on the Internet. Locality matters, global is good but locality is critical. Internetworking is used to do things that are intrinsically local. It is nice to be able to communicate with someone in Scotland, but people really are interested in what happens in their town and neighborhood.

Deregulation gives us multiple players - several phone companies with boundaries -are addressing issues of the effects of deregulation on policy, cost, and scope of service. It isn't necessarily a bad idea to throw away existing phone companies if there are many more to chose from.

Shared vs. dedicated infrastructure is an issue. Utilities exist because it makes no sense to have competition in some areas. There are 4 companies that have wires to the house - phones, cable, ISP, and power. He asked for a show of hands of people with direct Internet access. About 9 people or close to 1/3 of the group have direct Internet access. Each company is getting into other businesses. Phone companies are delivering data, cable companies are delivering data and phone service, and power companies are interested in delivering anything that makes sense and is profitable. The best way to do it is to have an information utility with fiber optics and copper delivering information and power to all homes and businesses. Power is transmitted on copper wires, information is delivered on fiber to the home.

It took 100 years for the phone companies to get the connections and plugs right - the final design is the RJ 11 jack. When the information utility is connected to homes it needs to add a data meter for each customer, to measure how much of each type of thing going in is used - power, data, groceries, local and long distance phone calls, etc. Wireless transmission has limitations - it is very good for short distances, but not for longer ones. It's best for transmissions of less than a mile but it is not widely used in the US yet. It should be more popular in the future.

The important concept is having truly open competition and offering many different ways to do things.

Palo Alto is a good place for the Internet interchange because there is lots of fiber here already, a facility was available, and the location was favorable. The Internet Exchange building on Bryant is full and has no room for expansion, so the next step is to double prices for rent. That was done, and the tenants just paid the higher rates. With the city fiber ring plus the Internet exchange already here, anyone can set up as a major user of very high speed communications anywhere in the city. How do we hook it up to homes? It takes time, but if we work together it will take less time than if we just wait for things to happen.

Internet technology is used to do distinctly non-global things. There are physical limits to the phenomena. Data can only be sent so fast - send it farther, like to and from a satellite, and it will take longer for a reply than talking face to face. Public non-global networks are important. Governments enable shared infrastructure with fewer problems.

The information utility consists of the user (home) utility , and businesses which offer services. Community networking considers that the Internet is a big collection of networks, homes have networks, the home connects to the network and is part of it. Community networks are about ownership - who, how many, what do they do? Anyone connected to the Internet is part of it.

One model is to connect to ISPs from each home. Another is to wire houses together so they can talk to each other free and not use the ISP. If there is a local hub, homes can be interconnected, but while it is fine technically to tie into local hubs, it has real administration problems. There should be at least 2 hubs in the cityh for better service and Internet connections.

A good community network has low cost communication with local resources like schools, libraries, etc. High speed access is tied to ISPs for global access. Any customer can run local information services at low or no cost, and ISPs must be paid for long distance communications.

We could set up local hubs for the community, plus a central hub that connects to ISPs and to other organizations. The central hub should be a community resource center. It can be done cheaply. There is a major technical problem with routing which can be resolved. The central hub can be built and run by any organization or individual. The cost structure is an open issue. It is easier to build it technically than to figure out how to make it pay. Letting some company build the local network and run it is the way the power and phone infrastructures were built, and it took 100 years to recover from them. It is possible to grant monopolies but it is not the best solution. If it is built by the public, it will work more for the community.

Are there security problems? No, security really isn't an issue. In a community network security should be left to the members of the community, and they can decide how much is needed. Brian has lots of experience with firewalls and knows how to keep secure. He created many of the original firewalls.

Cost for residential access can be as low as $50/month for just bandwidth at 100 MBPS assuming the entire city is wired and everyone is connected. The 100% take estimate is based on great benefits for the power company which will cut rates in exchange for better information on power use. Service charges will be extra, but there will be free content such as neighborhood interactions. That pays for capital and operating costs over a 3 year period. How long should it take to get everyone hooked up? No more than 7 years, but 3 is better. The city tends to like to let others build it rather than do it themselves, because it's safer and low risk. Models can be worked out that are not a tax burden and that provide the services to everyone with low risk. John Kelly noted the cost estimate for wiring all homes and businesses for high speed access would be about $55 million at those estimates, assuming 31,000 homes, businesses and stores in Palo Alto. Brian said that was in the range of his cost estimates.

How big a test of high speed access is big enough? About 1%, or maybe 250 users, but that should include businesses, schools, homes, a wide range of users. City should have an interest in the bit paths as it does in sewers and roads. There are about 50 places that are creating community networks with public funding all over the world - Stockholm, Burbank, Torrance, Anaheim, Glascoe KY, many others. If Palo Alto doesn't move fast we won't be in the first 100 cities to be actively wired.

Internal nets for companies and universities are wired with the kinds of hubs being proposed for Palo Alto. Burbank and Helsinki already are wiring their cities this way.

The fundamental principal of the Internet is that by sharing infrastructure we get better communications. Where there is sharing there are boundaries. All parties must agree on the location and nature of the boundaries. Experience in the Internet has created a consensus of where to draw boundaries. Better communications enable new and enhanced communications. Everyone stands to benefit.