[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Multimedia Moeglich Machen...
- to: debate@fitug.de
- subject: Multimedia Moeglich Machen...
- from: uzs106@ibm.rhrz.uni-bonn.de
- Date: Wed, 4 Dec 1996 17:14:41 +0100
- Comment: This message comes from the debate mailing list.
- Sender: owner-debate@fitug.de
Nach dem langen File der Kommunikationsverhinderer hier einer ueber
die Foerderung. Die Kommission war mal sehr stark im Bereich
von wireless, cordless etc. Vielleicht koennte sie auch mal eine
Empfehlung nach folgendem Muster machen, was die Freigabe bestimmter
Frequenzen betrifft. Eine Kompetenz wird sich schon finden lassen.
Ausserdem passts prima in die Initiativen zu "Kleine und mittlere
Unternehmen", das Thema hatten wir hier schon mal...
Was den deutschen Beitrag zu Rigos paper betrifft, stelle ich fest, dass
uns Schneider, Terra und Co schlicht belogen haben, was die angebliche
Befuerchtung von rechtlichen Schritten betrifft, oder waren sie wirklich
so doof, wie sie sich dargestellt haben?? Ausserdem hat der Vertreter
der Bundesregierung die Durchsetzbarkeit des Konzeptes zu optimistisch
dargestellt, wie wir an ULFs paper sehen konnten, namentlich der Reaktion
auf Nolte. Letzteres koennte Rigo ja mal auf jener Liste verbreiten, die
Nichtreaktion auf Nolte ist ja das gerade Gegenteil der Darstellung der
Bundesregierung. Oder ist Zundel mittlerweile geblockt??
Wie auch immer, was man mit ein paar freien Frequenzen anfangen koennte,
eines von vielen Szenarien, das Telephonnetz zu umgehen, und mehr:
Web page: HTTP://ksgwww.harvard.edu:80/iip/beyer.html
beyer
The Rooftop Community Network:
Free, high-speed network access
for communities
Dave Beyer, Mark D. Vestrich and JJ Garcia-Luna-Aceves
Introduction
This chapter describes a dramatic departure from the conventional means of
delivering high-performance end-user connections to Internet services. The
Rooftop Community Network (or "Rooftop Network") offers not just a new
technology, but a new economic model that relieves the current dependence on
phone or cable companies to propagate local access infrastructure. The Rooftop
Network uses innovative wireless technology to allow deployment of fast,
robust, community networks, which are constructed entirely by the end-users,
and which are free of monthly operating charges.
A user joins a Rooftop Community Network by installing a Rooftop Connection Kit
which couples a high-speed, digital radio with intelligent, packet-switching
software. A Rooftop Community Network is a self-managing web of peer radios in
which every Rooftop Connection can serve as a repeater in its Community
Network. Thus the Rooftop Community Network approach carries the original,
distributed Internet model to the next revolutionary generation by extending
the switching infrastructure all the way out to the end-users.
The Rooftop Community Network also enables its users to seamlessly connect to
the rest of the Internet by conforming to the Internet protocol standards. Each
user’s interconnect costs may be sharply reduced by sharing access to
these Internet links among a number of users in the community. A Rooftop
connection only uses network resources when there is actually data to transmit
or receive due to its packet-switching technology. This delivers continuous
connectivity for all users while still optimizing overall network performance.
Continuous connectivity means email is delivered instantly and every user can
publish information without tying up a dedicated line (with its associated
cost).
Sample applications that a Rooftop Community Network enables include:
<*> Untariffed, high-speed access to corporate client/server applications
by telecommuting employees working from home.
<*> Shared access to databases by multiple sites in a metro area such as a
hospital with its affiliated clinics and doctors offices.
<*> High-speed "Public Access Networks" or "FreeNets" that have no monthly
charges for access to local content.
<*> Rich, networked, educational applications linked to local schools and
public libraries.
<*> For-profit service providers offering the Rooftop Network as an
alternative, high-speed access means for their subscribers while
bypassing the local phone or cable company loop.
This chapter provides a brief introduction to the concept, technology,
applications, and policy issues of concern for the Rooftop Community Network.
The Concept
With the emergence of the Internet and the current deregulation environment,
communication infrastructures are becoming increasingly distributed. The
Rooftop Community Network further extends this trend by bringing the
communication infrastructure all the way down to the local community and
household.
To join a Rooftop Network, a user would purchase and install a Rooftop
Connection Kit. The main components of the kit are:
<*> An "Internet Radio" consisting of a high-speed (256-kbps to 2-Mbps)
digital radio for an unlicensed band, and an embedded microcontroller
that runs intelligent wireless network and Internet routing software,
and
<*> An antenna, smaller and simpler than a typical TV antenna, to be
mounted on the user’s roof, along with a cable (like that for
cable TV) to connect the antenna to the Internet Radio.
After installing the antenna, the Internet Radio would be connected to the
user’s computer or local computer network in one of three ways: 1) over a
simple serial connection to a single computer; 2) over a link to a local area
network such as Ethernet; or 3) over a short-range wireless link to desktop or
mobile laptop computers, an alternative that promises to become increasingly
attractive as the price of radios continues to decline.
Free Networking within the Community
A Rooftop Community Network is a web of peer radios, all automatically
participating in forwarding traffic for the network (see Figure 1). Each user
with a Rooftop Connection is part of its local Rooftop Community Network and
can send traffic to any other user in the network, for free. Unlike a cellular
network, there are no base stations that need to be installed or maintained.
Self-managing software protocols (discussed below) automatically control the
routing of packets across multiple links between their source and
destination(s). These protocols require no user configuration or intervention.
They ensure fairness so that each user receives an adequate share of network
bandwidth. The software also helps to ensure security so that traffic is safe
from eavesdroppers and malicious users. By using RF bands that have been
dedicated for unlicensed use by the Federal Communications Commission (FCC),
there are no license fees to acquire spectrum or operate the radio. Equipment
purchase is the only expense. Hence, there are no airtime charges to operate a
Rooftop Network.
<Image>
Figure 1 The Rooftop Community Network
The Rooftop Community Network serves as an extension to the world-wide Internet
or to any other private data network that uses the Internet protocols. Thus,
users are free to use all of the same, familiar, well-supported, and rapidly
growing base of network software applications that have become commonplace on
the Internet including standard web browsers, email applications, and news
readers.
Examples of applications enabled by the free community communication of a
Rooftop Network include:
<*> Community education networks. These may support a wide range of
applications from rich networked, multi-user, educational role-playing
games linking homes, local schools, and public libraries, to simply
supporting free email between students doing their homework and
teachers or professional mentors in their home,
<*> High-speed "Public Access Networks" or "FreeNets". Rooftop Community
Networks can further stimulate the current explosion of local FreeNets
by eliminating the cost and limitations of the local-loop phone lines,
and by making it easy for all of its users to be publishers of
information, due to their continuous connectivity to the network.
<*> Business telecommuter networks. Businesses can reduce the cost of
telecommuters by eliminating monthly access charges. Good corporate
citizens can support their communities by sharing their infrastructure
to seed neighborhood networks.
<*> Shared access to community databases. These may include access to
local Government records such as the minutes to town assembly meetings
or building regulations, or shared access to medical information among
multiple hospital sites and their affiliated clinics and doctors
offices.
The Rooftop Network also has a unique combination of characteristics which
makes it ideal for disseminating community information:
LocalityBroadcasts can be naturally limited to the geographic radius
appropriate for the message,ImmediacyMessages are delivered immediately (unlike
classified ads for instance),FreeThere are no fees or monthly charges for
communication within the Rooftop Network.
Examples of community information that can take advantage of these
characteristics include: Neighborhood messaging(e.g., requests for assistance
in locating a lost pet; invitations to upcoming garage sales); Community
events(e.g., announcements of upcoming fairs; distributions of the Town Hall
Meeting Minutes; emergency notifications); and Commercial events (e.g.,
"blue-light special" announcements at the local grocery store; solicitations of
customers’ opinions concerning which items to stock). Virtual
"subscriptions" can be used to help ensure that users only receive such
messages from information types and sources to which they have "subscribed".
Continuous, High-speed Internet Access at Low Cost
Rooftop Community Network users will also benefit from high-speed, continuous
(24-hours/day), low cost access to the rest of the Internet. With a local
Rooftop Community Network amortizing the cost of a single fast link to the
Internet, these Internet-access costs can be shared. The Rooftop Connection
which serves as a relay (or "router") into the rest of the Internet is called an
AirHead, short for Air-to-Internet-Router.
Continuous Connectivity. Unlike dial-up connections, a Rooftop Connection only
uses network resources when actually transmitting or receiving user data. This
enables continuous connectivity for all users without bogging down the network.
Continuous connectivity allows instant delivery of email rather than requiring
the user to solicit for new email. Continuous connectivity also allows a user
to become a provider of information without requiring a dedicated phone line
for the purpose. For example, a home could post descriptions, pictures, the
highest bid, and viewing hours for items in an on-line garage sale; the town
weekly newspaper could consist largely of links to articles, editorials, and
up-to-date classified ads located on the computers in people’s homes; or,
a family could post their family tree with web links to other family trees.
Email is free. Electronic mail remains the most valuable application on the
Internet. As an added incentive for new members, Rooftop Networks could promote
that all email traffic from anywhere on the Internet will be delivered for free.
For example, a traffic load of 150 kBytes of email, per member, per day, for
free would equate to 50, 3-kByte messages per member each day. For a group of
25 Rooftop Network users sharing an AirHead’s 128-kbps link to the
Internet, this would consume approximately 0.5% of the AirHead’s link
capacity.
Key Enabling Factors
Four ongoing trends make Rooftop Community Networks feasible:
1 The exploding, wide-spread popularity of the Internet, and the
accompanying availability of, and familiarity with, a wide array of
network applications that facilitate communication and information
sharing.
2 Recent development of high performance, low cost digital radios, which
have begun to bring these capabilities within reach of typical users.
3 The availability of unlicensed, wireless spectrum , enabling the
ad-hoc deployment and propagation of untariffed, high-speed Rooftop
Networks with no licensing costs or regulatory delays.
4 The advancement of software protocols for intelligent wireless
networks and the Internet, permitting the rapid and easy deployment of
self-managing, wireless Rooftop Networks, and the seamless integration
of these networks with the Internet.
Below we discuss these last two trends of particular importance for Rooftop
Community Networks.
Availability of Unlicensed Wireless Spectrum
In 1985, the FCC modified its regulations to permit unlicensed digital
communications that meet certain "spread-spectrum" waveform requirements in the
"ISM" (Industrial, Scientific, and Medical) RF bands of 902-928 MHz,
2400-2.483.5 MHz, and 5725-5870 MHz. The maximum output power permitted by a
transmitter was limited to 1 Watt, and the maximum antenna gain at the
transmitter was also limited (to 6 dBi when transmitting at the maximum power).
This ruling marked the FCC’s first significant step towards allowing the
introduction of new, unlicensed wireless networks such as the Rooftop Community
Network, and was sufficient to support the development and deployment of the
first of such networks to begin to test the general concepts [xx].
Intelligent Wireless Network Software
The software in a Rooftop Community Network is based on a technology known as
Distributed Packet Radio, or DPR. DPR protocols manage the self-organization,
routing, and security within a Rooftop Community Network. Standard Internet
protocols are used to handle the interface between Rooftop Networks and the
Internet.
The underlying concepts of DPR were largely developed during the 70’s and
80’s in research and development programs funded by the U.S. Advanced
Research Projects Agency (ARPA). These programs were given the task of
demonstrating a robust, secure, self-organizing, and highly adaptable
communication technology by capitalizing on the synergistic characteristics of
two new technologies: 1) the flexibility and reliability of spread-spectrum,
digital radios; and 2) the adaptivity and efficiency of packet-switching data
communications.
These R&D programs [xx], along with related efforts in the Amateur Packet Radio
[xx], and Internet [xx] communities, have resulted in a mature technology that
provides the self-managing, secure, efficient, and asynchronous packet-based
characteristics needed for the successful introduction of Rooftop Community
Networks.
Spurring the Internet Radio Industry
Current unlicensed, digital radios are designed for local area networks and
simple point-to-point or point-to-multipoint applications. In these topologies,
the real-time processing capacity of the radio is typically limited to simple
link reliability and channel-access protocols, with the more sophisticated
routing and control functions left to wired routers. However, to support
Rooftop Networks, these sophisticated functions are required of the Internet
Radio including packet forwarding, distribution of routing information,
security protocols, congestion avoidance, remote network management, as well as
more intelligent link and channel access protocols to deal with the more
complex, multihop radio environment of the Rooftop Network. In addition, to
ensure efficient use of the available spectrum, Internet Radios must be
designed to give the embedded protocol software finer control of the radio
characteristics (e.g., transmit power, frequency, processing gain).
Although current technology would permit the development of an integrated
Internet Radio with a high-volume, end-user price of $500 to $750, the price of
current Internet Radios, with board- or system-level integration of components,
varies from around $3,000 to $5,000. Introduction of an integrated Internet
Radio with an end-user price of under $1,000 is needed before large-scale
adoption of Rooftop Community Networks by average home consumers can begin to
be realized.
In addition, a key missing ingredient needed to energize the Internet Radio
industry, and help generate these lower cost Internet Radios, is a standard
"Internet Radio API" (Application Programming Interface) between the radio
platform hardware and the embedded network control software. The emergence of
such an API would permit the independent, un-coupled development of the control
software and radio hardware, allowing organizations to focus on what they do
best, and providing multiple sources for these two key components of the
Internet Radio system.
Scaling to Large Rooftop Networks
There are three main technical issues concerning the scaling of Rooftop
Networks to thousands or tens of thousands of users (or "nodes") in a regional
area: routing within the Rooftop Network, routing between the Rooftop Network
and the rest of the Internet, and RF spectrum congestion. These issues are
briefly discussed in the following sections.
Routing within Large Rooftop Networks
As the number of nodes in a Rooftop Network increases, so does the size of the
network routing tables and the number of control messages needed to update
them. One way to cope with an increasing number of nodes is to assign addresses
to them in a way that many routing table entries can be aggregated into a
single entry. For example, fixed nodes can be assigned an address that has some
correlation with their geographic location. By doing so, a node that is outside
a given neighborhood may only have to store a single entry in its routing table
corresponding to all of the nodes in an entire neighborhood, rather than an
entry for each node in the neighborhood. Nodes in the neighborhood will have
entries corresponding to all of the other nodes in their neighborhood. This
form of hierarchical routing is an adaptation of the hierarchical routing
scheme first proposed by McQuillan [xx] and recently extended to more efficient
routing algorithms [xx]. Also, to avoid "traffic congestion" in a large Rooftop
Network, traffic between distant nodes (e.g., greater than 6 hops apart) is
routed to nearest AirHeads for forwarding over the wired Internet.
Routing Between Large Rooftop Networks and the Internet
To allow a Rooftop Network to connect with the rest of the Internet without
forcing each node to maintain large routing tables to hold the many addresses
known for the rest of the Internet, the AirHeads can mask the majority of the
routing information from the rest of the Rooftop Network. The AirHeads can
collaborate to decide who should advertise a route to an aggregate of address
ranges in the Internet. Each AirHead then distributes routing information
messages to a few large address ranges in the Internet, rather than a large
number of information messages corresponding to very small address ranges. A
Rooftop Network node chooses the best match between the Internet address it
needs to contact and the address ranges advertised by AirHeads to select which
AirHead to use. Therefore, AirHeads are the only members of the Rooftop Network
that need to handle complex routing tables to connect to the Internet. A
similar "route aggregation" strategy by the AirHeads will also help to limit
the amount of routing information that the rest of the Internet must learn in
order to route traffic to particular nodes within the Rooftop Network.
RF Spectrum Congestion in Large Rooftop Networks
Typically, all of the nodes in a Rooftop Network will share the same overall RF
band for their transmissions. Thus, as the size of the Rooftop Network
increases, the traffic demand within this shared band will increase. This
raises the concerns of reduced network throughput per user, and of increased
likelihood that simultaneous transmissions will "collide". To avoid these
problems, a Rooftop Network can use a combination of methods including:
reducing the power of each transmission to only that needed to reach the
intended destination; scheduling transmissions within each local area to avoid
collisions; and synchronizing the source and destination of each transmission to
switch to a waveform that is less correlated with simultaneous transmissions
outside this local area. Also, by combining the above methods with sufficient
"spread-spectrum" techniques in the Internet Radio, and accounting for typical
path loss characteristics, it can be shown that Rooftop Networks remain
effective even with an infinite number of randomly distributed nodes [xx].
Regulation and Policy
Three specific areas of regulation and policy could have a significant impact
on the rate of deployment of Rooftop Community Networks. Two of these are local
issues, zoning and local taxation. The third is federal regulation of RF
spectrum.
Local Antenna Regulations
Local zoning would effect Rooftop specifically in the area of antenna
installation. There is already an explosion of new antennas occurring in
municipalities for a wide range of applications. Network operators will install
tens of thousands of PCS antennas in the next few years. Institutions will
install various microwave antennas of all shapes and sizes. Individuals will
install millions of small satellite television receivers, too. Local
governments are trying to cope with this flood, attempting to address citizen
concerns about aesthetics, plus health and safety.
For example, Medina, Washington, Bill Gates’ home town, recently obtained
a federal injunction, preventing several wireless network operators from
installing PCS antennas on a local commercial building. This location provides
particularly critical radio coverage of the bridge which is the main commuter
route between downtown Seattle and its eastern suburbs across a lake. The issue
in this case is neighborhood aesthetics. An affluent residential area would
prefer not to have an unsightly "antenna farm" populating its quaint downtown.
In this type of case, and others around the country, these antennas are
principally being installed and maintained by for-profit network operators.
However, antennas for Rooftop Community Networks will be installed by
individual citizens or local institutions such as schools and libraries.
Nonetheless, there are closer examples.
Direct Broadcast Satellite (DBS) antennas provide the closest parallel today,
to deployment of Rooftop Community Networks by individual citizens. The number
of these units in the US alone is over one million. These small DBS dishes are
installed for personal use by their owners. Even so, some multiunit
developments and certain subdivisions attempt to restrict or control their
deployment, again on aesthetic grounds.
The appearance of a Rooftop antenna is quite unobtrusive. A typical Rooftop
antenna might be a dipole type antenna. This is a fiberglass rod one inch in
diameter and three feet long mounted in a small bracket bolted to a roof.
Another might be a directional antenna, which approximates the general
appearance of the common TV antenna, but is significantly smaller. Either type
would be far less obtrusive than either a DBS dish or a regular television
antenna.
Thus, to maximize rapid, widespread deployment of Rooftop Community Networks,
we would strongly propose that any zoning regulations for its antennas be no
more onerous than those applying to television antennas installed for personal
use.
Local Taxation
Some municipalities around the country are attempting to impose sales taxes on
wireless PCS or cellular network operators, and on Internet Service Providers
(ISPs). Local efforts to tax Rooftop Community Networks would clearly have a
chilling effect on their deployment, even though they pose a curious case.
There are no access fees for traffic across the local Rooftop Community
Network. That traffic is free of charge because it uses the shared, local,
wireless infrastructure. Thus, the volume or revenue basis for any sort of
traffic-based tax would be exceedingly elusive.
The question of who the taxed party ought to be also becomes highly ambiguous.
The owner-operators of the local infrastructure are its users, who receive
nocash remuneration for participating in the network. The supervisory
institution for any given community network may be either a for-profit firm or
a non-profit institution such as a school or library. Any revenue either might
receive from network participants would most likely be for access to the rest
of the worldwide Internet and as such would probably be interstate or even
international traffic. It would be quite dubious for a local government to tax
this traffic.
We strongly suggest that the benefits in citizen participation and productivity
to be gained from rapid, widespread deployment of Rooftop Networks in a
community, far outweigh any slight revenue that could be generated from
attempting to tax them locally. Forward-looking municipalities should in fact
try to stimulate and possibly even subsidize their growth, rather than inhibit
network growth through short-sighted, regressive taxes.
Federal Radio Communication Regulations
In the federal realm, the principal regulatory issue affecting successful
deployment of Rooftop Community Networks is availability of suitable spectrum.
As a technical issue, the core wireless network protocols can run on any
intelligent packet-switch radio, operating at any frequency. From a practical
point of view, however, users will make a utility decision to deploy a Rooftop
Community Network by comparing the price and performance of Connection Kits
against other alternatives, wired or wireless. Thus, the location and size of
the frequency band available to operate Rooftop Networks, significantly gates
their appeal.
The selection of frequency has its biggest impact on the cost of Rooftop
Connection Kits to the user. In general, the higher the frequency, the more
expensive the components required to build Internet Radios. The size of the
band available affects the maximum data rate, maximum range, and ability to
effectively share the spectrum with other similar systems in the same area.
Useful, affordable Rooftop Community Networks can be deployed today in the
Industrial, Scientific and Medical, ISM bands, designated as unlicensed, shared
frequencies by the FCC. Using these existing bands, Connection Kits can be
highly competitive with ISDN adapters and even cable modems. However, all
technologies improve at rapid rates.
In the future, users will demand multimegabit speeds to access the Internet. To
achieve these rates, a Rooftop Network would require two to three hundred
megahertz of spectrum, with specifications similar to those that Apple Computer
proposed in its original submission to the FCC’s SUPERNet proceedings [xx].
The authors strongly endorse the general concept to create a new unlicensed
data band that would support multimegabit data rates over ranges of several
kilometers. We have repeatedly seen that spectrum auctions are dominated by
either large corporations, sometimes foreign, or financial operators that have
access to large pools of Wall Street capital. In the interests of economic
diversity, it would seem extremely sensible to provide one segment of spectrum
where communities and individual entrepreneurs could have the opportunity to
experiment with more innovative or even radical ideas that could deliver
significant economic and civic benefits to a broad array of citizens. It is
ironic that as we approach the end of the 20th Century, we see the US
Government using a 19th Century approach, namely auctions. We strongly suggest
that other, innovative business models should also be tried to create
breakthrough industries for the 21st Century .
________________________________________________________________________________
Acknowledgment
The writing of this chapter was supported, in part, by Small Business
Innovative Research (SBIR) project number DAAB07-96-C-D010.
________________________________________________________________________________
Notes
[TBD]
References
Beyer, D., Vestrich M., and Nguyen B. (1996). Distributed Packet Radio: What is
it? What good is it? How does it work? How can it help you today?" Rooftop
Communications Corp. white paper.
Beyer, D., Frankel M., et al (1989). Packet Radio Network Research,
Development, and Application. Proceedings of the 1989 SHAPE Conference on
Packet Radio, Amsterdam. Summary also published in MILCOM Conference, 1989.
Beyer, D. (1990). Accomplishments of the DARPA SURAN Program. Proceedings of
the 1990 MILCOM Conference. Monterey, California.
McQuillan, J. (1974). Adaptive Routing Algorithms for Distributed Computer
Networks, BBN Technical Report 2831.
Murthy, S. and Garcia-Luna-Aceves J.J. (1997). Loop-Free Internet Routing Using
Hierarchical Routing Trees, to appear Proc. IEEE INFOCOM 97, April 1997.
Shepard, T (1995). Decentralized Channel Management in Scaleable Multihop
Spread-Spectrum Packet Radio Networks, MIT Doctoral Thesis.
... [TBD]
... [TBD]