[2098] in Commercialization & Privatization of the Internet
NREN: whither connectivity
daemon@ATHENA.MIT.EDU (Brian Lloyd)
Sun Jan 26 16:19:22 1992
Date: Sun, 26 Jan 92 13:15:01 PST
From: brian@lloyd.com (Brian Lloyd)
To: com-priv@psi.com
Reply-To: brian@lloyd.com
One of the things that led into the development of the NetBlazer is
the concept that some connectivity (9.6Kbps to 64Kbps) is MUCH better
than no connectivity. As I work with BARRnet to provide both high and
low bandwidth services to BARRnet's varied membership it is clear that
very few yet need the massive bandwidth proposed by proponents of
gigabit capacity networks.
One of the things that people seem to be forgetting are patterns of
access. Most of the requirements are for local and regional capacity,
not contenental and global capacity. If you use the average campus or
corporation as a model, the demand for bandwidth is local and the
requirements for intercampus or intercompany capacity is more limited,
at least using the current crop of applications (I know, new
applications MAY change this).
I fear that the emphasis on gigabit networking will make it difficult
for elementary and secondary schools to afford basic connectivity. I
am working with my local school district to bring basic Internet
connectivity to all the schools. To keep costs down we are using
dial-up IP using PPP. The entry cost for the school is low because
they can use existing hardware (a castoff PC, a modem, and Phil Karn's
KA9Q IP/TCP package operating as a router). This they can afford or,
if they can't, they can often find castoff PC's from parents and/or
local businesses. This is enabling technology; gigabit networking
isn't. Once we have network literate students we have a groundswell
of demand that will make the gigabit networks happen. How? It will
become clear to the network service providers that they need
additional capacity becuse people will be purchasing service --
something that any business understands. They will add capacity as a
matter of course.
Let me give you an analogy: the highway system. Often the first thing
that happens is to add basic connectivity (a two lane road). This
road is instrumented (car counters) to determine the existing traffic
load and load patterns. When the load reaches a preset average level,
the road is expanded to provide greater capacity. The highway
department does NOT start out by building an 8-lane superhighway on
the off chance that, if it is there people MAY use it, or that someone
MIGHT produce a car or truck that can somehow make better use of an
improved road someday. "If you build it they will come," works just
fine in the movies for baseball diamonds in the middle of nowhere. I
am not sure that it is applicable to efficient, cost-effective
networking.
Let's take the highway department approach to connectivity. Let's
provide inexpensive basic connectivity and then expand the
infrastructure to accommodate increased demand rather than spend our
money on infrastructure without leaving any left over to help those
who need basic connectivity.
Brian Lloyd, WB6RQN Lloyd & Associates
Principal and Network Architect 3420 Sudbury Road
brian@lloyd.com Cameron Park, CA 95682
voice (916) 676-1147 -or- (415) 725-1392
