Greetings and Happy New Year to all,
Thanks Dan for raising my pet topic, namely:
> It has been interesting reading about some of the accounts of people
> that are trying to establish email accounts and Internet connections in
> the third world, especially Ethiopia. But one of the things I was also
> hoping to read and get an understand for was what you guys who have
> actually witnessed these developments think are some of the possible
> Case in point. I hear about the privatization option for small to
> ISP's in Ethiopia. How would that work ???
Users would have to dial-up using the existing phone lines, or in the
case of institutional users, they could lease a line and thus have a
permanent connection to the ISP. In both cases, the physical wires
would still belong to the PTT. For the ISPs connection to the
Internet, it could go through the PTTs internet hub(s) to connect to
the outside world (or other ISPs within the country.) This scenario
would be the minimal level of deregulation. ISPs would be competing on
the basis of the speed, quality and number of modems, customer
service, and pricing.
Taking it one step further, the ISPs could be allowed to connect to
Internet routers outside, still using the leased lines of the PTT.
Then, they could compete on the basis of bandwidth too.
One more step would be to allow choice at the physical layer, e.g. the
ISP would have its own satellite receiver and transmmitter and connect
to the outside directly. Even in the local loop, wireless connections
direct from customer premises to the ISP are feasible, and could even
be cheaper than traditional phone lines.
The ISPs could also be allowed to connect directly to each other,
using lines leased from the PTT or built by themselves, thus
progressively building a national/regional Internet backbone.
Keep in mind that in all of the above, the PTTs traditional phone
business could still remain a profitable monopoly. Even in the
Internet business (which is comparatively very small), it would not
lose out: first it would get revenue from leasing lines, and second,
the more users there are the more others want to join ("positive
externality" for the economists out there), so the result for the
PTT-ISP would be a share of a much large pool of users rather than a
monopoly on a tiny and stagnant user base.
Of course, the reality would be more complex than this linear
progression. Indeed, the whole point of deregulation is precisely that
there is no single solution track. The market will decide the mix of
Internet services that will best statisfy the demands.
Without competitive forces in Internet services, a country is doomed
to remaining a technological backwater. The Internet business it too
fast and too vital to be left in the sole hands of someee bureacracy
demanding forms filled in triplicate every time you want to
sneeze. The problem is not the people who work there, in fact these
are probably the very same individuals who would be great innovative
service providers if there was a dynamic environment. The problem is
that in a government monopoly, no one is free to take the necessary
risks. The focus becomes control rather than growth. Every decision
probably has to go through a single chain of command. Each person
along the way is a bottle neck, with papers piling up on their desk.
10 months ago Ethiopia got one Internet node, at least two years later
than some countries which started using e-mail *after* Ethiopia got
PADIS. Today, it still has one. In the same ten months, the world as a
whole went from about 16 million hosts to about 25 million (see for
example http://www.nw.com or http://www.genmagic.com). At this early
stage, the Internet in Ethiopia should be growing many times *faster*
than the overall Internet, which is entering a maturing phase of
The experience described by David Dion is ominous. At this rate, it
won't be long before there is a 10 year waiting list for email
accounts (just like there is for phones already).
Perhaps the best argument is this. Looking around us, virtually
speaking, right here in this mailing list, if we could follow the path
of this email to every reader's computer, we would probably discover
more ways of connecting to the Internet than anyone of us has ever
imagined. You would travel through a global jungle populated with
strange creatures named LAN, WAN, TCP, IP, IPX, ATM, SNA, Sonet, WDM,
CSMA/CD, Ethernet, FDDI, Token ring, OC-3, T1, ISDN, ADSL, etc. who
knows there may even be some machine somewhere converting voltage
levels into drum beats and transmitting the bits on sound waves like
Africans of old. The beauty of this thing losely called the net is
that there is no real limit to the software, hardware, protocols,
architectures, etc. that you can put together to send bits. Let the