Figure 1. REUNA's Operation Centers and Points
of Presence
Figure 2. Connections diagram
1. Introduction
2. From BITNET to Internet: The Birth of REUNA
3. Chile's Telecommunications Evolution
4. The REUNA Project
5. The Finantial Model
6. Operation Centers: A Key to Success
7. Conclusion and Perspectives
REUNA's network is supported by three major operation centers located in Santiago (the Capital of Chile), Antofagasta (in the north) and Concepción (in the south). These operation centers concentrate the traffic from the points of presence in the capitals of 11 of the 13 Chile's administrative regions (See map below). The speeds of the connections have been recently updated so that the minimum speed is a 64/32 Kbps Frame Relay connection and the international link is a 512 Kbps IBS circuit to SURANet. The details are given in the connection diagram below.
Using this network REUNA connects over 40.000 users to the Internet among which are private commercial organisations as well as public, non- governamental organisations, private universities, public libraries, schools and individuals.
The growth of traffic to and from the Internet has been growing at impressive speeds, specially in the last months when the growth of new users became over 20% per month. The graph below shows the growth in the traffic of REUNA customers. We have excluded the traffic from member universities and CONICYT. We observe that the growth in this sector is largely over 20%, the average growth is indeed 46.6%.
Figure 2. Connections diagram
How can an Academic Network cope with the cost increase of the upgrades, the requested information services, the new infrestructure, the personel ?. A common answer in the academic sector has always been state support. Unfortunately (or fortunately) in Chile the governement is strongly against any form of subsidy, specially to the universities. Thus a self funding approach was established.
Due to this decision, from the very beginning REUNA started to search for external customers. As in many countries these customers were primarily other educational institutions, public organisations and non-profit as well as non- guvernamental organisations. This fact changed dramatically when the new structure of the Internet was adopted making possible a completely commercial approach. From that moment on we started a very aggresive plan to connect all kinds of organisations to the network, multiplying types of access and executing a marketing plan aiming to establish REUNA as the network of choice for commercial and non-commercial use of the Internet.
At present REUNA offers a variety of services including connectivity in several formas: dial-up, digital circuits, Frame Relay and ISDN. We also offer information services, training, consulting and have deals with several hardware providers to simplify the customers approach to the Net.
REUNA is self-funded already and pretends to be a major player in the open commercial competition that is starting to take place in Chile with the installation of several Internet Service Providers that will take place in the fisrt semester of this year.
2.- From BITNET to Internet: The Birth of REUNA
The birth of REUNA goes back to 1986 during
National Meeting of Academic Computing Centers
and Computer Science Departments of Chilean
Universities. Even though the idea of creating an
academic e-mail network could be traced back to
1984, it was only during the 1986 meeting that two
major facts ocurred. The first one was the
announcement by 4 computer science departments
of their intention to start a UUCP network instead
of the original idea of using X.25 based protocols.
This was very important because it was a concrete
step that needed almost no investment hence it was
really carried out in a short time. The second fact
was the donnation of IBM to several universities of
all the hardware, software and leased lines needed
to establish an SNA network among them. These
aparently oposed lines of action were collected
under the name of REUNA and plans for a gateway
were established. The gateway was working at the
beginning of 1987 giving birth to the first e-mail
network in Chile. This network included
universities from Antofagasta, Concepción,
Valparaíso, Santiago and Valdivia.
In December 1986 the UUCP part of the network started to communicate with UUCP thru INRIA in France and after some months of operation moved its entry point to UUNet in the USA. By april 1987 the gateway connecting both netwroks was operating and so REUNA was communicated with the world thru UUCP.
A more stable connection to the growing international networking community was established at the end of 1987 thanks to an agreement between NASA and the University of Chile. The agreement allowed chilean universities to share the use of a voice line connecting NASA's Santiago Tracking Station, operated by the University of Chile, and Goddard Space Flight Center in Bethesda, USA. The line was used for communications of the staff to the US during working hours and also during space missions. The line was thus essentially unoccupied during the nights and on weekends. A switch installed and operated by NASA personnel in Santiago and the US enabled the line for REUNA usage during those hours allowing to connect the University of Chile to the University of Maryland and from there to BITNET.
From the end of 1987 to 1991 the increase in traffic was exponential. The rate of growth was even larger than international average amounting to a factor of 3 every year.
Already in 1988 the idea of getting connected to the Internet (or to the NSFNet as was essentially known at that time) was present in REUNA. The main problem we faced was the cost of a 56 kbps link to the US, roughly US$ 12,000 per month. The idea of sharing costs was difficult to implement because the need for this new service was not well established in the academic sector. Also CONICYT was not interested at that time and was not willing to add that amount of money to their support for science and technology.
Until then REUNA was only a name for the group of universities involved in the development of the network with no official support and no involvement of university administrators. The idea of creating a consortia associating the interested parties appeared in 1990 when the recently appointed President of CONICYT invited the Rectors (Presidents) of all traditional universities to appoint official representatives to a commitee aimed to establish the bases for a more permanent development of an academic network.
After several months of work an agreement was reached between 19 of the 22 traditional universities and CONICYT. There would be a Non- Profit Organisation called REUNA that would develop, administrate, operate and commercialize an IP network in Chile connected to the Internet. The choice of Non-Profit versus Profit was essentially made to preserve the academic spirit of the network. Nevertheless, it was clear from the very beginning that the organisation should be self- funded and would sell its services commercially to other organisations in Chile.
REUNA's connection to the Internet was established on January 2, 1992 and the first year of operation was funded by CONICYT. Even though some non-member organisations got connected to the Internet during 1992, that year was essentially devoted to complete the network described in the Introduction, install the operation centers and train technicians from member institutions in the use and administration of the Internet.
The commercial start of REUNA was 1993. After a slow beginning the monthly growth rate of 13 % average was reached in 1993 and 17.5 % in 1994. The present rate is larger than 20% per month. This results is not only due to our marketing strategies or technical quality, it is also the effect of a large press coverage during the last year. Nowadays an article on the Internet appears everyday on the newspapers and even TV shows are talking about it.
REUNA is legally existing since August 1994. Since then the personnel has been multiplied by 3 and incomes by 8. Plans for 1995 aim to multiply the number of users and incomes by 4. Even considering the entrance of major commercial competitors. These figures are due to a collection of factors among which our finantial model appears to be crucial. Nevertheless, a major role in this growth has been played by an external factor: the impressive evolution of our telecommunications infraestructure.
Indeed some private telecom operators were
installed since 1984 in the sector of Packet
Switching. A hole in the legislation protecting Entel
and CTC had allowed the creation of ChilePac and
VTR, two X.25 service providers born from old
Telex companies, that could compete against
EntelData, the packet switching subsidiary of Entel.
In 1985 Entel and CTC were sold to the private
sector and a limited competition was allowed in the
long distance services. Telex Chile, the company
owning ChilePac installed their own satellite
service and started serving part of long distance
phone calls from CTC customers. The phone
company was the one determining how much traffic
would gto thru Entel and which one thru ChileSat.
It was the starting point of a series of private
investments in the telecommunications sector that
allowed several advances in a few years:
- Creation of three digital data networks in
Santiago in 1990. This was a crucial change for the
installation of networks, before that only voice
quality leased lines were installed for data
transmission and no guarantee was provided by the
operator. CTC Data Red, VTRNet and Teleductos
quickly installed a network of fiber that covered the
city making it possible nowadays to obtain a 64
Kbps link or faster in less than 30 days. Either
alone or associated with other operators the three
companies now offer the service country wide.
- Cellular phones are available since 1989 and
its use is widespread since 1992. Four companies
offer their services nationwide.
- VSAT services started to be installed in 1990
and are offered by three different companies.
Among protocols used are X.25 and TCP/IP.
- The telephone network is fully digital since
1992.
- ISDN services are available since 1993 and are
becoming an insteresting way to access Internet
services.
- Frame Relay networks are operating since
1994. An important cost reduction lead REUNA to
choose this service for its national network. Speeds
rangin from 32 Kbps to 2 Mbps are possible.
In 1994 deregulation was pushed to the limit
when long distance services were opened on a free
competition per call basis. This means that the final
customer can choose the long distance carrier for
any call he wishes to make either nationally or
internationally. No contract is necessary and the
carrier number is need in any long distance call. Of
course the customer can choose to have a contract
with one of the carriers but he can change at any
moment.
This deregulation produced at least three
interesting facts: a) Prices were dramatically
reduced; b) Investements in fiber optic
infraestructure have created a tremendous
telecommunications backbone in the country; c)
Investors from abroad were atracted by the business
oportunity. Among new investements several
companies are launching a PCS service in June.
The final deregulation step was announced last
January when local phone systems were opened for
competition. In July 1995 at least two companies
are starting a local phone service in Santiago, new
services will soon come in other cities.
It is clear that avalilability of
telecommunication services is key to the
development of any kind of MAN or WAN. Clearly
the development of the Internet in Chile is not only
due to the marketing strategies of REUNA, the
coverage given by the press or even the agreements
signed by the governement with the US in order to
expand the use of the Internet. The basis for the
development is this wonderfull telecommunications
infraestructure developed by the private sector.
The idea of FONDEF was to contribute with
funding for research, the development of an
infraestrcture or the establishment of a service that
would benefit not only the universities or the
research institutions, but mainly the private sector.
It was conceived as some kind of technology
transfer process. The grant was to be spent in 4
years at most at a clear proof of economic value
had to be given.
The consortia agreed to submit a project to
FONDEF in July 1991. This fact produced several
definitions inside REUNA. Among them: a) the
service had to be offered to the private sector; b) a
self funding structure had to be adopted to
guarantee economic value.
A brief summary of the finantials of the project
is:
a) Investment by FONDEF
b) Investement by the Consortia Members (4 years)
As a counterpart REUNA had to generate
incomes from the private sector so that the Internal
Return Rate was at least 13%. Indeed the economic
evaluation of the project lead to an IRR of 43% in a
10 years evaluation period. Moreover, the real
figures have been well over the estimates so that
the 43% IRR is really going to be reached in a 5
years period.
The project was approved in September 1992.
From a technical point of view the project
planned to install three Operation Centers in
Santiago, Antofagasta and Concepción and connect
all university members in a year period. Even
considering that bureaucratic process produced a
considerable delay in the provision of the
equipement, all technical goals were reached by
December 1993, two months after receiving the
routers and workstations.
The rapid growth of the Internet in the world
and the already described rates in Chile lead
REUNA to make a second investement in
equipement, marketing and personel. This time all
the investement is coming from the benefits of the
operation: US$ 400,000.
With the new investements more dial-up access
have been installed as well as ISDN for digital
access and E1 connections to the main carriers. The
project has gone well beyond its original goals.
These prices are certainly higher than those in
the US. So it is not surprising that final costs of
Internet access would be higher. Yet, the main
problem is not the local cost. The cost of a satellite
access to the US, cheaper than to Europe and same
that to any country in South America, is the real
high cost that has to be considered in developping a
funding model. The following is a list price for IBS
services:
Since most of our traffic goes to the
international link, which we pay completely, we
must consider the increase in the nessary
international badwitdth when defining our prices.
Thus, even if a new customer pays for his own link
to REUNA's operation center, our cost due to the
use of international links will not be fixed. The fact
that 95% of our traffic goes to the Internet makes it
possible for a 64 Kbps user to use as much as 61
Kbps of international traffic. A realistic flat rate
should consider a high price for it, specially when
the total international bandwidth is not large
enough to get substantial discounts for additional
64 Kbps portions.
Another important factor to consider is the fact
that a very low speed line, like a 2.4 Kbps is
absolutely not appropiate for IP access. Such a
portion of bandwidth could be priced at a flat rate
and a reasonable price, but who would like to
access a WWW server using such a link ?. A real
minimum access speed for a router connection
should be 19.2 Kbps. The point is that people
would love to have fast access to the Internet when
they want, but are not ready to pay the price of a
permanent unlimited traffic link at a high speed.
The only solution that we found to this puzzle
was to establish a per traffic fare. We are aware of
the fact that this implies to pay even for the traffic
that a user's ftp server generates. It is unfair to
charge a person who is willing to share information
with other people for free. Yes, but it is also unfair
that Chile's traffic to the Internet has to be paid by
Chile in both directions.
Of course a per byte charge is something people
would like to avoid since even a mistake during a
given month could produce a large bill that could
destroy the institution's budget. Our solution to that
problem was a mix of flat and per traffic rates: the
Virtual Bandwidth.
The customer chooses the amount of Megabytes
that its institution will be committed to use among a
set of possibilities. Let us say for example that he
chooses the upper limit of 160 Mbytes. During the
following 6 months the institution will be charged
at the flat rate corresponding to that limit without
regard to the real amount of international traffic
they use. After 6 months the average traffic is
calculated. If that average exceeds the limit by
10%, the institution will automatically be moved to
the next choice of virtual bandwidth. No payment is
made for the excess traffic during the preceding 6
months.
In this way, a customer can choose a virtual
bandwidth say corresponding to a physical 2.4
Kbps, and pay that amount, but his access can be
fast when he needs it.
In the table below we give examples of virtual
bandwidths and their current prices in US dollars.
An 18% VAT tax must be added.
The price strategy is completed under the
following conditions:
a) Given that 95% of the traffic goes to the
international link, there is no interest in charging
national traffic separetely. A uniform flat fee is
used. This contributes to incentivate national
traffic.
b) All News traffic as well as listserv's is
considered part of REUNA service, provided it is
peered thru REUNA's hosts.
c) E-mail is also considered a service and no
charge is made for that traffic provided it is routed
thru REUNA's mail exchangers.
This strategy has succeeded in at least two
aspects:
i) The increase in number of customers and
consequent traffic increase, as well as traffic
increase due to new internal users of the customer
institutions, has been followed by increase in
international and national bandwidth.
ii) Customer acceptance has allowed a greater
market share.
It is clear that our solution to volume charging
is not of universal use. Nevertheless we feel it is of
use in the case of high international costs and has
proved to be a successfull mean of atracting an
increasing amount of customers.
a) Availability. A 24 hours a day, 7 days a week
continuous operation is assured thru dedicated
personel. An extended working day going from 7
AM to 23 PM and an off hours alarm system is
possible because our operators work only in
providing Internet connectivity, no other service
must be assured. Operation centers separeted from
the member universities is key in providing this
service.
b) Support. Thru the operation centers our
customers receive the support they need, they do
not compete with other services. Again the
dedication of the personel to these tasks is crucial.
c) Information Services. Our Information
Department guides the users in navigation thru the
Internet making it possible to new comers to start
using it painlessly.
d) Accesibility. The diversified
telecommunications possibilities makes it easy to
reach REUNA from various platforms.
In the first three aspects we feel that a dedicated
technical personel has been the key to success. No
university computing center could successfully
carry out these tasks without weakening the
services given to their natural customers: the
academic comunity of their institution. As in every
university activity, the mixture of external and
internal services is usually a bad mix. From our
point of view, Internet Service Provision is an
activity that should be isolated from the academic
support.
On the other hand, the relationship between
REUNA and its university members has also been
crucial in the development of training seminars
with the cooperation of university staff. Both,
customers and university staff have benefitted from
these activities. During the last three years we have
developped several seminars in Santiago and cities
in the north and south of Chile. These activities
have been important not only in the training of
customers, but also in the obtention of new ones.
In this last point the legal structure of REUNA
is very important. Indeed a cooperation between
non-profit organisation is very natural thru a non-
profit structure. Additionally, in the case of
networking it is clear that manu resources are
needed to continue the upgrade of links, servers,
routers, etc. The obligation to re-invest the proffits
appears as a very convenient way to impose
upgrades on the network instead of using them for
other university activities. Under this philosophy
REUNA becomes a space for the development of
networking beyond the use of the Internet itself.
The attention atracted over the Internet by the
Press and service providers marketing strategies
have made it an attractive business oportunity for
carriers and other investors as cable companies,
new service providers are expected to appear during
1995 to serve the business community.
The largest journals of Chile are already
preparing their commercial edition over the
Internet. Experimental services should appear
during the second semester of 1995.
Information providers are rapidly becoming
connected to the Internet. A large amount of
chilean commercial information is expected to be
available on a per fee basis also during the second
semester of 1995.
Despite the competition from carriers and other
service providers REUNA feels that it is in good
position to maintain its rate of growth and keep a
reasonable market share in an expanded universe of
potential customers. If the number of customers is
only doubled every year there should be 1,000,000
users by the year 2000. There is room for many
providers!
Utreras got his Mathematical Engineering degree in
1975 from the University of Chile and his Doctor of
Engineering degree in 1979 from the University of
Grenoble in France.
Utreras has 7 years of uninterrupted work in the
networking area. He first conducted the connection
of the University of Chile to BITNET in 1987 and
then the connection of REUNA to the Internet in
1992. Its experience include the direction of
University computer centers and consulting in the
networking field.
3.- Chile's Telecommunications Evolution
As in many latin-american countries the
telecommunications companies were state owned
and essentially monopolic until 1985. Essentially
two companies served the whole country: Entel, the
long distance carrier and CTC, the phone company
which was not allowed to install its own long
distance facilities. Exceptionally, in the south of
Chile the telephone operator was private:
Telefónica del Sur, also in part of Santiago's
uptown a private operator, Telefónica Manquehue
was allowed to serve those customers. 4.- The REUNA Project
In 1991 the group of universities coordinated by
CONICYT was willing to establish an IP network,
but many member were reluctant to make the
necessary investement in equipment, personel and
infraestructure to deploy that network. Indeed at
that time the benefits of the Internet were not clear
enough to everybody so that priorities could be
changed in all the universities. It was necessary to
obtain some kind of support to start the network.
The support came from a recently established
National Fund for Private Sector-University
Relations: FONDEF (Fondo de Fomento).
Equipement provided US$ 700,000
Total Grant by FONDEF US$ 1,000,000
Infraestructure,Training,etc. US$ 300,000
US$ 900,000
Total Investement Money US$ 1,900,000
5.- The Financial Model
The telecommunications facilities described in
paragraph 3 are certainly very important and allow
the deployment of a good network. Yet the prices
are not as good as what can be got in the US. Below
we list some prices for illustrative purposes:
19.2 Kbps inside Santiago US$ 180 -US$ 234
64 Kbps inside Santiago US$ 234 -US$ 322
64 Kbps Santiago-Concepción US$ 700
64 Kbps Santiago-Antofagasta US$ 965
9.6 Kbps US$ 4.180
64 Kbps US$ 7.600
128 Kbps US$ 12.320
256 Kbps US$ 21.120
512 Kbps US$ 30.980
1.048 Kbps US$ 49.280
1.510 Kbps US$ 63.360
Virtual Bandwidth Cost
UT 5 US $ 45.43
UT 10 US $ 84.53
UT 20 US $ 121.90
UT 40 US $ 182.91
UT 80 US $ 307.62
UT 160 US $ 517.35
UT 320 US $ 870.08
UT 640 US $ 1,463.30
Where 1 UT equals 1 Mb during workin hours
and 2 Mb during non-working hours. This means
that during non-peak times the price of 1 Mb is
halfed.6.- Operation Centers: A Key to Success
REUNA has certainly been succesfull in starting
a comercial service in Chile. Competitors will soon
challenge its market share and future success will
depend on many factors, among which quality of
service will certainly be one of the most important.
We feel that REUNA is ready to face that
competition too. The main reason for this assurance
is that we are sure that our service satisfy the
following quality of service characteristics:7.- Conclusion and Perspectives
The creation of REUNA and the adoption of a
mixture of volume charging and flat rates has been
crucial to the development of Chile's Internet. With
the largest number of host/population in Latin-
America and a development rate of over 20% per
month this service is becoming a service of choice
for networking in Chilean community.
Author Information
Florencio I. Utreras is Executive Director of
REUNA and Full Professor of Numerical Analysis
in the Mathematical Engineering Department of the
University of Chile.